Luminous Landscape Forum
Equipment & Techniques => Digital Cameras & Shooting Techniques => Topic started by: ErikKaffehr on July 08, 2010, 04:53:25 pm
-
Hi,
There are a lot of discussion about DR (Dynamic Range) on this forum. It is quite obvious that persons with considerable experience have quite different opinion on the issue.
- One thing that I suppose we all agree on is that bigger is better as long as Depth Of Field is not an issue. A bigger sensor collects more photons and therefore will have better photon statistics. If we assume an MFDB sensor having the double area of a FX (Full Frame) sensor the theoretical advantage would be about a half stop.
- It is my understanding that normally the noise floor is dominated by shot noise (statistical variation of incident photons) and in the darks possibly by read noise. It is also my understanding that read noise is less on DSLRs than on MFDBs. The number of electrons each pixel can hold is about the same on sensors having the same pitch
So according to the above we would never see an advantage in excess of say one stop maximum on MFDBs over full frame DSLRs. DxO labs publishes detailed measurements that pretty much are consistent with the above observations.
It is my understanding that if we correctly expose right we would essentially have non specular highlights near saturation. In this case DR would show up as latitude for underexposure quite similar to increasing ISO. As a matter of fact, on MFDBs not having variable pre amplifiers underexposure would work identical to high ISO settings.
We would expect that if a camera A would have a 4 stop advantage in DR over another camera B it would perform identical to camera B at 16 times the ISO. So would an MFDB have a 4 stop advantage it would achieve the same image quality at 1600 ISO as the lesser camera at 100 ISO. But it seems that this is clearly not the case. MFDBs don't perform very well at high ISOs (except the PXX+ series).
Now, many experienced observers clearly see a 4-6 stop advantage with MFDBs over DSLRs. I don't have any issue with that, but I cannot understand where it is coming from.
It would be very nice if some could come up with a physically feasible explanation for MFDBs having a significant advantage in DR or samples clearly demonstrating the effect.
Best regards
Erik Kaffehr
-
Now, many experienced observers clearly see a 4-6 stop advantage with MFDBs over DSLRs. I don't have any issue with that, but I cannot understand where it is coming from.
Essentially one person is making these claims.
It would be very nice if some could come up with ........ samples clearly demonstrating the effect.
Wouldn't that be nice? Starting from clear experimental facts to build a theory?
- there is no theoretical reason why it should be more than one stop (not at image level, even less at pixel level),
- measurments show less than one stop,
- I have never seen samples showing more than one stop
I would propose we stop all DR discussion till the day somebody shows me on actual samples more than one stop difference in DR between a P65+ anda D3x files correctly exposed to the right. I will have zero problem to acknowledge I was wrong if presented with clear evidence.
And... please don't come and tell me it is hard to measure... if it is hard to measure then it is no more than one stop.
Cheers,
Bernard
-
but I cannot understand where it is coming from.
you have a chance - 3 slots left - http://luminous-landscape.com/forum/index....c=44724&hl= (http://luminous-landscape.com/forum/index.php?showtopic=44724&hl=)
-
- there is no theoretical reason why it should be more than one stop (not at image level, even less at pixel level),
- measurments show less than one stop,
- I have never seen samples showing more than one stop
I would propose we stop all DR discussion till the day somebody shows me on actual samples more than one stop difference in DR between a P65+ anda D3x files correctly exposed to the right. I will have zero problem to acknowledge I was wrong if presented with clear evidence.
And... please don't come and tell me it is hard to measure... if it is hard to measure then it is no more than one stop.
Cheers,
Bernard
Unless the laws of physics are suspended, you will not see such samples demonstrating a full stop difference in DR between these two cameras. The DXO (http://www.dxomark.com/index.php/eng/Camera-Sensor/Compare-sensors/%28appareil1%29/579|0/%28appareil2%29/485|0/%28onglet%29/0/%28brand%29/Phase%20One/%28brand2%29/Nikon) measurements, which place the noise floor for DR at a S:N of 1:1, show that the D3x has more DR than the P65+. If you set the noise floor higher, the P65+ would come out better, since the DR at the higher noise floor would be limited more by shot noise than read noise. The D3x apparently has less read noise, giving it an advantage where the noise floor is set very low.
Figure 12 of Emil Martinec's essay (http://theory.uchicago.edu/~ejm/pix/20d/tests/noise/noise-p2.html#SNR-DR) allows determination of the DR for a given ISO and noise floor for the Canon 1D3. One merely notes where the S/N curve for a given ISO crosses the x-axis for a given S/N and then reads off how many stops that represents on the x-axis. Where shot noise predominates, the slope of the curve approaches 0.5, whereas it approaches 1 where read noise predominates. Similar curves are available for other cameras, but apparently not for the P65+.
-
Thanks for the suggestion, would be interesting.
Best Regards
Erik
you have a chance - 3 slots left - http://luminous-landscape.com/forum/index....c=44724&hl= (http://luminous-landscape.com/forum/index.php?showtopic=44724&hl=)
-
Hi,
I think that there are more than one. Mark Dubovoy says about six stops, but he also mentioned that Jeff Schewe came up with four stops. I think Michael Reichmann came up with a similar conclusion. Anders HK doesn't really discuss DR but sees a real difference.
I'd also like to see an explanation to the difference in high ISO performance.
Some thing I'd point out is that most compare with Canons and at least according to DxO the Canons are significantly worse contenders than the Nikon D3X.
When Michael Reichmann compared 5DII, Alpha 900 and D3X he did not test for DR but he found image quality to be quite similar at low ISO. So even when the Nikon D3X has a significantly better image quality than the Sony Alpha 900, Michael could not readily see the difference.
Another discussion, http://www.imx.nl/photo/leica/camera/page176/s2part4.html (http://www.imx.nl/photo/leica/camera/page176/s2part4.html) really shows a difference between the D3X and Sony Alpha 900. I don't understand where that difference is coming from, either, but it's very much visible. There has been a lot of negative remarks on Erwin's testing, suggesting he does good lens tests and bad camera tests. Erwin may even test for DR in a future article.
Best regards
Erik
Essentially one person is making these claims.
-
When Michael Reichmann compared 5DII, Alpha 900 and D3X he did not test for DR but he found image quality to be quite similar at low ISO. So even when the Nikon D3X has a significantly better image quality than the Sony Alpha 900, Michael could not readily see the difference.
With all due respect to our host, his testing of the D3x was done during an antartica trip and less than 5% of his shooting was done with the D3x if I am not mistaken... Michael's focus at that time was not specifically DR by any means, and certainly not the DR of the D3x. Testing the D3x was one of his many goals during this trip and he had made it clear already that he was not that interested in that camera for various reasons.
Rightfully so. Snowy landscapes are definitely not the type of scene I would personnally pick to have a realistic understanding of the DR of a camera (I do a lot of snowy landscapes...). The only vaguely dark area is the sea and the sky where there is little detail to be found in the first place.
I have had access to A900 files and they are good for sure (probably a tad better colorwise in fact), but the shadows do not exhibit the same quality found in D3x files. There is also more noise in the mid-tones. I am telling you things the way they are, I would have zero issue with the A900 being better although it is 3 times cheaper... after a detailed computation of the cost for me (think system, lenses,...) the D3x ended up being cheaper than a switch to Sony anyway.
There has been extended discussions on why the D3x, although it is based on the same sensor as the A900, could have more DR. Simply put it is for the same reason Leaf back owners have been claiming that they backs have more DR than a ZD based on the same sensor. DR is controlled by shadow noise and the quality of the processing electronics plays a huge role here. The same can be seen in among DX cameras where the D90 has more DR than other cameras using the same Sony sensor.
Cheers,
Bernard
-
Hi Bernard!
I have no issue with the D3X being better than the Sony Alpha 900. In addition Nikon has Live View and Sony don't. The two things I wanted to point out was that difference between the two is clearly significant in DxO testing, but it was not really obvious in Michael Reichmann's review. Michael compared Sony Alpha 900, Nikon D3X and Canon 5DII in studio, but he was not discussing DR. To be more specific:
DxO (whom both you and me seem to trust) say that you need about 5 points on the DxO mark to see a visible difference. The Nikon D3X and the Sony Alpha are 10 points apart, but the difference is still not very visible in Michael Reichmann's studio shots. The Nikon D3X and the Phase One P65 score quite evenly on DxO but now we suddenly have 4-6 stops in DR advantage in some reports.
Michael Reichmann is more cautious in his own writing:
He says that
- P65+ i significantly better than Leica M9
- Leica is significantly better than Alpha 900
- Alpha 900 is significantly better than Panasonic GH1
He essentially states that there are visible and significant differences, doesn't say it's DR.
I actually don't understand how we can judge dynamic range from a print. Any print has significant manipulation of the gradation curve, we may add clarity and so on. Prints have a DR of about 7 stops so if we have more than that we need to do some tone mapping. Putting 12 stops into a DR of 7 stop without manipulation would result in a very flat print.
Regarding where the advantage is coming from would be interesting to know. It seems that the D3X has no external ADCs (?). Fourteen bit readout seems slow compared with twelve bits. Does the D3X have a different and slower on chip ADC? Someone suggested that they could take two separate readouts using different settings of on chip pre amps, but Emil Martinec says that would give a visible signature in raw files and that isn't there.
Now, if Nikon came up with a really unique idea they may probably not tell.
Also, camera electronics only affect readout noise. If we look at the Noise diagram the D3X has constantly less noise than the Sony, curves on Nikon and Sony are very similar and quite different from say Canon. But Nikon is simply better. My guess would be differences in CGA and microlenses. Another observation is that in Erwin Puts tests the Nikon is much sharper. I'd presume that is a difference in optical low pass filter, or possibly Nikon being better focused.
Best regards
Erik
With all due respect to our host, his testing of the D3x was done during an antartica trip and less than 5% of his shooting was done with the D3x if I am not mistaken... Michael's focus at that time was not specifically DR by any means, and certainly not the DR of the D3x. Testing the D3x was one of his many goals during this trip and he had made it clear already that he was not that interested in that camera for various reasons.
Rightfully so. Snowy landscapes are definitely not the type of scene I would personnally pick to have a realistic understanding of the DR of a camera (I do a lot of snowy landscapes...). The only vaguely dark area is the sea and the sky where there is little detail to be found in the first place.
I have had access to A900 files and they are good for sure (probably a tad better colorwise in fact), but the shadows do not exhibit the same quality found in D3x files. There is also more noise in the mid-tones. I am telling you things the way they are, I would have zero issue with the A900 being better although it is 3 times cheaper... after a detailed computation of the cost for me (think system, lenses,...) the D3x ended up being cheaper than a switch to Sony anyway.
There has been extended discussions on why the D3x, although it is based on the same sensor as the A900, could have more DR. Simply put it is for the same reason Leaf back owners have been claiming that they backs have more DR than a ZD based on the same sensor. DR is controlled by shadow noise and the quality of the processing electronics plays a huge role here. The same can be seen in among DX cameras where the D90 has more DR than other cameras using the same Sony sensor.
Cheers,
Bernard
-
An issue here is the almost total lack of hard technical information from the manufacturers. Compare that with film, where you have highly detailed specification sheets from Kodak, Ilford and so on for all their emulsions, with characteristic curves and everything you need.
The effect of different implementations of the same physical sensor would also be really interesting, beacause then we could examine how on-board processing of the image data affects the end result. For example, the same Kodak KAV 39000 sensor is used by various MF digital backs, but as far as I know no-one has compared say the Phase P45, Hass HD-39 or CFV-39 in terms of the RAW data they produce at a given ISO.
John
-
Compare that with film, where you have highly detailed specification sheets from Kodak, Ilford and so on for all their emulsions, with characteristic curves and everything you need.
And then... Each film maker had its own grain measurement, and we're essentially in the same boat : if not as desirable, noise has a bit the same subjective qualities as film grain, and therefore is hard to quantify.
My 2 (euro) cents worth :
- yes, electronics quality (read noise) seems to control at least as much quality as "laws of physics" (shot noise), if not more,
- moreover, what all measurements don't show so far is noise quality (http://theonlinephotographer.typepad.com/the_online_photographer/2010/03/what-tests-dont-tell-you.html) : is there any pattern in noise, such as banding or blotching, making it suddenly far more objectionable even if noise level (RMS) stays the same?
-
I'll try to be as constructive as I can.
I've put a disclaimer somewhere in Lu-La about that topic.
OP's question: Why is so that persons with considerable experience have quite different opinion on the issue?
IMO, simply because all these apparently contradictory statements have a part of truth.
I don't like at all those "in between lines" sarcastic comments on Mark that are diseminated each time this topic appears.
Putting constantly an etiquette on someone's under the pretext that he was wrong in an article is not a practise I enjoy reading,
very teenager behaviour. So as we are in between men I guess, it would be nice if ones for a while we could bring that debate ahead
without pointing Mark's "heresy". Because he might not be as wrong as we think and let me explain this point.
The DxO are NOT trustable on a real basis. Until you can not recognise that fact and take the DxO measurements as only an indication of a data,
you will not be able to accept the subjectivity and apparent contradictions.
So you are trying to understand something that seems absurd and questions will emerge over and over again without being answered
because the reference you are taking is not enough.
DxO is very similar to horsepower. They measure rigurously, in a scientific way, a value called DR.
A car can have 200 horsepowers, a camera can have 11 stops DR. This is a value, based on a sort of standard of measurement.
DxO is not even a worldwide standard like for example temperature measurement. DxO is french people who decided to do some testings following a specific
methodology. Just an observation. But anyway, let's say they are trustable. They are trustable only to bring some datas that have been measured in a specific
laboratory condition.
Now, we all know that horsepower is just one element of a car's performance. There are many other parameters that can make the differences.
In those, the conditions play an important role also. DxO is not trustable, not because they are bad, but because it simply can not explains
where and when those DR differences happen. They only give a usefull indication exactly like horsepower does. But the field is another story and can contradict these facts.
If you follow the formula one championship, maybe the most tech and demanding sport, you will see that those facts happen all the time: measurements that are not
trustable in this or that circuit, who are alterated by temperature, humidity etc...
If you relly on DxO to understand why those things happen, it is like you would relly on horsepower as a fix data to understand a fact, and of course, all that falls appart.
If the most tech demanding sport is actually unable to explain the differences by measurements, I doubt a little french comitee doing testing can give you any clew.
Then, there is another factor wich is the capture and the post production.
I think that the pictures in this thread are showing an impressive DR. Tipically the situation where you would prefer having a P65 than any other gear.
6 stops? No, of course not.
I don't know the D3x capability so I won't speak about it.
But then, there is the post-production task. What's called the room. The real information contained and the recuperation capacity.
Should we call that DR ? maybe that's the clew.
But in this area, the differences are indeed big.
In controled light, I doubt that the DR differences are significatives. But in others situations, they could be.
So all those concepts are subjectives, each case being a case in itself and maybe we should also consider the use of the 2 letters DR.
Cheers.
-
- moreover, what all measurements don't show so far is noise quality (http://theonlinephotographer.typepad.com/the_online_photographer/2010/03/what-tests-dont-tell-you.html) : is there any pattern in noise, such as banding or blotching, making it suddenly far more objectionable even if noise level (RMS) stays the same?
Absolutely, not to mention hideous painterly effects at micro detail level.
Cheers,
Bernard
-
Then, there is another factor wich is the capture and the post production.
Ideally, two captures of the same scene with of two different cameras will get the same treatment, so post production could be set aside in a "fair" DR comparison.
But yes, as the DR limit is essentially the one of noise in the shadows (we know how to expose don't we?), postprocessing and more precisely noise treatment can play a big role (LR users have been given a good notch more of DR with LR3 eg) and unbalance some uncareful comparisons, as well as greatly emphasize the noise structure problem (banding is not only awfully ugly, it is also much harder to reduce in postprocessing).
-
So all those concepts are subjectives, each case being a case in itself and maybe we should also consider the use of the 2 letters DR.
Just one question then, how do you define the "real DR" that differs from the "lab DR" of DxO?
If we are able to see 2, 4 or 6 stops of DR gap along the "real DR" axis, it must have a definition and be a measurable entity, so how is it defined?
Thanks.
Cheers,
Bernard
-
Just one question then, how do you define the "real DR" that differs from the "lab DR" of DxO?
If we are able to see 2, 4 or 6 stops of DR gap along the "real DR" axis, it must have a definition and be a measurable entity, so how is it defined?
Thanks.
Cheers,
Bernard
That's exactly what I think is the problem.
How do we stritcly have to use a term.
I do not have the answer to your question, but it seems that DR embrasse something too large and therefore confuse.
Remember that horsepowers have also the same issue: there are 2 measurements.
The Data horsepower and the real horsepower. They are not the same.
This has been used in car engineering to define 2 concepts that where shocking together.
That would be a great idea if we could invent a similar terminology.
Cheers.
Ps: somethinmg similar also exists in audio I think. I'm not an expert at all but I remember having heard about real power and data power.
-
Another discussion, http://www.imx.nl/photo/leica/camera/page176/s2part4.html (http://www.imx.nl/photo/leica/camera/page176/s2part4.html) really shows a difference between the D3X and Sony Alpha 900. I don't understand where that difference is coming from, either, but it's very much visible. There has been a lot of negative remarks on Erwin's testing, suggesting he does good lens tests and bad camera tests. Erwin may even test for DR in a future article.
And none of the negative remarks I have seen have been accompanied with actual criticism on his methodology, and consist of unsupported dismissals, hand-waving and thinly veiled ad hominems.
-
Just one question then, how do you define the "real DR" that differs from the "lab DR" of DxO?
If we are able to see 2, 4 or 6 stops of DR gap along the "real DR" axis, it must have a definition and be a measurable entity, so how is it defined?
Thanks.
Cheers,
Bernard
DxO does have the requisite information if you know where to look:
http://luminous-landscape.com/forum/index....c=42158&hl= (http://luminous-landscape.com/forum/index.php?showtopic=42158&hl=)
It is a bit exasperating to continually hear this complaint about DxO. The points about noise character are well taken however. They can be quantified by a more refined analysis of the noise spectrum, but DxO does not do this (at least, they don't present any results on their website).
-
Emil,
What is your point on my assumption that a significant advantage (like 4-6 stop) in DR for MFDBs would also be seen as good high ISO performance?
The ideas that exposing to the right we would have highlight near saturation so the extensive DR would be on the shadow side, essentially allowing for underexposure which is essentialy the same as fake ISOs.
I started this discussion because I would like to see any feasible explanation for having several steps of advantage of MFDBs over DSLRs, or is it the emperors new cloths?
Best regards
Erik
DxO does have the requisite information if you know where to look:
http://luminous-landscape.com/forum/index....c=42158&hl= (http://luminous-landscape.com/forum/index.php?showtopic=42158&hl=)
It is a bit exasperating to continually hear this complaint about DxO. The points about noise character are well taken however. They can be quantified by a more refined analysis of the noise spectrum, but DxO does not do this (at least, they don't present any results on their website).
-
DxO does have the requisite information if you know where to look:
http://luminous-landscape.com/forum/index....c=42158&hl= (http://luminous-landscape.com/forum/index.php?showtopic=42158&hl=)
It is a bit exasperating to continually hear this complaint about DxO. The points about noise character are well taken however. They can be quantified by a more refined analysis of the noise spectrum, but DxO does not do this (at least, they don't present any results on their website).
The problem is you are arguing about what amounts to religion with science, which is bound to fail
-
It is my understanding that if we correctly expose right we would essentially have non specular highlights near saturation. In this case DR would show up as latitude for underexposure quite similar to increasing ISO. As a matter of fact, on MFDBs not having variable pre amplifiers underexposure would work identical to high ISO settings.
Erik,
This is my understanding too.
Jonathan Wienke took the trouble a few years ago to produce a DR chart which defined the concept in terms of the legibility of text, in order to cut through the subjective confusion.
Is there anyone who would claim that a camera which could not provide legible text at a specific underexposure, has a higher DR than another camera which can produce legible text in the same circumstances, using lenses of comparable quality?
The chart consisted of progressivley smaller text and numbers on different colored backgrounds.
The procedure, when comparing the DR of two cameras, was to start with a full ETTR exposure of the chart from each camera, then progressively underexpose shots of the chart with both cameras.
For example, if camera A shots which were underexposed 12 stops still revealed legible text of a certain size on the chart, which camera B shots could reveal as legible at no greater underexposure than 10 stops, then camera A could reasonably be considered to have a 2 stop DR advantage, provided that the same text at -12 & 1/3rd EV (using Camera A) and -10 & 1/3rd EV (using camera B ), were both illegible.
This method also allowed DR testing at either the pixel level or the 'equal FoV' level. To compare DR at the pixel level, the distance to the chart would be adjusted so that the shots from each camera would consist of the same number of pixels.
However, in my view, the DR at equal FoV is more significant. The camera with the higher pixel count may have a DR advantage which should not be ignored.
DXO Mark shows the D3X as having a slightly higher DR than the P65+ at the pixel level, but a slightly lower DR at equal print sizes.
It would be interesting to compare results using Jonathan Wienke's method.
-
Feppe,
Emil is scientist and I am an engineer ;-)
Best regards
Erik
The problem is you are arguing about what amounts to religion with science, which is bound to fail
-
But then, comes immediatly a question.
If the D3x is that impressive, and in many aspects supposely matches or equal the highest MF backs,
considering its much lower price, its better distribution and universality, and that can even shoot video,
why the MF users are not selling their gears like hot-dogs and go all for the D3x?
(not talking about the one who use regularly view and tech cameras but MF cameras)
-
Hi,
The idea with this discussion is not compare D3X to P65+ or any other combination, but to find out if there is a reasonable explanation to the perceived 4-6 stop benefit in DR of MFDBs over DSLRs.
Best regards
Erik
But then, comes immediatly a question.
If the D3x is that impressive, and in many aspects supposely matches or equal the highest MF backs,
considering its much lower price, its better distribution and universality, and that can even shoot video,
why the MF users are not selling their gears like hot-dogs and go all for the D3x?
(not talking about the one who use regularly view and tech cameras but MF cameras)
-
Is there anyone who would claim that a camera which could not provide legible text at a specific underexposure, has a higher DR than another camera which can produce legible text in the same circumstances, using lenses of comparable quality?
It would be interesting to compare results using Jonathan Wienke's method.
IMHO, the above has little to do with DR. In engineering literature DR is normally understood to be a "point property", as opposed to a "neighbor hood property". For e.g., color is a point property - you have a well defined notion of the color at a given pixel. Texture is a neighborhood property - can you have texture at a single pixel? DR has generally been understood in terms of the dynamic range of a given pixel (point property). Attempts to extrapolate the meaning of DR from point property to neighborhood property have typically resulted in confusion. Some cases may still have merit, for example the (neighborhood property) notion of "image DR/SNR", and more importantly, "image scale dependent DR/SNR". However, the problem is that such notions are not constructed properly, with flawed assumptions, such as that erroneous notion that SNR for a natural image varies as square root of pixel summed/averaged, which adds further to the confusion.
Jonathan Weinke, or for that matter anybody else, is free to define a notion of "image quality", but relating that measure of IQ to DR, such as the above claim of "text-based DR" is not what has traditionally been understood in engineering circles.
Joofa
-
Removed.
-
Hi,
The idea with this discussion is not compare D3X to P65+ or any other combination, but to find out if there is a reasonable explanation to the perceived 4-6 stop benefit in DR of MFDBs over DSLRs.
Best regards
Erik
Correct, I was way off topic with this one.
Cheers.
-
Hi,
I don't argue with your view. On the other hand I'm not necessarily only interested in a technically accepted definition of DR but also perception of DR. Essentially in anything "for real" that can matter to a trained observer and that can be explained by theory or proven by samples.
Also, I think that Mark Dubovoy was talking about DR on details with full texture.
Best regards
Erik
IMHO, the above has little to do with DR. In engineering literature DR is normally understood to be a "point property", as opposed to a "neighbor hood property". For e.g., color is a point property - you have a well defined notion of the color at a given pixel. Texture is a neighborhood property - can you have texture at a single pixel? DR has generally been understood in terms of the dynamic range of a given pixel (point property). Attempts to extrapolate the meaning of DR from point property to neighborhood property have typically resulted in confusion. Some cases may still have merit, for example the (neighborhood property) notion of "image DR/SNR", and more importantly, "image scale dependent DR/SNR". However, the problem is that such notions are not constructed properly, with flawed assumptions, such as that erroneous notion that SNR for a natural image varies as square root of pixel summed/averaged, which adds further to the confusion.
Jonathan Weinke, or for that matter anybody else, is free to define a notion of "image quality", but relating that measure of IQ to DR, such as the above claim of "text-based DR" is not what has traditionally been understood in engineering circles.
Joofa
-
That's exactly what I think is the problem.
How do we stritcly have to use a term.
I do not have the answer to your question, but it seems that DR embrasse something too large and therefore confuse.
I do not think that "dynamic range" is a singular entity, and I'd compare it more to terms like "intelligence" for which there is no singular measure, or a word like "game" (thank you Prof. Wittgenstein) in which there is no singularly necessary quality (or jointly sufficient quality) that any thing has to have in order to be a 'game'. In actuality, it is practical characteristics of the data and its use that give us the impression of quality. Sometimes we cluster a number of our impressions under the term "dynamic range" though in the end, it is our ability to make practical use that serves our interests.
If a file is malleable, and able to sustain significant amount of adjustment without falling apart, that is a practical benefit. In the case of the D3x, the extreme low noise in the blacks gives the files an important practical benefit in all kinds of adjustments. This is quite different from the painstaking care that needs to be taken with, say, 5DII files, which show all kinds of strange noise in the shadows, which must be hidden or otherwise mitigated in practical use. If what the D3x has is "dynamic range" it's the kind I'm most glad to have.
-
Hi,
Just two comments:
1) Nikon D3X has much better DR than Canon 5DII according to DxO, this may be consistent with what you see
2) The DR measured by DxO doesn't say about the character of the noise just about the noise present
Best regards
Erik
I do not think that "dynamic range" is a singular entity, and I'd compare it more to terms like "intelligence" for which there is no singular measure, or a word like "game" (thank you Prof. Wittgenstein) in which there is no singularly necessary quality (or jointly sufficient quality) that any thing has to have in order to be a 'game'. In actuality, it is practical characteristics of the data and its use that give us the impression of quality. Sometimes we cluster a number of our impressions under the term "dynamic range" though in the end, it is our ability to make practical use that serves our interests.
If a file is malleable, and able to sustain significant amount of adjustment without falling apart, that is a practical benefit. In the case of the D3x, the extreme low noise in the blacks gives the files an important practical benefit in all kinds of adjustments. This is quite different from the painstaking care that needs to be taken with, say, 5DII files, which show all kinds of strange noise in the shadows, which must be hidden or otherwise mitigated in practical use. If what the D3x has is "dynamic range" it's the kind I'm most glad to have.
-
DxO does have the requisite information if you know where to look:
http://luminous-landscape.com/forum/index....c=42158&hl= (http://luminous-landscape.com/forum/index.php?showtopic=42158&hl=)
It is a bit exasperating to continually hear this complaint about DxO. The points about noise character are well taken however. They can be quantified by a more refined analysis of the noise spectrum, but DxO does not do this (at least, they don't present any results on their website).
As Emil has indicated in earlier posts, the engineering definition of DR places an upper limit on the DR defined by other criteria. The read noise and full well capacity (which determines shot noise) can be incorporated into a noise model that describes fairly well the observed noise characteristics of a camera, as Roger Clark has demonstrated on his web site.
The KODAK KAF-50100 is used in the Hasselblad H3DII-50, and its read noise is 12.5 electrons at the specified read out rate and the full well capacity is 40,300 electrons. The Nikon D3 (for which I have data) has a full well of 65568 electrons and a read noise at the base ISO of 17.63 electrons. One can plug these values into a spreadsheet using Clark's model and get the shown values. One can read the DR from the table as the exposure in f/stops from saturation for a given S:N. For a S:N of 1, the DR of the KAF-50100 is about 11.75 stops, which is in excellent agreement with the DXO measurement of 11.35 stops. The DR of the D3 at S:N = 1 is 12 stops as compared to the DXO value of 11.92 stops. The red values show where DR is limited by read noise.
[attachment=23050:SNR_DR.gif]
[attachment=23051:DXO_DR.GIF]
This is, of course, per pixel DR. The KAF-50100 has many more pixels with a smaller full well than that of the D3. The KAF-50100 collects 2.6 x more electrons than the D3, as can be determined by multiplying the full well times the pixel count. With the Hasselblad, one could print a larger picture at a given print resolution as compared to the Nikon, which after all is one of the main reasons for a MFDB over a dSLR. Alternatively, one could downsize to the pixel count of the Hasselblad to that of the Nikon and gain dynamic range by pixel binning. This is the rationale of the DXO print DR.
[attachment=23052:DXO_DR_print.gif
Although the D3 was a breakthrough for Nikon, it's DR at base ISO is not as great as it could be, since the read noise at low ISO is relatively high, but improves with higher ISO. As Emil has explained, this accounts for the flattening of the DR curve at low ISOs and is due to limitations in the electronics downstream to the sensor. This limitation is not present for the Hasselblad (or the Nikon D3x).
This discussion does not take pattern noise or the noise spectrum into account, but I would venture to hypothesize that a 4-6 stop improvement in DR of a MFDB as compared to the Nikon D3x or other top dSLRs is unlikely. The experts must be observing something other than DR when comparing the two formats. It would be interesting for them to demonstrate their claims with data or actual pictures. Until then, I remain skeptical.
-
As Emil has indicated in earlier posts, the engineering definition of DR places an upper limit on the DR defined by other criteria. .... This is, of course, per pixel DR. .... Alternatively, one could downsize to the pixel count of the Hasselblad to that of the Nikon and gain dynamic range by pixel binning. This is the rationale of the DXO print DR.
Do you know how DXO is calculating print DR?
An issue with the "downsizing to the pixel count" problem is that, as I mentioned before, many authors, and it would appear Emil Martinec included (from the link that you cite in your message), have taken the notion of DR as in the case of a flat field and tried to extrapolate that on a natural image, thinking that the "scaling of noise" is valid in that case also, which it is not IMHO. Because, they appear to calculate the effect of such downsizing on noise statistics only, and not realizing that the signal is also getting affected (blurred), and both S and N in the ratio S/N are varying as opposed to just N as they seem to assume.
A more extensive model of DR/SNR variation in downsampling is needed and it is not too difficult to work out such models.
Joofa
-
Hello Bill,
So you also say that the two sensors are about one stop apart, and this is what would be expected. Also that this is in agreement with DxO data. I'm not really surprised.
Best regards
Erik
As Emil has indicated in earlier posts, the engineering definition of DR places an upper limit on the DR defined by other criteria. The read noise and full well capacity (which determines shot noise) can be incorporated into a noise model that describes fairly well the observed noise characteristics of a camera, as Roger Clark has demonstrated on his web site.
The KODAK KAF-50100 is used in the Hasselblad H3DII-50, and its read noise is 12.5 electrons at the specified read out rate and the full well capacity is 40,300 electrons. The Nikon D3 (for which I have data) has a full well of 65568 electrons and a read noise at the base ISO of 17.63 electrons. One can plug these values into a spreadsheet using Clark's model and get the shown values. One can read the DR from the table as the exposure in f/stops from saturation for a given S:N. For a S:N of 1, the DR of the KAF-50100 is about 11.75 stops, which is in excellent agreement with the DXO measurement of 11.35 stops. The DR of the D3 at S:N = 1 is 12 stops as compared to the DXO value of 11.92 stops. The red values show where DR is limited by read noise.
[attachment=23050:SNR_DR.gif]
[attachment=23051:DXO_DR.GIF]
This is, of course, per pixel DR. The KAF-50100 has many more pixels with a smaller full well than that of the D3. The KAF-50100 collects 2.6 x more electrons than the D3, as can be determined by multiplying the full well times the pixel count. With the Hasselblad, one could print a larger picture at a given print resolution as compared to the Nikon, which after all is one of the main reasons for a MFDB over a dSLR. Alternatively, one could downsize to the pixel count of the Hasselblad to that of the Nikon and gain dynamic range by pixel binning. This is the rationale of the DXO print DR.
[attachment=23052:DXO_DR_print.gif
Although the D3 was a breakthrough for Nikon, it's DR at base ISO is not as great as it could be, since the read noise at low ISO is relatively high, but improves with higher ISO. As Emil has explained, this accounts for the flattening of the DR curve at low ISOs and is due to limitations in the electronics downstream to the sensor. This limitation is not present for the Hasselblad (or the Nikon D3x).
This discussion does not take pattern noise or the noise spectrum into account, but I would venture to hypothesize that a 4-6 stop improvement in DR of a MFDB as compared to the Nikon D3x or other top dSLRs is unlikely. The experts must be observing something other than DR when comparing the two formats. It would be interesting for them to demonstrate their claims with data or actual pictures. Until then, I remain skeptical.
-
That's exactly what I think is the problem.
How do we stritcly have to use a term.
I do not have the answer to your question, but it seems that DR embrasse something too large and therefore confuse.
Remember that horsepowers have also the same issue: there are 2 measurements.
The Data horsepower and the real horsepower. They are not the same.
This has been used in car engineering to define 2 concepts that where shocking together.
Fred,
Hum... I am afraid that won't cut it.
If Mark tells us he sees 6 stops gap in DR what we have is a clearly measurable entity, otherwise why write "6"? So there must be a clear definition.
Is "6" a metaphor for "a lot"? Fine, but then how does "a lot" translate in practical terms? What is the characteristic of these files showing "a lot" more DR?
Cheers,
Bernard
-
DxO does have the requisite information if you know where to look:
http://luminous-landscape.com/forum/index....c=42158&hl= (http://luminous-landscape.com/forum/index.php?showtopic=42158&hl=)
It is a bit exasperating to continually hear this complaint about DxO. The points about noise character are well taken however. They can be quantified by a more refined analysis of the noise spectrum, but DxO does not do this (at least, they don't present any results on their website).
Emil,
I am not questioning DxO. I am just trying to understand what some of the folks here mean when they say that the DxO figure doesn't reflect the actual real world DR they see in their files.
Until now the only concrete aspect that has been mentioned is the look of noise. So in essence does "the 6 stops more DR" just point to the fact that the 5DII has banding issues in deep shadows?
Cheers,
Bernard
-
But then, comes immediatly a question.
If the D3x is that impressive, and in many aspects supposely matches or equal the highest MF backs,
considering its much lower price, its better distribution and universality, and that can even shoot video,
why the MF users are not selling their gears like hot-dogs and go all for the D3x?
(not talking about the one who use regularly view and tech cameras but MF cameras)
Off topic indeed, but so much fun.
If the backs are so much better, why are like new second hand 39 megapixel backs selling at the same price point as the D3x?
More seriously, there are many valid reasons why people might want to keep using backs regardless of their DR:
- ADDED: they have signficantly more resolution (I had stupidely omited the obvious )
- They have the platform and lenses and like them,
- They like the look of AA filter less files,
- They are used to subtle DoF effects related to the large size of the sensor,
- Backs can be mounted on a variety of plarforms (including LF cameras),
- The sensor is easier to clean,
- Backs feel/look more profesional,
- It creates a differentiator relative to other photographers unable to afford a back,
- They like thethered shooting for which backs have been optimized,
- They have a personnal relationship with their back dealer that they would be unable to find with Canon/Nikon,
- They are able to meet Phaseone R&D guys and to influence their directions but wouldn't be able to do the same with Canon/Nikon,
- ...
That is what I find most irritating about this whole DR story. Backs don't need to oversell their DR using mythical and un-provable qualities, they have enough going for themselves already.
Cheers,
Bernard
-
Do you know how DXO is calculating print DR?
http://www.dxomark.com/index.php/eng/Our-publications/DxOMark-Insights/Detailed-computation-of-DxOMark-Sensor-normalization (http://www.dxomark.com/index.php/eng/Our-publications/DxOMark-Insights/Detailed-computation-of-DxOMark-Sensor-normalization)
Cheers,
Bart
-
Do you know how DXO is calculating print DR?
http://www.dxomark.com/index.php/eng/Our-publications/DxOMark-Insights/Detailed-computation-of-DxOMark-Sensor-normalization (http://www.dxomark.com/index.php/eng/Our-publications/DxOMark-Insights/Detailed-computation-of-DxOMark-Sensor-normalization)
Cheers,
Bart
Hi Bart,
Thanks for the link. IMHO, it appears that DXO is making the same mistake that I have pointed out in my earlier message, i.e., just to consider noise statistics in image size reduction. For a flat field it will work. However, for a natural, real, image, that is not correct as I see it, since the signal is also affected in the reduction operation and as I said before both S and N in S/N ratio are changing. Please note that SNR will typically improve in downsizing, however, the noise only statistics only sets the upper bound on that number as it corresponds to a flat field. For a natural image, one can't ignore the change in signal in addition to noise in such SNR calculations.
Joofa
-
Anyone seriously interested in pursuing this topic should first read and understand the excellent post by ejmartin Dynamic Range and DxO (http://http://luminous-landscape.com/forum/index.php?showtopic=42158&hl=). Continuing from there ...
Noise at extremely low signal levels is likely to be dominated by sensor electronics readout.
[blockquote]Suppose you are interested in "Engineering DR", defined by (maximum signal) / (signal equivalent to dark noise). If your camera electronics has pixel readout noise of 40 electrons and each pixel has full well (saturation) capacity of 40000 electrons, then your DR is (40000/40) = 1000 = 60db = 10 f/stops. The camera manufacturer can improve this DR for you either by lowering electronics noise or by increasing pixel full well capacity. [/blockquote]
Noise at moderate signal levels is likely to be dominated by photon statistics.
[blockquote]Suppose you are interested in "Photographer DR", defined by S/N>16 in the darkest pixels of the image which still contain acceptable detail. Then you require those dark pixels to contain at least 256 electrons. If the pixels have maximum full well capacity of 40000 electrons, then your DR is (40000/256) = 156 = 44db = 7.3 f/stops. The camera manufacturer can improve this DR for you by increasing the pixel full well capacity. Increasing the pixel count will also improve DR, assuming you are interested in "print DR" rather than "pixel DR". [/blockquote]
On a slight tangent to this topic, DR can also be improved by merging multiple exposures, at least for stationary subjects and a tripod mounted camera. I think some clever camera manufacturer should provide a one-button press function which captures and merges two exposures taken with different sensor integration times. This could easily beat the DR of any existing camera.
-
Emil,
What is your point on my assumption that a significant advantage (like 4-6 stop) in DR for MFDBs would also be seen as good high ISO performance?
Yes you have that correct, and it's a very good point to make.
-
IMHO, the above has little to do with DR. In engineering literature DR is normally understood to be a "point property", as opposed to a "neighbor hood property". For e.g., color is a point property - you have a well defined notion of the color at a given pixel. Texture is a neighborhood property - can you have texture at a single pixel? DR has generally been understood in terms of the dynamic range of a given pixel (point property). Attempts to extrapolate the meaning of DR from point property to neighborhood property have typically resulted in confusion. Some cases may still have merit, for example the (neighborhood property) notion of "image DR/SNR", and more importantly, "image scale dependent DR/SNR". However, the problem is that such notions are not constructed properly, with flawed assumptions, such as that erroneous notion that SNR for a natural image varies as square root of pixel summed/averaged, which adds further to the confusion.
Jonathan Weinke, or for that matter anybody else, is free to define a notion of "image quality", but relating that measure of IQ to DR, such as the above claim of "text-based DR" is not what has traditionally been understood in engineering circles.
Joofa
On the contrary, it has a lot to do with DR. Engineering DR is quite close to the range of tonal values over which pixel SNR exceeds unity, which condition is closely tied to the ability to extract details at the pixel scale.
-
Emil,
I am not questioning DxO. I am just trying to understand what some of the folks here mean when they say that the DxO figure doesn't reflect the actual real world DR they see in their files.
Until now the only concrete aspect that has been mentioned is the look of noise. So in essence does "the 6 stops more DR" just point to the fact that the 5DII has banding issues in deep shadows?
Cheers,
Bernard
I suspect that may be some of it. Pattern noise is a big problem with low ISO shadows in Canon DSLR's, much less so with current Nikons, especially the D3x. I have also recently discovered that there is a strong interaction among color separation in the CFA, sensor noise characteristics, and color space conversion. dcraw uses a simple matrix profile for converted demosaiced raw color to the output color space. Here is the Red channel of a 5D2 file, ISO 100, from Imaging-Resource -- before and after color space conversion to sRGB, using dcraw with my own demosaic algorithm:
(http://theory.uchicago.edu/~ejm/pix/20d/posts/dpr/5D_Rchannel_cam-vs-sRGB.gif)
Canon has recently adopted a strategy of making their R channel color filter more or less Yellow in sRGB terms. The loss of saturation in R requires a boost in color saturation during color space conversion, which exacerbates noise. I suspect that commercial converters, to the extent that they don't exhibit this problem, are doing some chroma NR under the hood. Certainly ACR has less pattern noise in R than the above example when output to sRGB.
Edit: People might also be interested in a comparison of the Red channel of this image for different DSLR's:
(http://theory.uchicago.edu/~ejm/pix/20d/posts/dpr/RedBanding_5D2-1Ds3-D3x.gif)
I chose the R channel because it will show the greatest effect -- typical tungsten white balance will amplify the R channel by a stop or more, and reveal noise more than G or B. There is also the camera-dependent effect of sRGB conversion, as mentioned above
-
On the contrary, it has a lot to do with DR. Engineering DR is quite close to the range of tonal values over which pixel SNR exceeds unity, which condition is closely tied to the ability to extract details at the pixel scale.
I don't think so. First of all we have to be follow the engineering tradition on how DR has been defined. I don't think so that I have come across the notion of a "text-based-DR" exactly in the terms as described here. Such notions, of course, do exist, but as I said in the realm of image quality (IQ). IQ and DR are not necessarily the same and we should not mix them.
Secondly, you can take a "high-text-DR" image and just slightly blur the boundaries of the text leaving everything the same. The text will be difficult to read now. Has the DR changed now? If it has, does it mean that the so called "image DR" is only defined on the edges of the text? What about the rest of the image? Does it not contribute to the image DR? Unfortunately, IMHO, if you tread this path there will be no shortage of such questions to answer.
The notion of DR is pretty clear in engineering circles and I don't think that such "text-based-DR" is how it has been traditionally treated in the image sensor community.
Joofa
-
Hi,
Resolution alone my by a differentiator.The flexibility you mention is a great plus. I also suspect that many MF lenses are pretty good. That said, my experience with my Pentax 67 is that the lenses have some shortcomings, like quite intensive chromatic aberrations. The Pentax 67 has produced a few stunning images (using scanned slide film), but if you pixel peep the image files quality falls apart. It's my understanding that Hasselblad switched to self designed lenses built by Fuji Film (?) because the new lenses have better performance.
In some cases it's important to have professional equipment like other pros. For Landscape shooters that probably does not matter. My guess is resolution counts a lot.
If we go back to your favorite topic (sorry for the pun), the Nikon D3X is considered to be to expensive. So there is a tendency that P65+ get compared with Canon 1DsIII and Canon 5DII.
Now, according to DxO at least, the Nikon D3X has a very real advantage in DR over Canon 5DII. I don't know how Nikon achieves that performance. Lloyd Chambers has gone into quite intensive comparisons between Canon 5DII and Nikon D3X and finds Nikon vastly superior on detail in darks. There is also a tendency that folks have Canon stuff. Than they go to Phase when they build a new system. It's not a bad strategy, if you can afford it. Most Phase lenses (And Hasselblad HC-lenses) seem to be impressive and not that expensive. Especially Canon seems to have some struggle in the short end.
One area that I think gets to little attention is flare and the camera itself. In a well designed optical system all light not contributing to the image should be absorbed. This is an are where MF equipment may actually excel, this however has nothing to do with sensor design. Lens and system flare is very hard to measure.
Best regards
Erik
Off topic indeed, but so much fun.
If the backs are so much better, why are like new second hand 39 megapixel backs selling at the same price point as the D3x?
More seriously, there are many valid reasons why people might want to keep using backs regardless of their DR:
- They have the platform and lenses and like them,
- They like the look of AA filter less files,
- They are used to subtle DoF effects related to the large size of the sensor,
- Backs can be mounted on a variety of plarforms (including LF cameras),
- The sensor is easier to clean,
- Backs feel/look more profesional,
- It creates a differentiator relative to other photographers unable to afford a back,
- They like thethered shooting for which backs have been optimized,
- They have a personnal relationship with their back dealer that they would be unable to find with Canon/Nikon,
- They are able to meet Phaseone R&D guys and to influence their directions but wouldn't be able to do the same with Canon/Nikon,
- ...
That is what I find most irritating about this whole DR story. Backs don't need to oversell their DR using mythical and un-provable qualities, they have enough going for themselves already.
Cheers,
Bernard
-
Emil,
I see your point, I think. Actually I think that DxO also discusses this, or a similar issue. The samples you show are "red channel" on sensor and red channel on an sRGB image generated using all channels?
http://www.dxomark.com/index.php/en/Our-pu...vs.-Nikon-D5000 (http://www.dxomark.com/index.php/en/Our-publications/DxOMark-Insights/Canon-500D-T1i-vs.-Nikon-D5000)
I presume that it's possible to aim for high ISO by having CFAs having broad spectral transmission characteristics and fix the colors in post?
Best regards
Erik
I suspect that may be some of it. Pattern noise is a big problem with low ISO shadows in Canon DSLR's, much less so with current Nikons, especially the D3x. I have also recently discovered that there is a strong interaction among color separation in the CFA, sensor noise characteristics, and color space conversion. dcraw uses a simple matrix profile for converted demosaiced raw color to the output color space. Here is the Red channel of a 5D2 file, ISO 100, from Imaging-Resource -- before and after color space conversion to sRGB, using dcraw with my own demosaic algorithm:
(http://theory.uchicago.edu/~ejm/pix/20d/posts/dpr/5D_Rchannel_cam-vs-sRGB.gif)
Canon has recently adopted a strategy of making their R channel color filter more or less Yellow in sRGB terms. The loss of saturation in R requires a boost in color saturation during color space conversion, which exacerbates noise. I suspect that commercial converters, to the extent that they don't exhibit this problem, are doing some chroma NR under the hood. Certainly ACR has less pattern noise in R than the above example when output to sRGB.
-
Off topic indeed, but so much fun.
If the backs are so much better, why are like new second hand 39 megapixel backs selling at the same price point as the D3x?
More seriously, there are many valid reasons why people might want to keep using backs regardless of their DR:
- They have the platform and lenses and like them,
- They like the look of AA filter less files,
- They are used to subtle DoF effects related to the large size of the sensor,
- Backs can be mounted on a variety of plarforms (including LF cameras),
- The sensor is easier to clean,
- Backs feel/look more profesional,
- It creates a differentiator relative to other photographers unable to afford a back,
- They like thethered shooting for which backs have been optimized,
- They have a personnal relationship with their back dealer that they would be unable to find with Canon/Nikon,
- They are able to meet Phaseone R&D guys and to influence their directions but wouldn't be able to do the same with Canon/Nikon,
- ...
That is what I find most irritating about this whole DR story. Backs don't need to oversell their DR using mythical and un-provable qualities, they have enough going for themselves already.
Cheers,
Bernard
Bernard,
I agree 100% with your post here.
-
Hi,
Summing it up this far...
We had some very knowledgeable persons chiming in.
- It is my understanding that there seems to be some consensus that we cannot see any reasons for MFDBs having more than maximum one stop advantage in DR over DSLRs.
- Posters seem to regard DxO data as relevant, at least if we dig a bit deeper in the data.
- It seems also quite obvious that Nikon D3X has pushed the envelope in DSLRs and Canon have some problem with pattern noise.
- Emil Martinec confirms my initial suggestion that extended DR in MFDBs should also result in good high ISO performance.
- Emil also looks at the behavior of the CGA (Color Grid Array), Canon seems to have a CGA design on recent designs that does not map smoothly on sRGB primaries.
- There is some discussion on resolving textures within the concept of DR but no consensus.
Bernard points out that there are plenty of reasons for using MFDBs over DSLR aside from the purported advantage in DR. A very good point.
Thanks for good input and a civilized discussion. Looking forward to the continued discussion.
Hi,
There are a lot of discussion about DR (Dynamic Range) on this forum. It is quite obvious that persons with considerable experience have quite different opinion on the issue.
- One thing that I suppose we all agree on is that bigger is better as long as Depth Of Field is not an issue. A bigger sensor collects more photons and therefore will have better photon statistics. If we assume an MFDB sensor having the double area of a FX (Full Frame) sensor the theoretical advantage would be about a half stop.
- It is my understanding that normally the noise floor is dominated by shot noise (statistical variation of incident photons) and in the darks possibly by read noise. It is also my understanding that read noise is less on DSLRs than on MFDBs. The number of electrons each pixel can hold is about the same on sensors having the same pitch
So according to the above we would never see an advantage in excess of say one stop maximum on MFDBs over full frame DSLRs. DxO labs publishes detailed measurements that pretty much are consistent with the above observations.
It is my understanding that if we correctly expose right we would essentially have non specular highlights near saturation. In this case DR would show up as latitude for underexposure quite similar to increasing ISO. As a matter of fact, on MFDBs not having variable pre amplifiers underexposure would work identical to high ISO settings.
We would expect that if a camera A would have a 4 stop advantage in DR over another camera B it would perform identical to camera B at 16 times the ISO. So would an MFDB have a 4 stop advantage it would achieve the same image quality at 1600 ISO as the lesser camera at 100 ISO. But it seems that this is clearly not the case. MFDBs don't perform very well at high ISOs (except the PXX+ series).
Now, many experienced observers clearly see a 4-6 stop advantage with MFDBs over DSLRs. I don't have any issue with that, but I cannot understand where it is coming from.
It would be very nice if some could come up with a physically feasible explanation for MFDBs having a significant advantage in DR or samples clearly demonstrating the effect.
Best regards
Erik Kaffehr
-
I don't think so. First of all we have to be follow the engineering tradition on how DR has been defined. I don't think so that I have come across the notion of a "text-based-DR" exactly in the terms as described here. Such notions, of course, do exist, but as I said in the realm of image quality (IQ). IQ and DR are not necessarily the same and we should not mix them.
Secondly, you can take a "high-text-DR" image and just slightly blur the boundaries of the text leaving everything the same. The text will be difficult to read now. Has the DR changed now? If it has, does it mean that the so called "image DR" is only defined on the edges of the text? What about the rest of the image? Does it not contribute to the image DR? Unfortunately, IMHO, if you tread this path there will be no shortage of such questions to answer.
The notion of DR is pretty clear in engineering circles and I don't think that such "text-based-DR" is how it has been traditionally treated in the image sensor community.
Joofa
Of course not Joofa. Image quality also has a lot to do with lens quality. The text based concept of DR is a practical, photographic method of comparing the DR capabilities of different cameras by examining detail in the deepest shadows. It's not a method of placing a precise figure on the camera's DR, in terms of dB or range of f/stops.
As a photographer, my interest in the DR of a camera relates to its ability to produce detail in the shadows. The legibility of text in a severely underexposed image, especially a colored text on a background of a different color, seems to me a good indication of the quality of detail.
If camera A boasts a higher DR than camera B, in terms of a certain engineering definition, but camera B produces clearer or more legible text when images from both cameras are compared at 10 stops underexposure (at base ISO) using lenses of equal quality, then camera B is the camera I want, as regards DR capability.
-
Emil,
I see your point, I think. Actually I think that DxO also discusses this, or a similar issue. The samples you show are "red channel" on sensor and red channel on an sRGB image generated using all channels?
http://www.dxomark.com/index.php/en/Our-pu...vs.-Nikon-D5000 (http://www.dxomark.com/index.php/en/Our-publications/DxOMark-Insights/Canon-500D-T1i-vs.-Nikon-D5000)
I presume that it's possible to aim for high ISO by having CFAs having broad spectral transmission characteristics and fix the colors in post?
Best regards
Erik
According to DXO, the red channel in the Canon 500D was more sensitive to green than red with D50 illumination, requiring a high matrix coefficient for that channel. The P65+ is only slightly better than the Canon in this area. And the D3x has more favorable characteristics for the red channel.
[attachment=23060:D65_Color.gif] [attachment=23061:D3xColor.gif]
-
Emil,
I see your point, I think. Actually I think that DxO also discusses this, or a similar issue. The samples you show are "red channel" on sensor and red channel on an sRGB image generated using all channels?
http://www.dxomark.com/index.php/en/Our-pu...vs.-Nikon-D5000 (http://www.dxomark.com/index.php/en/Our-publications/DxOMark-Insights/Canon-500D-T1i-vs.-Nikon-D5000)
Yes it's much the same point. The example shown is red channel, after demosaic but before color space conversion (dcraw lets you output that); and red channel, after both demosaic and color space conversion. If it were the red channel on the sensor, one would have only half the resolution of the other. DxO was making a general point about noise, in my example I was somewhat more interested in getting at why the 5D2 has so much more pattern noise at low ISO. Interestingly, if you look at color response at DxO, the P65+ also has a lot of sRGB green in the cameras R channel, while the D3x has much less; the coefficients in the color space transformation are much smaller for the D3x. Where the P65+ wins, however, is that its white balance coefficients are smaller; the camera is much more equally sensitive to sRGB R,G, and B than the D3x. So the R channel after color space conversion has to be amplified more in the D3x, again bringing more noise. This presumably comes at a cost in some transmissivity of the G and B filters for the Phase One back. It seems that there are some engineering tradeoffs to be considered.
I presume that it's possible to aim for high ISO by having CFAs having broad spectral transmission characteristics and fix the colors in post?
Best regards
Erik
When one does that, the demosaiced image has more recorded photons, so less luminance noise but little chromaticity, and one must boost the saturation in the color space transformation post-demosaic; so one has less luminance noise at the expense of poorer color, and more chroma noise. One can of course mitigate the chroma noise in post-processing, and/or in the converter as part of the color space conversion routine, but at the cost of some color accuracy and color microcontrast.
-
Interestingly, if you look at color response at DxO, the P65+ also has a lot of sRGB green in the cameras R channel, while the D3x has much less; the coefficients in the color space transformation are much smaller for the D3x. Where the P65+ wins, however, is that its white balance coefficients are smaller; the camera is much more equally sensitive to sRGB R,G, and B than the D3x. So the R channel after color space conversion has to be amplified more in the D3x, again bringing more noise. This presumably comes at a cost in some transmissivity of the G and B filters for the Phase One back. It seems that there are some engineering tradeoffs to be considered.
Emil,
An excellent point which I noted when posting my images from DXO, but did not fully comprehend.
That there are engineering conflicts in sensor design are pointed in this paper (http://spiedl.aip.org/getabs/servlet/GetabsServlet?prog=normal&id=JEIME5000018000002023002000001&idtype=cvips&gifs=yes&ref=no), which does a computer simulation for RGB and CMY sensor design. The authors suggest a RGB sensor having two red sensors for each green and blue sensor. While such a sensor would reduce noise on a statistical basis, one must wonder how it would affect perceived noise, taking into account the spectral sensitivities of human vision, where the eye is most sensitive to green. I understand that this is why the Bayer array has two green for each red and blue sensor.
Depending on the type of photography being performed, the most important illuminants are daylight and tungsten. With Illuminant A (tungsten), red is increased and there is better balance between the red and green channels, but blue is decreased and the white balance multiplier and blue matrix coefficient must be increased, aggravating noise in this channel.
Since the human perceptual system is least sensitive to blue, this would mitigate perceived noise in that channel.
-
What is always missed is these threads is the reality that how 'DR' plays out in a final printed image is the function of numerous variables that have nothing to do with the camera, but take the results out of the margin-of-error that might separate these machines. These factors include, roughly in order of importance:
1. Exposure
2. Raw conversion
3. Post-Pro
4. Monitor and printer calibration
5. Inkset
6. Media choice
Since the web and commercial mass repro have minuscule DR compared to fine art prints, I assume we are concerned only with the latter, because what the camera can capture at a theoretical level is of precisely zero use in the photographic process. To return to Ansel's metaphor, notes on a score which the instrument cannot play are hardly relevant to the quality of the performance.
Exposure: since no camera company has yet created a true "Expose to the Right" mode (and may mana rain upon him who does), there is really no way of knowing in real world work what is exactly to perfect exposure for any given scene. Indeed, since even tiny increases in exposure can have a dramatic effect on tonal differentiation in the image, many photographers couldn't actually agree on what level of exposure was 'just right' (ie: the point at which the ability to recover the slightest detail in the brightest highlight was lost). Moreover, in practical terms, some loss in the highlights may be perfectly acceptable (ie: cloud detail) for better shadows in a given, real-world scene.
The brings us to RAW conversion. Different RAW converters have varying methods and abilities of highlight recovery. Even positing the perfect user, LR, C1 and the maker's own software, will perform the alchemy of highlight recovery differently. Since highlights have so little 'colour' in them to our eyes, there is a lot of processing latitude to interpolate detail (and remember, *everything* off a bayer matrix is an interpolation*).
Post Pro: actual photographers don't shoot image to use them 'raw' from the camera, if such a thing even existed. We process them to suite our artistic taste or commercial expressive agenda. That manipulation of the image for tone, contrast, brightness, colour balance and saturation, etc, all impact on the date in the dark nether-regions -- specifically, the post-capture, post-conversion data.
The rest of the workflow: all of the calibration steps in the process will influence how the photographer sees the image, and thus how they choose to adjust it.....and thus how the tonal scale falls at the end of the day. Not that many people have truly excellent calibration control throughout their workflow (Mark happens to be one of them, btw).
Lastly, output matters. Since what we're talking about is making photographs (or so I assume), the user's media choice for comparison (and printer inkset) will have a significant impact on the way apparent DR is presented.
The upshot is this: real photographers can tell us what they see when they run images through their workflow from different and create final works of art. Theoreticians can tell us what numbers come out of machines at certain preliminary steps in that process under artificial conditions. As a photographer, the former interests me, the later interests me much less.
I have said here, and elsewhere, that working with a Phase back has given me a what I see to be a richer tonal range compared to 35mm dslrs. The post-moderns can explain to you why that sentence is of no scientific value. I can explain to you (elsewhere) why that meta-analysis is of no ultimate worth to the artist.
This explains my take of Irwin Puts work. I trust him implicitly on lens testing. That is a scientific process which he has mastered. And it tells us something. The same holds true, to a more limited extent, on things like his image-acutance conclusions (though publishing obviously motion-blurred or mis-focussed images has undercut his efforts of late in this regard).
But I have no idea if he aspires to be a photographic artist. I have not seen any real-world work of his that suggests he engages with the process for that purpose. That's not a criticism - I hope he gains satisfaction from what he does in his own way, because he has certainly enriched our understand greatly be his efforts. That said, unless and until he shows some body of works that demonstrates a real-world use of the tools to produce beautiful finished images, I can have no certainty that his methods of exposing, processing, viewing and printing are such that they can really tell me anything of much use about camera image quality.
When someone like Michael or James Russell or bcooter say something about characteristics of a camera's output, it is inherently subjective, but it matters because they know what they are talking about in a way that is more relevant to the end that photographers are pursuing - namely finished photographs.
- N.
ps. did I read here that Bernard got married to his D3s recently?
-
Nick,
The points you rise are absolutely valid.
I would like to make some points, however.
Erwin Puts did publish an image which probably was affected by camera vibration. As a matter of fact, Michael Reichmann was showing a image from D3X having similar problems. Now, in both cases the vibration was not relevant to the the test. I think Michael was discussing noise and Erwin Puts was discussing Longitudal Chromatic Aberration. If you test equipment unknown to you there is a possibility to make errors. Now, some vibrations are not relevant to axial chromatic aberration. Erwin may have pointed out the issue but I presume that he expects some IQ from his readers. Actually he expects a lot of IQ, and that is part of the problem with his writing.
I'm actually a bit upset about all this criticism. In my view folks try to share their experience and are doing a significant effort to achieve that. All they got in feedback is blame from folks that always know how to better but contribute little of their own. If you find an error please consider if it is relevant in context! To put things in perspective, I was actually considering renting a P65+ for testing, would cost like 3000 USD/day here in Sweden. Would I find a minor error, irrelevant to the the subject I try to investigate, would I pay another 3000 USD to do another test or just publish my findings? Please, get real!
I'd suggest that Erwin Puts may have explained the issue, but just to throw out information because of irrelevant objections is not going to help anyone!
The other issue is that the reason I started this discussion was that I wanted to find a scientifically feasible explanation to the purported differences we are supposed to see. You may consider sensors and raw converters to be just tools of the trade, but those tools are still to be developed, by scientists and engineers.
Best regards
Erik
What is always missed is these threads is the reality that how 'DR' plays out in a final printed image is the function of numerous variables that have nothing to do with the camera, but take the results out of the margin-of-error that might separate these machines. These factors include, roughly in order of importance:
1. Exposure
2. Raw conversion
3. Post-Pro
4. Monitor and printer calibration
5. Inkset
6. Media choice
Since the web and commercial mass repro have minuscule DR compared to fine art prints, I assume we are concerned only with the latter, because what the camera can capture at a theoretical level is of precisely zero use in the photographic process. To return to Ansel's metaphor, notes on a score which the instrument cannot play are hardly relevant to the quality of the performance.
Exposure: since no camera company has yet created a true "Expose to the Right" mode (and may mana rain upon him who does), there is really no way of knowing in real world work what is exactly to perfect exposure for any given scene. Indeed, since even tiny increases in exposure can have a dramatic effect on tonal differentiation in the image, many photographers couldn't actually agree on what level of exposure was 'just right' (ie: the point at which the ability to recover the slightest detail in the brightest highlight was lost). Moreover, in practical terms, some loss in the highlights may be perfectly acceptable (ie: cloud detail) for better shadows in a given, real-world scene.
The brings us to RAW conversion. Different RAW converters have varying methods and abilities of highlight recovery. Even positing the perfect user, LR, C1 and the maker's own software, will perform the alchemy of highlight recovery differently. Since highlights have so little 'colour' in them to our eyes, there is a lot of processing latitude to interpolate detail (and remember, *everything* off a bayer matrix is an interpolation*).
Post Pro: actual photographers don't shoot image to use them 'raw' from the camera, if such a thing even existed. We process them to suite our artistic taste or commercial expressive agenda. That manipulation of the image for tone, contrast, brightness, colour balance and saturation, etc, all impact on the date in the dark nether-regions -- specifically, the post-capture, post-conversion data.
The rest of the workflow: all of the calibration steps in the process will influence how the photographer sees the image, and thus how they choose to adjust it.....and thus how the tonal scale falls at the end of the day. Not that many people have truly excellent calibration control throughout their workflow (Mark happens to be one of them, btw).
Lastly, output matters. Since what we're talking about is making photographs (or so I assume), the user's media choice for comparison (and printer inkset) will have a significant impact on the way apparent DR is presented.
The upshot is this: real photographers can tell us what they see when they run images through their workflow from different and create final works of art. Theoreticians can tell us what numbers come out of machines at certain preliminary steps in that process under artificial conditions. As a photographer, the former interests me, the later interests me much less.
I have said here, and elsewhere, that working with a Phase back has given me a what I see to be a richer tonal range compared to 35mm dslrs. The post-moderns can explain to you why that sentence is of no scientific value. I can explain to you (elsewhere) why that meta-analysis is of no ultimate worth to the artist.
This explains my take of Irwin Puts work. I trust him implicitly on lens testing. That is a scientific process which he has mastered. And it tells us something. The same holds true, to a more limited extent, on things like his image-acutance conclusions (though publishing obviously motion-blurred or mis-focussed images has undercut his efforts of late in this regard).
But I have no idea if he aspires to be a photographic artist. I have not seen any real-world work of his that suggests he engages with the process for that purpose. That's not a criticism - I hope he gains satisfaction from what he does in his own way, because he has certainly enriched our understand greatly be his efforts. That said, unless and until he shows some body of works that demonstrates a real-world use of the tools to produce beautiful finished images, I can have no certainty that his methods of exposing, processing, viewing and printing are such that they can really tell me anything of much use about camera image quality.
When someone like Michael or James Russell or bcooter say something about characteristics of a camera's output, it is inherently subjective, but it matters because they know what they are talking about in a way that is more relevant to the end that photographers are pursuing - namely finished photographs.
- N.
ps. did I read here that Bernard got married to his D3s recently?
-
Since the web and commercial mass repro have minuscule DR compared to fine art prints, I assume we are concerned only with the latter, because what the camera can capture at a theoretical level is of precisely zero use in the photographic process.
Hi N.,
Your assumption is wrong, but you are not alone. The capture/input DR is recorded before gamma adjustment and other tonemapping. This means that shadow noise at the capture stage is amplified by postprocessing, and better S/N ratios in the shadows will translate to a higher overall perceived DR, also in the output.
The upshot is this: real photographers can tell us what they see when they run images through their workflow from different and create final works of art. Theoreticians can tell us what numbers come out of machines at certain preliminary steps in that process under artificial conditions. As a photographer, the former interests me, the later interests me much less.
As a photographer, I can tell you you're missing the point of this thread. There is no use in discussing the output quality in this thread unless one understands what to improve on the input side, and how that affects the output.
Cheers,
Bart
-
Since the web and commercial mass repro have minuscule DR compared to fine art prints,
When you say "the web" you're not talking about images viewed on computer screens are you? No way do fine art prints have more DR than computer screens. And while printing press has less DR than the best fiber-gloss inkjet prints, the typical magazine glossy still has better DR than fine art matte papers, so I'd hardly say it's minuscule.
Certainly, a fine art print is the gold standard in ultimate overall image quality; but to suggest it's the best way to judge the DR that a camera is capable of capturing seems misguided to me, since pretty much all DSLR's can capture more DR than an inkjet can reproduce.
-
Hi,
Jeff has several good points on this in my opinion. Prints have a DR of about 7 stops. My claim is that DSLRs have significantly more, like 12 stops. To fit a digital image to print we need to make some adjustment on tonality. A straight conversion would make for a very flat print.
LCD screens have significantly more DR than prints, perhaps as much as 9 stops.
Best regards
Erik
When you say "the web" you're not talking about images viewed on computer screens are you? No way do fine art prints have more DR than computer screens. And while printing press has less DR than the best fiber-gloss inkjet prints, the typical magazine glossy still has better DR than fine art matte papers, so I'd hardly say it's minuscule.
Certainly, a fine art print is the gold standard in ultimate overall image quality; but to suggest it's the best way to judge the DR that a camera is capable of capturing seems misguided to me, since pretty much all DSLR's can capture more DR than an inkjet can reproduce.
-
When you say "the web" you're not talking about images viewed on computer screens are you? No way do fine art prints have more DR than computer screens. And while printing press has less DR than the best fiber-gloss inkjet prints, the typical magazine glossy still has better DR than fine art matte papers, so I'd hardly say it's minuscule.
Certainly, a fine art print is the gold standard in ultimate overall image quality; but to suggest it's the best way to judge the DR that a camera is capable of capturing seems misguided to me, since pretty much all DSLR's can capture more DR than an inkjet can reproduce.
Jeff, my point is that no one visual artist actually views photos on their own computer screen as an end product, so it's kind of meaningless. I agree that most cameras capture more DR than inkjet can reproduce.
Bart, we are at cross-purposes. What I am saying is that the getting the optimal exposure in any scene, and then converting is to a usable imagine is a process so fraught with subjective variability (in real world applications) that tests/debates of this nature which purport to be scientific are of little of no practical use and, that the inherently subjective judgement of really good visual artists who have mastered the medium is equally or more meaningful. The quest for scientific certainty of comparison in this realm is amusing at best, misleadingly useless at worst.
- N.
-
Nick,
I disagree. Science and technology has taken photography to where it is today. It may be argued that films, sensors and so are just tools, but I would argue that the technology we have is more like the paint than the brush.
Another point is that we can buy a really good tool for creating pictures for perhaps 1000 USD. Somewhat improved tools are available for perhaps 3000 USD, like the Canon 5DII and the Sony Alpha 850/900. The MFDB discussion is about backs that costs up to 30 kUSD. We need to ask what you get for ten times the price. Some folks can afford a 30 kUSD back, some can't. For many of us we can afford it if we really want/need. In my view it is important that we have as much and as correct information available so users can spend their money wisely.
Best regards
Erik
Jeff, my point is that no one visual artist actually views photos on their own computer screen as an end product, so it's kind of meaningless. I agree that most cameras capture more DR than inkjet can reproduce.
Bart, we are at cross-purposes. What I am saying is that the getting the optimal exposure in any scene, and then converting is to a usable imagine is a process so fraught with subjective variability (in real world applications) that tests/debates of this nature which purport to be scientific are of little of no practical use and, that the inherently subjective judgement of really good visual artists who have mastered the medium is equally or more meaningful. The quest for scientific certainty of comparison in this realm is amusing at best, misleadingly useless at worst.
- N.
-
What is always missed is these threads is the reality that how 'DR' plays out in a final printed image is the function of numerous variables that have nothing to do with the camera, but ...
One variable that does relate to the camera is the camera DR. It takes quite an artist to produce a print that has more DR than was captured. On the other hand, a good photographic artist can map a high dynamic capture to something that is pleasing on a print.
Since the web and commercial mass repro have minuscule DR compared to fine art prints, I assume we are concerned only with the latter, because what the camera can capture at a theoretical level is of precisely zero use in the photographic process. To return to Ansel's metaphor, notes on a score which the instrument cannot play are hardly relevant to the quality of the performance.
As pointed out by others, the DR of an sRGB device (as with the web) is greater than that of a print. The best prints have a DR of 275:1 or about 8 stops (see Karl Lang (http://www.adobe.com/digitalimag/ps_pro_primers.html)--Rendering the print: the Art of Photography). An 8 bit sRGB space has values of 1..255 which represent values of 0.222:255 when referred back to scene radiance, corresponding to 10 stops. As Karl explains, the art of photography is to render a high radiance image to what appears good on the print. Your Adams reference is not apt. If you have no score (i.e no image with a range of radiances), you have nothing to print. What the camera can capture is of paramount importance, and the job of the artistic photographer is to map the radiance of the capture to what can be shown in a print.
The effective DR of an 8 bit sRGB color image is limited by banding, not noise. According to Greg Ward (http://www.anyhere.com/gward/hdrenc/hdr_encodings.html), an 8 bit sRGB image has a practical DR of 1:0.025, 40:1 or about 5.3 stops. This does not represent the ratio of the brightest to the darkest part of the image in the usual sense of dynamic range, but to the limited granularity of the encoding that causes noticeable differences in color as one goes from one level to the next in the scale. Banding can be diminished by noise, which dithers the image. This introduces a paradox: high DR requires low noise, but noise can dither banding and increase the useful DR.
Exposure: since no camera company has yet created a true "Expose to the Right" mode (and may mana rain upon him who does), there is really no way of knowing in real world work what is exactly to perfect exposure for any given scene. Indeed, since even tiny increases in exposure can have a dramatic effect on tonal differentiation in the image, many photographers couldn't actually agree on what level of exposure was 'just right' (ie: the point at which the ability to recover the slightest detail in the brightest highlight was lost). Moreover, in practical terms, some loss in the highlights may be perfectly acceptable (ie: cloud detail) for better shadows in a given, real-world scene.
If you have a working knowledge of the relationship of the camera luminance and RGB histograms to the raw data (this involves the headroom allowed for exposure latitude as well as the white balance multipliers), it is not difficult to obtain good ETTR exposure. Digital capture is linear, and a small difference in exposure will not make a dramatic effect on tonal differentiation as long as you don't blow channels. Highlight recovery can correct for minor overexposure and the exposure control in the raw converter can correct for slight underexposure. To sum up, I hope you are a better artist than a scientist.
-
Jeff, my point is that no one visual artist actually views photos on their own computer screen as an end product, so it's kind of meaningless.
I think points can be more effectively made without erroneous statements of fact, such as prints having more DR than screens. But even so, I disagree with your premise. The fact that a print is the final output does not make the DR in the original capture meaningless. If it did, why would it matter if a MFDB has more DR or not?
What I am saying is that the getting the optimal exposure in any scene, and then converting is to a usable imagine is a process so fraught with subjective variability (in real world applications) that tests/debates of this nature which purport to be scientific are of little of no practical use and, that the inherently subjective judgement of really good visual artists who have mastered the medium is equally or more meaningful. The quest for scientific certainty of comparison in this realm is amusing at best, misleadingly useless at worst.
So if I'm understanding you, the various steps that take place in going from camera capture to final print are the reason why "scientific" DR numbers don't agree with really good artists who have mastered the medium and claim that there's a huge difference in DR when there really isn't? I fail to see the logic in that. And that's my problem with the folks who dismiss the scientific DR measurements. They don't offer any explanation for why those tests are wrong or what they fail to consider. All we get is vague statements, but you can never pin them down and get concrete answers or examples. It comes across as "I know what I know, facts be damned".
If you want to say you get a better final print from MFDB, that's one thing. There could be various reasons for that, but a huge gap in DR is not one of them.
-
The capture/input DR is recorded before gamma adjustment and other tonemapping. This means that shadow noise at the capture stage is amplified by postprocessing, and better S/N ratios in the shadows will translate to a higher overall perceived DR, also in the output.............
There is no use in discussing the output quality in this thread unless one understands what to improve on the input side, and how that affects the output.
Cheers,
Bart
Well said, Bart. There's a lot of processing that takes place between a RAW capture and the final print. I remember a few years ago attempting linear conversions to retrieve maximum highlight detail from certain images, before ACR hit the market. I soon gave up on it. It was too difficult to get a satisfactory tonal result across the whole image when starting with a linear conversion.
Issues also raised in this thread about the DR limitations of the the monitor, the limitations of web-based jpegs, and particularly the DR limitations of the print, are complete red herrings.
When attempting to compare a particular aspect of camera performance, such as DR, it's standard practice that one should at least attempt to keep the processing similar, such as noise reduction, sharpening, saturation and vibrancy, hue and WB, and of course proof settings with regard to paper/printer profile, when making a print.
There are differences in the way certain RAW converters handle image files from different models of cameras. I know, for example, that Bibble can produce slightly sharper results from my Canon RAW images than ACR. But I know also that this slight edge in resolution is at the expense of noise. ACR images are a little softer, or blotchier, but have clearly less noise. If I apply some noise reduction to the Bibble conversion, it looks pretty close to the ACR conversion.
Nevertheless, there are subtle differences, and anyone may prefer one particular RAW converter to another based on personal taste, and even image type, such as portrait or landscape.
However, the issue in this thread is not about subtle differences in DR due to differences in the default noise reduction of certain RAW converters, but is about huge differences in DR as reported by certain reviewers.
What's the explanation?
There are a number of explanations that I think may all be true to some degree. I'll list a few, but please don't think I'm pointing the finger at anyone. We're all human, but some of us are more scientifically rigorous than others. And even the scientifically rigorous are not neccessarily scientifically rigorous all the time, in all circumstances. Even Einstein made some flaws of judgement, perhaps due to his religious proclivities. (I'm thinking of, "God does not play dice", in relation to the theory of Quantum Mechanics).
1. MFDB manufacturers are struggling to be viable. Their products are ridiculously expensive in relation to the increased performance over a good 35mm DSLR, and they need all the good publicity they can get. A bit of hyperbole from a few reputable photographers is much appreciated.
2. Image is important in more ways than one. The professional photographer needs to impress his clients. An expensive MFDB system may do the trick.
3. Less discerning photographers tend to follow the 'big boys'. If their successful mentors or idols are using particular equipment, they will tend to follow and buy the same equipment. It's all too human. There's a lot of irrational behaviour in human society. When their purchasing decision has been made, mortgaging their house to buy a complete MFDB system, any criticism of their excessive expenditure will be met with strong, subjective statements supporting the superior performance of their equipment.
There can be no objective testing in such circumstances, because such testing would reveal such a marginal increase in performance, disproportionate with the price paid, it would cause distress. We all like to kid ourselves on occasions, including me.
4. There's a tendency to group the performance of all 35mm DSLR together, and compare an unspecified experience of 35mm performance with MFDB, which may exclude the best 35mm performer, the Nikon D3X, but include the best MFDB performer, whatever that may be.
Should I continue? I don't want the thread to be closed.
I'll add that I've never found much difficulty in determing the DR capability of a camera. Everyone who can afford a good camera probably lives in a house or a flat, doesn't he/she? I mean, he/she who owns a P65+ or D3X is not likely to be a street dweller.
In which case, just photograph your living room on a sunny day and expose for the brightest clouds out of your window. It's terribly simple. If you want to be really thorough, you could place a few very detailed objects and artifacts in the living room, even a newspaper.
Having exposed correctly to get all the cloud detail, using your MFDB and D3X in ETTR mode, then examine the detail in your living room. The camera that provides the greater detail in your living room has the better DR.
The fact that a true ETTR may be difficult to achieve, is another red herring. Just do it. If the clouds are blown, take another shot, and another shot, till it's right.
-
But then, comes immediatly a question.
If the D3x is that impressive, and in many aspects supposely matches or equal the highest MF backs,
considering its much lower price, its better distribution and universality, and that can even shoot video,
why the MF users are not selling their gears like hot-dogs and go all for the D3x?
(not talking about the one who use regularly view and tech cameras but MF cameras)
Well, the June issue of Nihon Camera has a couple of full-page landscape images from the D3x and the Pentax. And on seeing them, I prefer the D3x image. Both are Jpeg images. Which is a bit surprising since I own both a Phamiya and a D3x, and the Phamiya outresolves the D3x considerably.
Edmund
-
Ray,
There is also a healthy advantage in resolution. It matters if you make large prints...
Best regards
Erik
Well said, Bart. There's a lot of processing that takes place between a RAW capture and the final print. I remember a few years ago attempting linear conversions to retrieve maximum highlight detail from certain images, before ACR hit the market. I soon gave up on it. It was too difficult to get a satisfactory tonal result across the whole image when starting with a linear conversion.
Issues also raised in this thread about the DR limitations of the the monitor, the limitations of web-based jpegs, and particularly the DR limitations of the print, are complete red herrings.
When attempting to compare a particular aspect of camera performance, such as DR, it's standard practice that one should at least attempt to keep the processing similar, such as noise reduction, sharpening, saturation and vibrancy, hue and WB, and of course proof settings with regard to paper/printer profile, when making a print.
There are differences in the way certain RAW converters handle image files from different models of cameras. I know, for example, that Bibble can produce slightly sharper results from my Canon RAW images than ACR. But I know also that this slight edge in resolution is at the expense of noise. ACR images are a little softer, or blotchier, but have clearly less noise. If I apply some noise reduction to the Bibble conversion, it looks pretty close to the ACR conversion.
Nevertheless, there are subtle differences, and anyone may prefer one particular RAW converter to another based on personal taste, and even image type, such as portrait or landscape.
However, the issue in this thread is not about subtle differences in DR due to differences in the default noise reduction of certain RAW converters, but is about huge differences in DR as reported by certain reviewers.
What's the explanation?
There are a number of explanations that I think may all be true to some degree. I'll list a few, but please don't think I'm pointing the finger at anyone. We're all human, but some of us are more scientifically rigorous than others. And even the scientifically rigorous are not neccessarily scientifically rigorous all the time, in all circumstances. Even Einstein made some flaws of judgement, perhaps due to his religious proclivities. (I'm thinking of, "God does not play dice", in relation to the theory of Quantum Mechanics).
1. MFDB manufacturers are struggling to be viable. Their products are ridiculously expensive in relation to the increased performance over a good 35mm DSLR, and they need all the good publicity they can get. A bit of hyperbole from a few reputable photographers is much appreciated.
2. Image is important in more ways than one. The professional photographer needs to impress his clients. An expensive MFDB system may do the trick.
3. Less discerning photographers tend to follow the 'big boys'. If their successful mentors or idols are using particular equipment, they will tend to follow and buy the same equipment. It's all too human. There's a lot of irrational behaviour in human society. When their purchasing decision has been made, mortgaging their house to buy a complete MFDB system, any criticism of their excessive expenditure will be met with strong, subjective statements supporting the superior performance of their equipment.
There can be no objective testing in such circumstances, because such testing would reveal such a marginal increase in performance, disproportionate with the price paid, it would cause distress. We all like to kid ourselves on occasions, including me.
4. There's a tendency to group the performance of all 35mm DSLR together, and compare an unspecified experience of 35mm performance with MFDB, which may exclude the best 35mm performer, the Nikon D3X, but include the best MFDB performer, whatever that may be.
Should I continue? I don't want the thread to be closed.
I'll add that I've never found much difficulty in determing the DR capability of a camera. Everyone who can afford a good camera probably lives in a house or a flat, doesn't he/she? I mean, he/she who owns a P65+ or D3X is not likely to be a street dweller.
In which case, just photograph your living room on a sunny day and expose for the brightest clouds out of your window. It's terribly simple. If you want to be really thorough, you could place a few very detailed objects and artifacts in the living room, even a newspaper.
Having exposed correctly to get all the cloud detail, using your MFDB and D3X in ETTR mode, then examine the detail in your living room. The camera that provides the greater detail in your living room has the better DR.
The fact that a true ETTR may be difficult to achieve, is another red herring. Just do it. If the clouds are blown, take another shot, and another shot, till it's right.
-
Actually, personally I don't give much of a monkey's for DR. All I really worry about is blown highlights, and that is a matter for correct exposure. Working in B/W, most times the last thing I want is lots of shadow detail. I usually like my shadow areas to be black, or nearly so, so I often a clip a lot of DR out of the print quite deliberately.
But one advantage of my ridiculously expensive MF digital back, which I don't see mentioned here, is the abilty to make really severe crops (or selectives, as we used to call them). I can chop just a section out of the frame and still have 20MP, quite enough for an A3 print or more. Which is nice to have.
John
-
Ray,
There is also a healthy advantage in resolution. It matters if you make large prints...
The increased resolution of MFDBs is a distinct advantage for large prints, but one can gain similar resolution by stitching with a dSLR as demonstrated by Bernard with his exquisite photos of Japanese mountainous sites. Stitching can only be used for relatively static objects and is a pain, but so would be backpacking MFDB gear into the mountains. Photography involves a number of judiciously chosen compromises. Contrary to the suggestion of an injudicious earlier post, Bernard has chosen what is best for his purposes and is not married to his D3x .
As shown by Ctein, Roger Clark and others, the resolution of the even MFDBs is insufficient for really large high resolution prints and stitching, a large format scanning back or large format film may be required.
Regards,
Bill
-
Actually, personally I don't give much of a monkey's for DR. All I really worry about is blown highlights, and that is a matter for correct exposure. Working in B/W, most times the last thing I want is lots of shadow detail. I usually like my shadow areas to be black, or nearly so, so I often a clip a lot of DR out of the print quite deliberately.
But one advantage of my ridiculously expensive MF digital back, which I don't see mentioned here, is the abilty to make really severe crops (or selectives, as we used to call them). I can chop just a section out of the frame and still have 20MP, quite enough for an A3 print or more. Which is nice to have.
John
John,
I could have written exactly the same lines. I have same experience with the DR and totally agree.
Generally I often clip a lot of DR also.
-
Hi,
The reason that this discussion was started was to find out if there is a physically feasible explanation for the purported 4-6 stop advantages in DR with MFDBs over DSLRs.
Regarding the importance of DR it's a different issue and your mileage may vary.
Best regards
Erik
Actually, personally I don't give much of a monkey's for DR. All I really worry about is blown highlights, and that is a matter for correct exposure. Working in B/W, most times the last thing I want is lots of shadow detail. I usually like my shadow areas to be black, or nearly so, so I often a clip a lot of DR out of the print quite deliberately.
But one advantage of my ridiculously expensive MF digital back, which I don't see mentioned here, is the abilty to make really severe crops (or selectives, as we used to call them). I can chop just a section out of the frame and still have 20MP, quite enough for an A3 print or more. Which is nice to have.
John
-
But one advantage of my ridiculously expensive MF digital back, which I don't see mentioned here, is the abilty to make really severe crops (or selectives, as we used to call them). I can chop just a section out of the frame and still have 20MP, quite enough for an A3 print or more. Which is nice to have.
Probably it wasn't mentioned because the original intent of this thread was not to debate advantages and disadvantages of MFDB versus DSLR. I don't think anybody here is claiming the MFDB's have no advantages at all.
The question under debate is whether there's a huge gap in real world, usable dynamic range, and if so what explains that gap and why does it disagree with the engineering measurement of DR which shows the backs and DSLR's to be quite close in DR.
Whether it's Mark D's 6-stop comment, or the statements by Michael and some others that there's a 3-4 stop advantage, none of these claims have been backed up with any sort of compelling arguments based on facts, tests or even logic.
-
Probably it wasn't mentioned because the original intent of this thread was not to debate advantages and disadvantages of MFDB versus DSLR. I don't think anybody here is claiming the MFDB's have no advantages at all.
The question under debate is whether there's a huge gap in real world, usable dynamic range, and if so what explains that gap and why does it disagree with the engineering measurement of DR which shows the backs and DSLR's to be quite close in DR.
Yes Jeff and Eric
I was somewhat off-topic here, and my apologies. However, I know this debate is all terribly compelling and so on (and we seem to keep having it, in one form or another), but what I don't understand is why it seems to matter so much to all you chaps. I mean, what has it actually got to do with real-world photography? We have always had less DR in the negative and in the print than our eyes can see in the real world. A stop here or there between one film and another or one sensor and another may be really interesting to the anoraks of this world, but once you are out there climbing over hedges and struggling through the brambles, a bit of cloud cover to the north will fill the shadows and make more difference to your DR than the ruddy sensor ever will.
As photographers we spend our time working with and considering the quality of light. Sometimes we can control it, in a studio, sometimes we choose not to and grapple with the light nature bestows upon us. It seems to me that all the DSLRs and MF backs and films made today have plenty of DR for pictorial photography. Unless perhaps you always work in the middle of the Arizona desert at noon under a cloudless sky. A certain amount of dynamic compression is what makes a photograph look like a photograph - it is part of the style and visual language of of photography. Which is perhaps why these HDR images we see now look so horrible.
John
-
I was somewhat off-topic here, and my apologies. However, I know this debate is all terribly compelling and so on (and we seem to keep having it, in one form or another), but what I don't understand is why it seems to matter so much to all you chaps. I mean, what has it actually got to do with real-world photography? We have always had less DR in the negative and in the print than our eyes can see in the real world.
It matters a lot because it opens up possibilities we never had before. It matters alot because it allows a level of realism in my images I never had before. It matters a lot because I can shoot at ISO levels I never could before with such a level of quality. You cannot be serious in preferring the limited dynamic range of slide film, can you? So, removing a weaker link in the chain matters, a lot.
A stop here or there between one film and another or one sensor and another may be really interesting to the anoraks of this world, but once you are out there climbing over hedges and struggling through the brambles a bit of cloud cover to the north will fill the shadows and make more difference to your DR than the ruddy sensor ever will.
You forgot to mention the use of a properly dimensioned lens hood, and no filters, and superior optics, but that's not the issue at hand.
Again, you are missing the point of this thread, it is about understanding the wild claims of superior DR where there is demonstrably none, or very little at best in a few isolated cases when downsampled.
Unless perhaps you always work in the middle of the Arizona desert at noon under a cloudless sky. A certain amount of dynamic compression is what makes a photograph look like a photograph - it is part of the style and visual language of of photography. Which is perhaps why these HDR images we see now look so horrible.
That is also not true, it's the poor tonemapping that gives HDRI a bad rep. You also fail to see that the huge interest in HDRI apparently suggests something was lacking, namely DR. That's why the debates are here, most of us are seeking to improve our image quality.
Cheers,
Bart
-
That is also not true, it's the poor tonemapping that gives HDRI a bad rep. You also fail to see that the huge interest in HDRI apparently suggests something was lacking, namely DR. That's why the debates are here, most of us are seeking to improve our image quality.
Cheers,
Bart
In my own opinion (and it is only mine) increased DR and "improved image quality" are not at all the same thing. At least, if we are talking about pictorial image quality. What is it that you want your pictures to look like? Exactly like reality? Surely reality is the starting point for art, not the destination.
John
-
What is it that you want your pictures to look like?
Hi John,
I don't want them to look like anything. I want them to evoke the emotion I want, without the technical hurdles or obstacles making my life more difficult than need be. Unnecessary noise, banding, blocked shadows, and other technical imperfections usually distract from the message being delivered.
Now, back to the topic.
It is my impression that some of the perceived differences attributed to DR, are actually caused by differences in the MTF. That's why an objective DR determination cannot fully explain the claims. There may also be a difference in the contribution of veiling glare.
Cheers,
Bart
-
Hmmm,
Better resolved texture? I was not thinking about that! Veiling glare is a quite obvious culprit. Add to that the issue Emil martinec discussed on CGA and color transformation and things may start to add up.
Best regards
Erik
It is my impression that some of the perceived differences attributed to DR, are actually caused by differences in the MTF. That's why an objective DR determination cannot fully explain the claims. There may also be a difference in the contribution of veiling glare.
Cheers,
Bart
-
Ray,
There is also a healthy advantage in resolution. It matters if you make large prints...
Best regards
Erik
Of course, Erik. The higher resolution from a sensor with a higher pixel count is very important when viewing large prints from the same distance you would normally view a small print.
I think it's interesting that there's a very good parallel in the Canon 35mm world, to the DXO comparison between the D3X and the P65+. It's the comparison between the 20D and the 5D2. I don't think anyone would claim the 5D2 has a 4-6 stop DR advantage over the 20D.
In both situations we have a larger format camera with the same pixel pitch as another smaller format camera, and in both situations the two different formats vary in size (or at least in area) by the same degree. Shooting the same scene with the 20D and 5D2, using the same lens, then cropping the 5D2 shot to the same FoV as the 20D shot, produces not only the same size of image but the same quality of image, approximately. Tonal range, color sensitivity, SNR at 18% grey, are all about the same. What differs slightly is the DR. The 5D2 has a slight edge in DR, perhaps due to a further narrowing of the gap between microlenses, but so small (about 0.25 of a stop) one might not notice it.
Likewise, using the same lens with the D3X and P65+ (or lenses of equal focal length and optical quality, as far as possible), then cropping the P65+ shot to the same FoV as the D3X shot, should result in images of similar quality. Perhaps the P65+ shot will be very marginally crisper due to its lack of an AA filter. However, tonal range, color sensitivity and even SNR at 18% grey should be very similar.
The significant difference between these two sets of comparisons, according to DXO Mark, is the way dynamic range varies. The D3X image should have a whopping 1.75 stops greater DR than the cropped P65+ image. Clearly the D3X is in a league of its own. The D3X exceeds the DR of the P65+ (at the pixel level) by a greater degree than the P65+ exceeds the DR of the 5D2 (at the pixel level).
But to raise your question again; how do the 4 to 6 stop claims arise when only a 1 stop DR advantage should apply in relation to the average full frame 35mm DSLR, according to DXO? I think perhaps we should look at the way C1 handles the P65+ RAW image. Maybe there's some magical algorithm in the C1 software that works only with Phase backs, creating an illusory 4 to 6 stops DR advantage . Or maybe owners of MFDB systems are so pissed-off at the fact that their expensive cameras have such lousy performance at high ISO, that they just can't resist stirring the pot and having a dig at owners of cameras which do have superb high-ISO performance, by occasionally making outrageous statements about the illusory 4-6 stop advantage which no-one seems to be able to demonstrate .
Seriously, I believe it's the case that most photographers who use MFDBs are professionals who constantly need to impress their clients. It would be counter-productive and a waste of time for any of them to demonstrate that a D3X really is capable of delivering more DR than a P65, or that the DR advantage of a P65 over a Canon 35mm DSLR is only about 1 stop.
60mp is clearly much better than 21 or 24mp, and there's no doubt that on a 40"x60" print from a P65+, the veins on a model's eyeball are so much more impressive when viewed from a distance of a foot or so.
As Fred mentions, good solid blacks can be beautiful. If a professional really needs to photograph the interior of a room, simultaneously displaying the beautiful view through the window - flowers, green fields and fluffy white clouds all correctly exposed - he's probably in a position, as a professional, to bring in a truck-load of lighting, or even apply a film gel to the glass of the window to reduce the brightness of the outside scene. The result is likely to be more impressive than a single shot from a D3X, at least on a large print. Also, if dynamic range is likely to be an issue, then a 1 stop advantage is not likely to be sufficient to produce a good, professional result. Bracketing of exposures and merging to HDR may be a minimum requirement, in the absence of additional lighting, whatever camera is used.
-
Despite that I gave up when curves and graphics started to show up, I find this thread really good, informative and as an example of
the high level and behaviour expected in this site.
Erik, the OP did a brillant moderator task to avoid out-off-topics paths. And that was not an easy topic that could have degenerated.
Many technical informations as I said, but IMO one point has not yet been answered:
In what frame should we all use the term DR, in order to be sure we are taking the same reference point.
In other words, what should we strictly understand by the DR term, and what is accepted by everybody as a relaible mesurement standard ?
If we can not reach a common field on a standard, then, weired differences will show up all the time.
IMO.
-
A certain amount of dynamic compression is what makes a photograph look like a photograph - it is part of the style and visual language of of photography.
Very true.
I believe that the very essence of the quest for DR is in fact a quest away from the issues of digital files we have been facing for years:
- sensors are linear and don't allow any highlight recovery,
- the transition to blown areas isn't always as smooth as with film,
- shadows can exhibit weird issues like banding,
- some digital files require good manipulation discipline not to reveal their digital nature in print,
- ...
We sort of all believe/claim to be on the fore front of digital image manipulation but most of us in fact just hope to be able to get rid of these issues. DR is perceived as the one stop solve it all magic word that encompases all that.
Once the issues are solved, we are indeed back to square one where what really matters is sweet light hitting the right spot in a scene. More often than not we won't need much DR to tap into that magic moment.
Now there are of course other reasons why we keep debating:
- Some claims made by moral autorities in this community are surprising which triggers a genuine desire to see proof,
- Gear discussions are fun and can be an efficient way to stay awake in some boring meetings,
- ...
Cheers,
Bernard
-
Ray,
Thanks for good comments. I would almost expect you pointing out the viewing distnace issue. I would add that it's working the way that we don't increase viewing distance proportionally to image size. Think of an IMAX or Omnimax theatre. One of the reasons to print big is that you can see the print at reasonable distance and still feel immersed in the picture.
One point I would like to make that peoople like Mark Dubovoy probably don't need to impress customers, more get satisfaction of doing what they do optimally. I also think they are serious in their writing.
On the other hand, Mark was quite confirmative on the 6 stop difference. I cannot explain how such a large difference could arise, therefore the topic.
Another issue is that DxO-mark is widely discredited by those who know better. I was most interested what perspective the folks familiar with image processing take on that issue.
My understanding is that it seems to be a common ground regarding DxO beeing pretty correct and usable and also that DR advantage of MFDBs over DSLR would not exceed one stop.
I also made the assumption that extensive DR would translate into good high ISO capability, and at least Emil Martinec seems to agree.
Emil and Bart also came up with some new ideas on effects which may be interpreted as DR related to CGA implementation, veiling flare and also MTF. Very well possible that all this adds up to a significant advantage in image quality.
Best regards
Erik
Of course, Erik. The higher resolution from a sensor with a higher pixel count is very important when viewing large prints from the same distance you would normally view a small print.
I think it's interesting that there's a very good parallel in the Canon 35mm world, to the DXO comparison between the D3X and the P65+. It's the comparison between the 20D and the 5D2. I don't think anyone would claim the 5D2 has a 4-6 stop DR advantage over the 20D.
In both situations we have a larger format camera with the same pixel pitch as another smaller format camera, and in both situations the two different formats vary in size (or at least in area) by the same degree. Shooting the same scene with the 20D and 5D2, using the same lens, then cropping the 5D2 shot to the same FoV as the 20D shot, produces not only the same size of image but the same quality of image, approximately. Tonal range, color sensitivity, SNR at 18% grey, are all about the same. What differs slightly is the DR. The 5D2 has a slight edge in DR, perhaps due to a further narrowing of the gap between microlenses, but so small (about 0.25 of a stop) one might not notice it.
Likewise, using the same lens with the D3X and P65+ (or lenses of equal focal length and optical quality, as far as possible), then cropping the P65+ shot to the same FoV as the D3X shot, should result in images of similar quality. Perhaps the P65+ shot will be very marginally crisper due to its lack of an AA filter. However, tonal range, color sensitivity and even SNR at 18% grey should be very similar.
The significant difference between these two sets of comparisons, according to DXO Mark, is the way dynamic range varies. The D3X image should have a whopping 1.75 stops greater DR than the cropped P65+ image. Clearly the D3X is in a league of its own. The D3X exceeds the DR of the P65+ (at the pixel level) by a greater degree than the P65+ exceeds the DR of the 5D2 (at the pixel level).
But to raise your question again; how do the 4 to 6 stop claims arise when only a 1 stop DR advantage should apply in relation to the average full frame 35mm DSLR, according to DXO? I think perhaps we should look at the way C1 handles the P65+ RAW image. Maybe there's some magical algorithm in the C1 software that works only with Phase backs, creating an illusory 4 to 6 stops DR advantage . Or maybe owners of MFDB systems are so pissed-off at the fact that their expensive cameras have such lousy performance at high ISO, that they just can't resist stirring the pot and having a dig at owners of cameras which do have superb high-ISO performance, by occasionally making outrageous statements about the illusory 4-6 stop advantage which no-one seems to be able to demonstrate .
Seriously, I believe it's the case that most photographers who use MFDBs are professionals who constantly need to impress their clients. It would be counter-productive and a waste of time for any of them to demonstrate that a D3X really is capable of delivering more DR than a P65, or that the DR advantage of a P65 over a Canon 35mm DSLR is only about 1 stop.
60mp is clearly much better than 21 or 24mp, and there's no doubt that on a 40"x60" print from a P65+, the veins on a model's eyeball are so much more impressive when viewed from a distance of a foot or so.
As Fred mentions, good solid blacks can be beautiful. If a professional really needs to photograph the interior of a room, simultaneously displaying the beautiful view through the window - flowers, green fields and fluffy white clouds all correctly exposed - he's probably in a position, as a professional, to bring in a truck-load of lighting, or even apply a film gel to the glass of the window to reduce the brightness of the outside scene. The result is likely to be more impressive than a single shot from a D3X, at least on a large print. Also, if dynamic range is likely to be an issue, then a 1 stop advantage is not likely to be sufficient to produce a good, professional result. Bracketing of exposures and merging to HDR may be a minimum requirement, in the absence of additional lighting, whatever camera is used.
-
Hmmm,
Better resolved texture? I was not thinking about that! Veiling glare is a quite obvious culprit. Add to that the issue Emil martinec discussed on CGA and color transformation and things may start to add up.
Best regards
Erik
Finally, we are looking at parameters other than DR in an attempt to explain the differences between the two formats. The DR theory is the emperor who has no clothes. Veiling glare has to do with the lens and the baffling in the camera box and not the format per se. Emil's comments on CGA and color transformation are right on and explain the DXO noted differences between the Canon 500D and Nikon D5000, where the Nikon has superior spectral sensitivity to the Canon; both cameras have similar relative sensitivities.
However, in the comparison of the Nikon D3x and Phase One P65+, the Nikon has better spectral sensitivity in the red channel, but an unfavorable relative sensitivity, requiring a larger white balance multiplier. The net effect is more or less a draw and the two cameras have a similar color sensitivity. As Emil noted, these are engineering compromises.
[attachment=23095:ColorSensPrnt.gif]
IMHO, it is a mistake to refer to dSLRs as a homogeneous entity, as if there is no difference between the d3x and a Digital Rebel. Likewise, there are significant differences between MFDBs. It makes more sense to compare the best dSLR (D3x) to the best MFDB (P65+). The P65+ and D3x have pixel sizes of 6 um and 5.8 um respectively, and the D3x has slightly better per pixel performance, but the P65+ has more pixels. Four to six f/stops of DR is 16 and 64 times respectively, and such an advantage is unlikely from a mere 2.5 x in sensor area.
Regards,
Bill
-
I believe that the very essence of the quest for DR is in fact a quest away from the issues of digital files we have been facing for years [...]
It is very true indeed that the very different tonal characteristic of the media made asked for an athletic gymnastic during the transition ; however, I'll rephrase that in a slightly different way (according to my personal history at least).
I shot slides in the last millennium, having no practical access to a color darkroom, and when transitioning to a dSLR a few years ago, I really felt like relieved of a burden : passing from a minilab print made from a slide (5-6 stops?) to an inkjet print made from an edited raw file (8-9 stops? these numbers are an uneducated guess only), my images could really stand much closer to what I saw especially in the shadows, which is very practical in my boringly figurative style (or lack thereof).
So I'd say the current quest for DR may have been triggered by the leap forward made by digital files in the color area : it showed that huge progress could be made, and showed also there is still some room for improvement - I'd say that my eyes have a DR of about 15 stops.
Veiling glare is a quite obvious culprit.
Sensor DR is a desirable thing but I'd also think that veiling glare is the weakest link relative to DR in many real-world situations, especially with wide-angle lenses, making sensor DR irrelevant past a certain point (which I've sometimes seen quoted in the 10-12 stops range, ie exactly where the DR war actually starts).
There are more and more lens reviews online, but I'd love to find one that measures or at least shows some things in this area. A shoot with a controlled (studio) lighting in or near the frame and details to unshovel in the shadows should do the trick?
-
Hi,
On the first issue I'd agree that neither MTF nor veiling glare is are sensor related issues. Both may help to explain a visible difference between camera systems A and B in rendering texture in the dark parts of the image.
I fully agree that we cannot see either DSLRs or MFDBs as homogenious entities, even if some contributors seem to take that liberty. Mark Dubovoy compared P65+ and Canon 1DsII. It certainly seems to be the case that Nikon D3X stands out amongs DSLRs. The question is also in this case, how did they achieve that. On the other hand, the advantage the D3X has is consistent with DxO data.
Best regards
Erik
Finally, we are looking at parameters other than DR in an attempt to explain the differences between the two formats. The DR theory is the emperor who has no clothes. Veiling glare has to do with the lens and the baffling in the camera box and not the format per se. Emil's comments on CGA and color transformation are right on and explain the DXO noted differences between the Canon 500D and Nikon D5000, where the Nikon has superior spectral sensitivity to the Canon; both cameras have similar relative sensitivities.
...
IMHO, it is a mistake to refer to dSLRs as a homogeneous entity, as if there is no difference between the d3x and a Digital Rebel. Likewise, there are significant differences between MFDBs. It makes more sense to compare the best dSLR (D3x) to the best MFDB (P65+). The P65+ and D3x have pixel sizes of 6 um and 5.8 um respectively, and the D3x has slightly better per pixel performance, but the P65+ has more pixels. Four to six f/stops of DR is 16 and 64 times respectively, and such an advantage is unlikely from a mere 2.5 x in sensor area.
Regards,
Bill
-
I was somewhat off-topic here, and my apologies. However, I know this debate is all terribly compelling and so on (and we seem to keep having it, in one form or another), but what I don't understand is why it seems to matter so much to all you chaps. I mean, what has it actually got to do with real-world photography? We have always had less DR in the negative and in the print than our eyes can see in the real world. A stop here or there between one film and another or one sensor and another may be really interesting to the anoraks of this world, but once you are out there climbing over hedges and struggling through the brambles, a bit of cloud cover to the north will fill the shadows and make more difference to your DR than the ruddy sensor ever will.
As photographers we spend our time working with and considering the quality of light. Sometimes we can control it, in a studio, sometimes we choose not to and grapple with the light nature bestows upon us. It seems to me that all the DSLRs and MF backs and films made today have plenty of DR for pictorial photography. Unless perhaps you always work in the middle of the Arizona desert at noon under a cloudless sky. A certain amount of dynamic compression is what makes a photograph look like a photograph - it is part of the style and visual language of of photography. Which is perhaps why these HDR images we see now look so horrible.
Personally I think there's more to photography than just the soft, filtered light of the magic hour, but that's what most Velvia photographers limited themselves to because they didn't really have a choice with such a contrasty film. And even in those conditions, the scene contrast often required using grad filters, which IMHO look stupid when their use is apparent (which is more often than not). The 6-stop range of slide film was not a positive, it was something that photographers had to make do with because they wanted the other benefits of slides.
I think all kinds of light can be "good" light depending on the conditions and your subject. Having more DR to handle the contrast in more scenes is a good thing. You can always increase contrast in post if you want, but you can't bring back what was lost so given I choice I'd rather have more DR than less.
You won't get any argument from me about overdone HDR, but that really has nothing to do with camera DR, that's just a matter of bad processing (and maybe poor taste).
Bottom line is that with the D3x, I don't have to bracket multiple exposures as often as with my old cameras, which is definitely a benefit.
-
Four to six f/stops of DR is 16 and 64 times respectively, and such an advantage is unlikely from a mere 2.5 x in sensor area.
The following comment is not necessarily related to the "DR difference" between D3X and P65+ per se, but for more general edification. Resolution is a reason for perceived DR difference, as you rightly point out. However, there are other issues that can cause DR difference even on the same technology. For example an important one is the slew rate (http://en.wikipedia.org/wiki/Slew_rate): the clocking used to read pixels on certain sensors limits the amount of signal change in a read out period, and hence, DR is substantially reduced. There are a few other factors also under the hood but I shall avoid going into them at this point but they can have considerable impact on DR, since I just wanted to point out that resolution difference is not the be all and end all of DR.
Joofa
-
Hi,
In my view:
There is a standard definition of DR and that is (Maximum Signal)/(Signal at SNR=1), this essentially translates to (Well capacity) / (read noise) both measured in electrons normally, where SNR=1 mean signal equals noise. This is based on signal processing theory. Now, it can be argued the image quality at SNR=1 is not very useful photographically, but we need to keep in mind that this would be very dark and probably much compressed in print. A very good reason to keep this definition of DR is that it is the accepted one.
As said before, the lower threshold here will be noisy. We could use a higher criteria, like SNR=16 (signal is 16 times noise), in this region readout noise would not dominate and shot noise would be more important. Shot noise is randomness of light so it is equal for all sensors of same size and quantum efficiency. Quadrupling the size of the sensor would give a factor of two on SNR all other parameters assumed to be constant. This is exactly where the expected "maxium one stop advantage" comes into play. So whatever criteria we choose the ration between a large sensor (MFDB) and a smaller one (DSLR) will be about the same.
In short:
- DR at SNR=1 is the normal definition used in signal processing.
- Would we choose another SNR, it would matter little, because noise will essentially depend on photon statistics.
I'm not an expert on this, just trying to put it simply!
Best regards
Erik
In what frame should we all use the term DR, in order to be sure we are taking the same reference point.
In other words, what should we strictly understand by the DR term, and what is accepted by everybody as a relaible mesurement standard ?
If we can not reach a common field on a standard, then, weired differences will show up all the time.
IMO.
-
Hi,
This example may demonstrate the need for DR:
[attachment=23101:_DSC5560.jpg]
Full image: http://http://echophoto.dnsalias.net//ekr/...KarerSee_01.jpg (http://echophoto.dnsalias.net//ekr/images/DRExample/KarerSee_01.jpg)
The trees are in shadow and the mountainside has snow illuminated by bright sunlight. Initially I intended to use HDR for the image but found that I could extract about the information I needed from the image having least exposure. Admittedly, it could be done better, the forrest behind the lake is a bit flat.
The image was shot on APS-C, Sony Alpha 100, no camera that really excels in DR...
Best regards
Erik
Yes Jeff and Eric
I was somewhat off-topic here, and my apologies. However, I know this debate is all terribly compelling and so on (and we seem to keep having it, in one form or another), but what I don't understand is why it seems to matter so much to all you chaps. I mean, what has it actually got to do with real-world photography? We have always had less DR in the negative and in the print than our eyes can see in the real world. A stop here or there between one film and another or one sensor and another may be really interesting to the anoraks of this world, but once you are out there climbing over hedges and struggling through the brambles, a bit of cloud cover to the north will fill the shadows and make more difference to your DR than the ruddy sensor ever will.
As photographers we spend our time working with and considering the quality of light. Sometimes we can control it, in a studio, sometimes we choose not to and grapple with the light nature bestows upon us. It seems to me that all the DSLRs and MF backs and films made today have plenty of DR for pictorial photography. Unless perhaps you always work in the middle of the Arizona desert at noon under a cloudless sky. A certain amount of dynamic compression is what makes a photograph look like a photograph - it is part of the style and visual language of of photography. Which is perhaps why these HDR images we see now look so horrible.
John
-
DR at SNR=1 is the normal definition used in signal processing.
Yes, this is the right definition to use. However, in practise both the max signal and the lowest perceptible signal has a relation with extrinsic circuit elements such as voltages applied, clocking, etc.
Would we choose another SNR, it would matter little, because noise will essentially depend on photon statistics.
The notion of SNR is not properly defined/interpreted for a single image, IMHO. If the count associated with a certain pixel is 20,000, then what is the noise? Is it sqrt (20,000)? Of course, not. Because, if you knew it was sqrt (20,000), you can just use this fact to calculate the actual signal, with some heuristic on the sign of noise. The issue is that for a given image and given pixel location the count is just a number. From a single number you can't figure out what is the noise. The reason the statistics generated by say, Roger Clark, can get away is that while the actual "signal" component associated with a count of 20,000 is unknown, it is typically not very far from 20,000, and Poisson statistics can help us here to take 20,000 as the signal, where in actuality it was not. However, this approximation will not work at low signal levels as sqrt becomes increasingly appreciable protion of the signal.
Then what is the meaning of "image SNR"? Well, what happens many times in statistics is that you use area statistics to figure out a point statistics. SNR for a given pixel in a single image (point property) is not known, but lets expand that notion to the area around it and examine the pixel values and try to reason the SNR of a fixed "patch of image" given the values of pixel intensities in neighboring pixels. That shall work if the local neighborhood is known to be kind of uniform, so that even if you can't figure out the "global image SNR" you can have at least some notion on the local SNR - the abuse of such concepts being the so called "shadow SNR", "high-light SNR", and what not.
Quadrupling the size of the sensor would give a factor of two on SNR all other parameters assumed to be constant.
I think it is about time we got rid of that myth. Since, that is not applicable to natural images, IMHO.
Joofa
-
Hi,
I don't mind to stand corrected, but I tried to answer a question from "fredjeang", as well as could. Would it be possible to give a short answer, decently good enough, that can be understood by a photographer without a junior degree in statistics or signal processing?
Best regards
Erik
Yes, this is the right definition to use. However, in practise both the max signal and the lowest perceptible signal has a relation with extrinsic circuit elements such as voltages applied, clocking, etc.
The notion of SNR is not properly defined/interpreted for a single image, IMHO. If the count associated with a certain pixel is 20,000, then what is the noise? Is it sqrt (20,000)? Of course, not. Because, if you knew it was sqrt (20,000), you can just use this fact to calculate the actual signal, with some heuristic on the sign of noise. The issue is that for a given image and given pixel location the count is just a number. From a single number you can't figure out what is the noise. The reason the statistics generated by say, Roger Clark, can get away is that while the actual "signal" component associated with a count of 20,000 is unknown, it is typically not very far from 20,000, and Poisson statistics can help us here to take 20,000 as the signal, where in actuality it was not. However, this approximation will not work at low signal levels as sqrt becomes increasingly appreciable protion of the signal.
Then what is the meaning of "image SNR"? Well, what happens many times in statistics is that you use area statistics to figure out a point statistics. SNR for a given pixel in a single image (point property) is not known, but lets expand that notion to the area around it and examine the pixel values and try to reason the SNR of a fixed "patch of image" given the values of pixel intensities in neighboring pixels. That shall work if the local neighborhood is known to be kind of uniform, so that even if you can't figure out the "global image SNR" you can have at least some notion on the local SNR - the abuse of such concepts being the so called "shadow SNR", "high-light SNR", and what not.
I think it is about time we got rid of that myth. Since, that is not applicable to natural images, IMHO.
Joofa
-
Hi Erik,
I think I did provide the short answer and I repeat it below:
The notion of SNR is not properly defined/interpreted for a single image, IMHO.
I.e., it is difficult to put the concept of SNR for a single natural image using the definition you provided unless some extra constraints are put in.
Joofa
-
The notion of SNR is not properly defined/interpreted for a single image, IMHO. If the count associated with a certain pixel is 20,000, then what is the noise? Is it sqrt (20,000)? Of course, not. Because, if you knew it was sqrt (20,000), you can just use this fact to calculate the actual signal, with some heuristic on the sign of noise.
Joofa,
Your post makes no sense to me. You do not seem to realize that probability and statistics apply to populations, not individuals. If the frequency of cancer in a certain population is 1 in 20, then the probability that a certain individual has cancer is 5%. However the patient either has cancer or doesn't have cancer.
Regards,
Bill
-
Joofa,
Your post makes no sense to me. You do not seem to realize that probability and statistics apply to populations, not individuals. If the frequency of cancer in a certain population is 1 in 20, then the probability that a certain individual has cancer is 5%. However the patient either has cancer or doesn't have cancer.
Regards,
Bill
Bill,
First of all in this case there is a notion of an "instantaneous" value of noise, i.e., each pixel count has a certain unknown noise value, which we don't know, but we do know its associated long term statistics. Secondly, I think you have just said what was my point. That the notion of sqrt in Poisson statistics applies to ensembles (populations) and not to specific count in a single image at a given pixel location (individuals).
Haven't we seen people taking sqrt of the count associated with a given pixel in an image as the "noise", even on this forum, and not realizing that the concept of sqrt applies to ensembles and not single numbers- i.e., if I am given a large number of the images of the same scene with same lighting conditions etc., then for a given pixel location I have a sequence of numbers in the temporal domain and terms such as mean/std-dev/etc. have a proper meaning.
For a given pixel in a single image we just have a number: a count. Which is just the sampled valued from an underlying distribution, and from a single number we can't infer any stuff such as underlying mean, std-dev, etc. But people do it all the time. And, they get away with it because for large values of count the sqrt is increasingly smaller proportion of that value of the count. So if we have 20,000 count just take that as the mean value of the underlying Poisson distribution, where as in actuality it may not be 20,000, but close to it, so sqrt is not that off.
Sincerely,
Joofa
-
Bill,
First of all in this case there is a notion of an "instantaneous" value of noise, i.e., each pixel count has a certain unknown noise value, which we don't know, but we do know its associated long term statistics. Secondly, I think you have just said what was my point. That the notion of sqrt in Poisson statistics applies to ensembles (populations) and not to specific count in a single image at a given pixel location (individuals).
Haven't we seen people taking sqrt of the count associated with a given pixel in an image as the "noise", even on this forum, and not realizing that the concept of sqrt applies to ensembles and not single numbers- i.e., if I am given a large number of the images of the same scene with same lighting conditions etc., then for a given pixel location I have a sequence of numbers in the temporal domain and terms such as mean/std-dev/etc. have a proper meaning.
For a given pixel in a single image we just have a number: a count. Which is just the sampled valued from an underlying distribution, and from a single number we can't infer any stuff such as underlying mean, std-dev, etc. But people do it all the time. And, they get away with it because for large values of count the sqrt is increasingly smaller proportion of that value of the count. So if we have 20,000 count just take that as the mean value of the underlying Poisson distribution, where as in actuality it may not be 20,000, but close to it, so sqrt is not that off.
Sincerely,
Joofa
Joofa,
The pixel with a count of 20,000 subtends and represents an area in the subject, and the count represents the luminance of that area during the time of the exposure. The illumination of the subject is subject to random variations following a Poisson distribution, and the sensor count representing this area would vary with repeated exposures even though the subject and exposure parameters are held constant. Shot noise is in the light and exists even before the light hits the sensor. However, the count of 20,000 is a point estimate of the mean illuminance and the square root of 20,000 is an estimate of the standard deviation. You are correct that the derived values are only estimates of these parameters, but as the sample size increases, the accuracy of these estimate increases. I think that these differences are well understood by most observers and you are splitting hairs. Look up Standard Error (http://en.wikipedia.org/wiki/Standard_error_%28statistics%29) and Standard Error of the mean.
If the luminance represented by several adjacent pixels is the same (as with a clear blue sky or uniformly illuminated wall), the image pixels would have differing values and this would appear as noise.
Regards,
Bill
-
The pixel with a count of 20,000 subtends and represents an area in the subject, and the count represents the luminance of that area during the time of the exposure. The illumination of the subject is subject to random variations following a Poisson distribution, and the sensor count representing this area would vary with repeated exposures even though the subject and exposure parameters are held constant.
Hi Bill,
What random variations and what repeated exposures? There is a single number here for a given pixel associated with a single exposure resulting in a single image. If I provide you with a single image and if the first pixel on that image has 20,000 counts, and then I ask you what is the noise value on that pixel then what is the answer? Now if you follow the path where people just take sqrts, then you would be tempted to believe that sqrt (20,000) is the exact value of noise on that pixel, so you have two answers, 20,000 - sqrt (20,000) = 19859, or, 20,000 + sqrt (20000) = 20,141. You can pick either 19859 or 20,141 and say that is the signal. Lets say I arbitrarily consider that 19859 is the right answer and declare victory, and, now this is the pure signal that I have figured out. I have no noise, then why do I worry about SNR now?
Or, you can take the right approach, which is that sqrt just represents a bound on some measure of variation of the pixel value (std. dev.) as you have rightly pointed out resulting from multiple exposures. But here is the catch, I just gave you a single image (single exposure); how do I determine std. dev from just one sample? The fact is that a given pixel in a single image is just one sample, and I can't have a technically-legal notion of SNR with it, at least the usual meaning.
The third option is pragmatic. I know that I have one sample, so I can't calculate std. dev. However, I do know that for relatively high illumination, if I had taken several images of the same scene instead of one, then I would have found that the value of the first pixel in each of them is varying a little about 20,000, but is always close to 20,000. I can now take this sequence and determine its mean and std. dev. and find that the std. dev is close to the sqrt of 20,000. So I have solved my problem, I shall not worry about acquiring several images to determine this std. dev., as an approximation I shall just take the sqrt of a single number, 20,000, which is the pixel count for the first pixel of a single image, and take that as an approximate value of the actual std. dev for noise statistics. However, with this understanding that this number, sqrt (20,000), is not the approximate value of noise on the first pixel in the first image, but an average value of noise measure on that pixel if I had acquired a large number of images. So, I still don't know the exact value of noise on the first pixel in the first image, but I know that it will rarely exceed sqrt (20,000).
Still this approximation will only work for higher illumination and not be good for low light images.
Shot noise is in the light and exists even before the light hits the sensor. However, the count of 20,000 is a point estimate of the mean illuminance and the square root of 20,000 is an estimate of the standard deviation. You are correct that the derived values are only estimates of these parameters, but as the sample size increases, the accuracy of these estimate increases.
There is no sample size issue here. We have N=1 sample size. A single number 20,000, which is the count associated with the first pixel in the first image. And, since I don't have multiple images so I can't observe a sequence for N >1. Of course, there is an easy way out, instead of temporal direction, if I collect pixel counts spatially, then even in a single image I have a whole bunch of samples (millions). But here is the catch? Do all of these samples come from the same underlying Poisson distribution? The problem is that for a natural image the answer is no. So I can't take the std. dev. of all pixels in an image and declare that to be the std. dev. of the sequence represented by pixel counts associated with the first pixel of multiple images, if I had acquired more than one images for the same scene.
I think that these differences are well understood by most observers and you are splitting hairs. Look up Standard Error (http://en.wikipedia.org/wiki/Standard_error_%28statistics%29) and Standard Error of the mean.
What splitting hairs? Bill, there is only a single sample here, so the notion of accuracy increase, standard error are not even applicable!
If the luminance represented by several adjacent pixels is the same (as with a clear blue sky or uniformly illuminated wall), the image pixels would have differing values and this would appear as noise.
Now we are onto something. And I did point out this thing as expanding the notion of a point statistic to an area statistic in an earlier message. However, as I said before, this has only local utility, because for it to extend to the full image would mean that it is almost a flat field, and not a natural, real image we capture using our cameras - images of people, vegetation, cats, dogs, sky, water, anything ....
Sincerely,
Joofa
-
Hi,
I don't get your point.
Let's assume that we have a uniform surface evenly illuminated that is imaged on say 100x100 pixels, that is a 10000 pixels. These 10000 pixels are independent in exposure. So we have 10000 exposures with average value of 20000 and assumed Stddev of 141. So Signal is 20000 , noise 141 and SNR = 141. Let's reduce exposure to 200, no we have noise about 14 from photo statistic but we need also to take readout noise into account.
If we assume readout noise to be 10 electrons and add noise in quadrature we get a noise level of 17 electrons so SNR would be 200/17 that is about 12.
Simplicistic?
Best regards
Erik
Hi Bill,
What random variations and what repeated exposures? There is a single number here for a given pixel associated with a single exposure resulting in a single image. If I provide you with a single image and if the first pixel on that image has 20,000 counts, and then I ask you what is the noise value on that pixel then what is the answer? Now if you follow the path where people just take sqrts, then you would be tempted to believe that sqrt (20,000) is the exact value of noise on that pixel, so you have two answers, 20,000 - sqrt (20,000) = 19859, or, 20,000 + sqrt (20000) = 20,141. You can pick either 19859 or 20,141 and say that is the signal. Lets say I arbitrarily consider that 19859 is the right answer and declare victory, and, now this is the pure signal that I have figured out. I have no noise, then why do I worry about SNR now?
Or, you can take the right approach, which is that sqrt just represents a bound on some measure of variation of the pixel value (std. dev.) as you have rightly pointed out resulting from multiple exposures. But here is the catch, I just gave you a single image (single exposure); how do I determine std. dev from just one sample? The fact is that a given pixel in a single image is just one sample, and I can't have a technically-legal notion of SNR with it, at least the usual meaning.
The third option is pragmatic. I know that I have one sample, so I can't calculate std. dev. However, I do know that for relatively high illumination, if I had taken several images of the same scene instead of one, then I would have found that the value of the first pixel in each of them is varying a little about 20,000, but is always close to 20,000. I can now take this sequence and determine its mean and std. dev. and find that the std. dev is close to the sqrt of 20,000. So I have solved my problem, I shall not worry about acquiring several images to determine this std. dev., as an approximation I shall just take the sqrt of a single number, 20,000, which is the pixel count for the first pixel of a single image, and take that as an approximate value of the actual std. dev for noise statistics. However, with this understanding that this number, sqrt (20,000), is not the approximate value of noise on the first pixel in the first image, but an average value of noise measure on that pixel if I had acquired a large number of images. So, I still don't know the exact value of noise on the first pixel in the first image, but I know that it will rarely exceed sqrt (20,000).
Still this approximation will only work for higher illumination and not be good for low light images.
There is no sample size issue here. We have N=1 sample size. A single number 20,000, which is the count associated with the first pixel in the first image. And, since I don't have multiple images so I can't observe a sequence for N >1. Of course, there is an easy way out, instead of temporal direction, if I collect pixel counts spatially, then even in a single image I have a whole bunch of samples (millions). But here is the catch? Do all of these samples come from the same underlying Poisson distribution? The problem is that for a natural image the answer is no. So I can't take the std. dev. of all pixels in an image and declare that to be the std. dev. of the sequence represented by pixel counts associated with the first pixel of multiple images, if I had acquired more than one images for the same scene.
What splitting hairs? Bill, there is only a single sample here, so the notion of accuracy increase, standard error are not even applicable!
Now we are onto something. And I did point out this thing as expanding the notion of a point statistic to an area statistic in an earlier message. However, as I said before, this has only local utility, because for it to extend to the full image would mean that it is almost a flat field, and not a natural, real image we capture using our cameras - images of people, vegetation, cats, dogs, sky, water, anything ....
Sincerely,
Joofa
-
Personally I think there's more to photography than just the soft, filtered light of the magic hour, but that's what most Velvia photographers limited themselves to because they didn't really have a choice with such a contrasty film. And even in those conditions, the scene contrast often required using grad filters, which IMHO look stupid when their use is apparent (which is more often than not). The 6-stop range of slide film was not a positive, it was something that photographers had to make do with because they wanted the other benefits of slides.
I think all kinds of light can be "good" light depending on the conditions and your subject. Having more DR to handle the contrast in more scenes is a good thing. You can always increase contrast in post if you want, but you can't bring back what was lost so given I choice I'd rather have more DR than less.
You won't get any argument from me about overdone HDR, but that really has nothing to do with camera DR, that's just a matter of bad processing (and maybe poor taste).
Bottom line is that with the D3x, I don't have to bracket multiple exposures as often as with my old cameras, which is definitely a benefit.
Jeff
Thanks for the thoughtful reply. Gosh, did slide film really only have 6 stops of DR? I always seemed to do alright with it provided that I kept the sun over my shoulder, and I never bothered to bracket. But here's an interesting point, which might just be relevant to your topic -
Back then (in the '70s and '80s) I was shooting both 35mm and MF. But it always seemed to me that the MF transparencies had more exposure latitude (and hence DR) than the 35mm. The MF slides certainly seemed less fussy about metering and my percentage of keepers was higher. However, they couldn't have had more DR, in reality, because it was exactly the same emulsion in both cameras (Ektachrome). The picture editors back then always insisted on MF though.
Thinking about the main thrust of this topic, I can't personally believe that my MF digital back has huge amounts more DR than a 35mm DSLR. After all, 4 stops, let alone 6 stops, is a massive difference - you would see it straight away. It would be like using fill-in flash or having a huge reflector set up all the time. The shots from my MF back just look like photographs - the balance between shadows and highlights is pretty much what I would expect from any camera.
What I do notice though, is that these 39 MP MF files seem to have a great deal of latitude in post-processing, which is where the confusion may arise. Not in the highlight areas so much, where it is still easy to blow them and recovery be of no use, but in the shadow areas. It depends on the specific subject and file, of course, but it is possible to push the shadow areas in some cases by +2 EV or more and the result will be amazingly good, at least at 50 and 100 ISO. But this is not strictly DR, is it? DR is measured from an unedited exposure, as I understand it.
John
-
Hi,
This example may demonstrate the need for DR:
[attachment=23101:_DSC5560.jpg]
Full image: http://http://echophoto.dnsalias.net//ekr/...KarerSee_01.jpg (http://echophoto.dnsalias.net//ekr/images/DRExample/KarerSee_01.jpg)
The trees are in shadow and the mountainside has snow illuminated by bright sunlight. Initially I intended to use HDR for the image but found that I could extract about the information I needed from the image having least exposure. Admittedly, it could be done better, the forrest behind the lake is a bit flat.
The image was shot on APS-C, Sony Alpha 100, no camera that really excels in DR...
Best regards
Erik
Erik
With all due respect, I think this is a silly example. There are obvious limits to what photography can successfuly achieve. If one was a painter, you might set your easel up and do something with this. As a photographer, I would take one look at this subject and think "No way, forget it". If my eyes were unsure, my meter would set me straight. In this case, you don't need more DR, just more commonsense
John
-
In this case, you don't need more DR, just more commonsense
Or HDR ...
Cheers,
Bart
-
What I do notice though, is that these 39 MP MF files seem to have a great deal of latitude in post-processing, which is where the confusion may arise. Not in the highlight areas so much, where it is still easy to blow them and recovery be of no use, but in the shadow areas. It depends on the specific subject and file, of course, but it is possible to push the shadow areas in some cases by +2 EV or more and the result will be amazingly good, at least at 50 and 100 ISO. But this is not strictly DR, is it? DR is measured from an unedited exposure, as I understand it.
Hi John,
That's correct, it's tonemapping, but having a lot of capture DR helps to avoid the image falling apart at the shadows. It also helps achieving better quality at higher ISOs.
Cheers,
Bart
-
Back then (in the '70s and '80s) I was shooting both 35mm and MF. But it always seemed to me that the MF transparencies had more exposure latitude (and hence DR) than the 35mm. The MF slides certainly seemed less fussy about metering and my percentage of keepers was higher. However, they couldn't have had more DR, in reality, because it was exactly the same emulsion in both cameras (Ektachrome). The picture editors back then always insisted on MF though.
I don't believe this is the case, John. The larger format will usually tend to deliver higher dynamic range especially when the emulsion or pixel quality is the same. When the emulsion or sensor design is not the same, as in the examples of the D3X and P65+, one having a CMOS sensor and the other a CCD, then anomalies can occur. The additional area of the P65 sensor should allow for a greater DR than the smaller sensor in the D3X, but this is apparently not the case because of differences in technology and design.
The principle here, when comparing identical compositions of course, is that any image detail on the the larger format must result from greater illumination (greater number of photons) than the same detail on the smaller format, whether such detail be in the shadows, mid-tones or highlights. It cannot be otherwise, provided the same exposure is applied in each case. If the same exposure is applied, then each unit of area in both formats, say each square mm of film, must receive the same amount of light to be correctly (or equally) exposed. The greater illumination applied to the same detail (which of course covers a larger area on the larger format emulsion), results in a cleaner and more detailed MF image when both images are compared at the same size on monitor or print.
-
Ray,
Thanks for good comments. I would almost expect you pointing out the viewing distance issue. I would add that it's working the way that we don't increase viewing distance proportionally to image size. Think of an IMAX or Omnimax theatre. One of the reasons to print big is that you can see the print at reasonable distance and still feel immersed in the picture.
Hi Erik,
I would put it another way. I'd claim that, more frequently, we do not reduce viewing distance in proportion to the smaller image size of an A4 or A3 or A2 print that may hang on our wall. It's why I bought the 24" wide Epson 7600 a few years ago. It's my experience that, in the average house and sitting room, where there may hang a number of A4 or A3 photos, one does not appreciate the fine detail contained in such prints on most occasions when one happens to glance in the direction of one of those photos. One generally has to make a deliberate effort to get off one's chair and walk right up to the print to view it from a distance of about 1.5x its diagonal.
This is why I prefer to hang small prints in a hallway. I can view them from an appropriately close distance every time I pass them, and on occasions I may linger a few seconds to admire the texture of a rock or tree trunk in the foreground. I would never attempt to place a large panorama in a hallway unless it could withstand close scrutiny without appearing fuzzy.
I remember well the early days of the introduction of the HDTV standard. There were different groups lobbying for different resolution standards. In those early days, large HD screens were horrendously expensive and the group lobbying for the 720p standard had a good point. They would demonstrate that in order to see the difference between 720p and 1080p on a 'then' affordable but still expensive big screen TV of 32" or so, one would have sit no further than a metre away from the screen, which is far closer than most people would want to sit.
However, that point of view of the 720p lobby has proved to be rather short-sighted. Big HD screens are now affordable, which is why I recently bought a 12th generation 65" Panasonic plasma TV. It's better to be stuck on a 1080p standard for a few decades than a 720p standard.
Incidentally, when I sit about 2.5 metres from my 65" Panasonic I get a noticeably 'immersive' experience from a good quality 2mp image. If the image were significantly higher resolution than 2mp, then in order to appreciate that higher resolution, I would have to sit closer than 2.5 metres and would then have to turn my head from left to right to see the entire image clearly. I think the same principle applies in the cinema. Despite the screen being relatively huge, you still wouldn't want to sit closer than 1.5x the screen diagonal for an immersive experience.
Now to get back to the DR issue. Having more pixels on the same sensor does not help DR capability in any way, as I understand. In fact it probably hinders it slightly because the total read noise can be greater, unless there is some compensating technological development in other areas, and there usually is of course. I imagine if Phase One had produced an MF full frame 24mp back instead of the 60mp P65+ FF DB, the D3X might not have had a DR advantage.
I personally don't have any trouble determining which of my cameras has the higher dynamic range, unless the DR of the cameras I'm testing is very similar. If anyone makes a claim that camera 'A' has 4 to 6 stops greater DR than camera 'B', but is unable or unwilling to demonstrate such differences with visual examples, then I think such claims can be taken with a grain of salt.
A 4 to 6 stop difference in DR, or even half of that difference, should be very easy to demonstrate. One might disagree over a 1/4 stop, or a 1/3rd stop or even 1/2 a stop either way, but 2 stops or more?? No way!!
I can appreciate to a certain extent the objection of the busy professional who has done his research into his need for a particular DB, and who may already own lots of fine MF lenses. If DR is not an issue for him and he has lots of good reasons for using MF equipment, why should he take the time to demonstrate clearly and precisely the DR differences between his DB and any 35mm equipment he happens to own? What's in it for him? To do a thorough job requires care and patience. If one doesn't do a thorough and meticulous job, getting the ETTR exactly right with both cameras, the FoV and lens quality favouring neither one camera nor the other, the focussing and DoF the same, and the lighting and the scene exactly the same in both shots, then one's test will be criticised and considered flawed.
On the other hand, exhortations from such professionals directed at people like us, to persuade us to test the DBs for ourselves and get first-hand experience, are also not practical, unless one is fairly sure beforehand that one is in need of that additional performance and resolution of a DB regardless of the DR issue. Assuming there's a store available that hires out the latest DB's and 35mm gear with appropriate lenses, it would be an expensive exercise that could hardly be justified in order to settle just the DR issue.
If a 4 stop increase in DR (over FF 35mm) were sufficient reason to persuade me (and many others) to go the very expensive MF route, I'm sure the MFDB sales reps would be falling over themselves to demonstrate this DR difference. I wonder why they are not? My guess is because they can't.
Making claims of a 4-6 stop DR advantage then becomes merely a sales ploy to encourage people to visit their nearest MF dealer to check out these extraordinary claims for themselves. Having taken the trouble to do this, and after getting the opportunity to handle a Phase MFDB system and perhaps realise it's not as heavy and cumbersome as one imagined, and having seen close-up examples in the showroom of the marvelous detail that a larger format and high pixel-count camera is capable of, sheer material greed may take over, (I want it. Bugger the DR) and then there's the possibility of a sale. Oops! Have I given the game away?
Wouldn't I look foolish if I were to spend the price of a new Canon 5D2 body in order to hire a P65+ back, MF camera body with lens, plus a D3X and lens in order to confirm that a bunch of guys with PhDs at DXO Labs were actually right. I'm not that silly, ya know!
-
Bart,
The image I posted has actually detail in the foreground but not on the tree trunks. So I'm pretty satisfied with it. I tried to use HDR but had problems with double contours, color fringing and mapping in general. The only advantage I found with HDR was less noise in the foreground.
Best regards
Erik
Or HDR ...
Cheers,
Bart
-
Hi,
I don't get your point.
Let's assume that we have a uniform surface evenly illuminated that is imaged on say 100x100 pixels, that is a 10000 pixels. These 10000 pixels are independent in exposure. So we have 10000 exposures with average value of 20000 and assumed Stddev of 141. So Signal is 20000 , noise 141 and SNR = 141. Let's reduce exposure to 200, no we have noise about 14 from photo statistic but we need also to take readout noise into account.
If we assume readout noise to be 10 electrons and add noise in quadrature we get a noise level of 17 electrons so SNR would be 200/17 that is about 12.
Simplicistic?
Erik,
A bit simplistic, but it does get the point across. Joofa wants to work with single pixels in a "real" scene. To get the information in a single flat field of 100x100, he would need 10,000 exposures and would wear out his shutter in the process. Even then, the validity of the data would be questionable, since the camera would likely moved a bit on the tripod, and his single pixel would represent another area in the image.
In practice, one would take duplicate exposures of a flat field and use a program such as ImagesPlus or Iris to split out the red, blue, and green channels for separate analysis and work with raw data. Each channel has a different DR and other characteristics. The green channel is most often chosen for analysis, since the eye is more sensitive to green and the Bayer array has 2 green pixels for each red and blue pixel. One would then crop to a representative area near the center, say 200 x 200 pixels, and then subtract one image from other after adding an offset to prevent negative numbers. This removes the effect of nonuniform illumination and the variations in sensitivity among the individual pixels. As Emil has pointed out in his excellent treatise on noise, PRNU (pixel response non-uniformity) is a major contributor to noise, especially at higher exposures. One would then divide the standard deviation of the subtracted channels by the square root of two, since the measured SD represents noise for two exposures. Neglecting read noise, which is minimal at high exposure values, this would give the shot noise. To obtain the noise at various exposures, one needs to take multiple pairs of data by varying the shutter speed.
For Canon cameras that do not clip read noise, RN can be determined by taking a very short exposure with the lens cap on. Nikons clip read noise and more complicated methods are needed. As you mentioned noise adds in quadrature.
See Roger Clark (http://www.clarkvision.com/imagedetail/evaluation-1d2/index.html) and Peter Facey (http://www.brisk.org.uk/photog/d3index.html) for details.
This analysis is rather cumbersome, but can be carried out by a serious photographer and I have performed such an analysis for my Nikon D3 after a few missteps. It is an interesting exercise, but the analysis has already been done by professionals at DXO:
[attachment=23126:Phase1_65_.gif]
Regards,
Bill
-
This analysis is rather cumbersome, but can be carried out by a serious photographer and I have performed such an analysis for my Nikon D3 after a few missteps. It is an interesting exercise, but the analysis has already been done by professionals at DXO:
Hi Bill,
I agree, it is tedious but one does learn a thing or two about one's specific camera body, and about rigid testing conditions and procedures. I did a similar analysis (http://www.openphotographyforums.com/forums/showthread.php?t=4771) for my Canon 1Ds3 when there was no reliable info available. My conclusion was and engineering DR of 11.3, DxO recorded 11.22 on their body. The results are in enough of an agreement for me to trust the DxO data.
Cheers,
Bart
-
An issue here is the almost total lack of hard technical information from the manufacturers. Compare that with film, where you have highly detailed specification sheets from Kodak, Ilford and so on for all their emulsions, with characteristic curves and everything you need.
Kodak provides plenty of technical information for its sensors,
http://www.kodak.com/global/en/business/IS...l?pq-path=14425 (http://www.kodak.com/global/en/business/ISS/Products/Fullframe/index.jhtml?pq-path=14425)
Dalsa provides plenty of technical information for most of it sensors, but not the "exclusive" ones like the 60MP:
http://www.dalsa.com/sensors/Products/sensors.aspx (http://www.dalsa.com/sensors/Products/sensors.aspx)
Sony provides specs for many of its smaller "digicam" sensors but not for its SLR sensors,
and Canon and Panasonic provide none, AFAIK.
There is a clear pattern: spec. sheets are provided for sensors being offered openly to any potential customer (customers being camera makers, not us camera buyers), but not for sensors used only in-house, or limited special somewhat exclusive relationships between sensor maker and camera maker.
With film the customers are photographers, so we are provided with all the specs.
-
Let's assume that we have a uniform surface evenly illuminated that is imaged on say 100x100 pixels, that is a 10000 pixels. These 10000 pixels are independent in exposure. So we have 10000 exposures with average value of 20000 and assumed Stddev of 141. So Signal is 20000 , noise 141 and SNR = 141. Let's reduce exposure to 200, no we have noise about 14 from photo statistic but we need also to take readout noise into account.
If we assume readout noise to be 10 electrons and add noise in quadrature we get a noise level of 17 electrons so SNR would be 200/17 that is about 12.
Hi Erik,
One has to consider what the spatial array of 100x100 pixels is describing. For simplicity I shall only talk about shot noise. If we ignore some differences in pixels, then under reasonable illumination on a uniform surface the 100x100 array of pixels is considered to be describing the variation on any single pixel in this array. I.e., instead of acquiring 100x100=10,000 images and then picking a fixed pixel location and figuring out its standard deviation you just take one single image and consider that the spatial block of 100x100 pixels is describing more or less the equivalent variation you would have encountered temporally on a fixed pixel location in 10,000 images. So this experiment is a convenience for shot noise calculation and we can work with just one image instead of 10,000 by thinking of equivalence of spatial and temporal variations.
However, in a real, natural image (say of your dog or cat) you can't do this in this manner. Since, each of the 100x100 pixels is now sampled from a different underlying distribution because the input light as received by the sensor is no longer uniform as before. You can still of course compute mean and std.dev. and other statistics on this 100x100 array, but such statistics are not describing the shot noise on any single pixel any more in the usual sense.
Joofa
-
Hi Bill,
I agree, it is tedious but one does learn a thing or two about one's specific camera body, and about rigid testing conditions and procedures. I did a similar analysis (http://www.openphotographyforums.com/forums/showthread.php?t=4771) for my Canon 1Ds3 when there was no reliable info available. My conclusion was and engineering DR of 11.3, DxO recorded 11.22 on their body. The results are in enough of an agreement for me to trust the DxO data.
Bart,
I agree that doing one's own testing is a very worthwhile procedure for the reasons you cite. I certainly learned a lot from my own tests, which were in agreement with those of Peter Facey and DXO. Roger Clark's findings for the 1DMII are also in agreement with DXO, as are my previously posted calculations for the Hasselblad using the Kodak spec sheets and Roger's noise model. It seems that the ones who disagree with DXO are those who have not done their own rigorous tests but are relying on subjective impressions gained from merely looking at pictures.
What is your expert opinion on the conflicting reports with regard to the dynamic range of MFDBs? Factors rather than actual DR must be involved.
Regards,
Bill
-
What is your expert opinion on the conflicting reports with regard to the dynamic range of MFDBs? Factors rather than actual DR must be involved.
Hi Bill,
As I said earlier, DR is only a (small) factor in the explanation of the perceived dynamic range. One other factor of importance, IMHO, is differences in MTF due to sensor array dimensions. A larger sensor array will result in a larger magnification factor of a given FOV on the sensor. A larger magnification will mean that a different part of the MTF response curve is used, a part with potentially a significant higher response. That higher response will allow to discriminate microcontrast that may otherwise result in loss of discrimination due to veiling glare or other optical contributions to the system MTF.
Here's what I mean:
(http://www.xs4all.nl/~bvdwolf/main/downloads/LargeSensorBenefit.png)
At the Nyquist frequency of a simulated smaller 'perfect' sensor, the MTF response is 56%. At that same resolution, due to larger magnification, a larger 'perfect' sensor will have 82% MTF response. That will allow for much more optical compromises before the same detail is drowning in the background detail. In addition the larger sensor array will have more resolution capability because output requires less magnification for same size output.
The complicating factors are of course things like the optical contributions, or rather detractions, from lenses and/or AA-filters. But also MTF killing factors like defocus and lack of DOF play a role. It makes it hard to demonstrate with cameras with different optical systems, that's why I isolated the sensor MTF because there are enough other factors to confuse an unambiguous analysis.
Hope that helps the discussion past the point of DR alone.
Cheers,
Bart
-
Bart,
That would be consistent with Mark Dubovoy's comment on full texture. You mentioned it before, and I thought it was relevant.
Another think I'm considering is that we need to sharpen more because of the OLP filter. If sharpening is done without a gradient mask it would also amplify noise. Also small thing add up.
Best regards
Erik
Hi Bill,
As I said earlier, DR is only a (small) factor in the explanation of the perceived dynamic range. One other factor of importance, IMHO, is differences in MTF due to sensor array dimensions. A larger sensor array will result in a larger magnification factor of a given FOV on the sensor. A larger magnification will mean that a different part of the MTF response curve is used, a part with potentially a significant higher response. That higher response will allow to discriminate microcontrast that may otherwise result in loss of discrimination due to veiling glare or other optical contributions to the system MTF.
Here's what I mean:
(http://www.xs4all.nl/~bvdwolf/main/downloads/LargeSensorBenefit.png)
At the Nyquist frequency of a simulated smaller 'perfect' sensor, the MTF response is 56%. At that same resolution, due to larger magnification, a larger 'perfect' sensor will have 82% MTF response. That will allow for much more optical compromises before the same detail is drowning in the background detail. In addition the larger sensor array will have more resolution capability because output requires less magnification for same size output.
The complicating factors are of course things like the optical contributions, or rather detractions, from lenses and/or AA-filters. But also MTF killing factors like defocus and lack of DOF play a role. It makes it hard to demonstrate with cameras with different optical systems, that's why I isolated the sensor MTF because there are enough other factors to confuse an unambiguous analysis.
Hope that helps the discussion past the point of DR alone.
Cheers,
Bart
-
As I pointed, I'm normally allergic to curves and graphics, (exept when it comes to my incomes )
But I think this thread is more and more interesting and I must say that valuable information is showing all the time.
Some explainations have already emerged and we are getting closer to understand the reasons of the OP questions.
That itself is a great accomplishment.
IMO, the reason why that has been possible here is that from the begining, instead of loading another war thread,
the attitude was: "ok, some experienced photographers are seeing something that contradict the measurements, why is so?"
This open mind to others views before condemning and the desire to understand are making this thread constructive.
I'm very happy to feel that and thanks again to Erik for his hability to conduct his "baby" (the thread),
and the great attitude of all the participant involved.
That really is an example of the high level of debate that Lu-La can have, independently of the divergences.
-
Hi,
I'm most thankful on both issues, and I also feel it has been a constructive discussion.
Best regards
Erik
and the great attitude of all the participant involved.
That really is an example of the high level of debate that Lu-La can have, independently of the divergences.
-
That would be consistent with Mark Dubovoy's comment on full texture. You mentioned it before, and I thought it was relevant.
That's why I added a graph to illustrate.
Another think I'm considering is that we need to sharpen more because of the OLP filter. If sharpening is done without a gradient mask it would also amplify noise. Also small thing add up.
Yes, it all adds up, although a good sharpening method can sharpen detail more than noise. In fact sharpening the output of an OLPF image sharpens quite well, without the more horrible low frequency aliasing-artifacts, while an unfiltered image can hardly be sharpened without exaggerating the aliasing artifacts. The shaded areas in my graph represent aliasing territory which folds back to the lower spatial frequencies. But then again, this only applies for in-focus high spatial frequencies. A larger sensor also means less DOF (or more diffraction when trying to get more DOF), and that's an effective MTF killer that could also require sharpening to compensate.
Things are not simple, we have to choose our compromises.
Cheers,
Bart
-
Kodak provides plenty of technical information for its sensors,
Dalsa provides plenty of technical information for most of it sensors, but not the "exclusive" ones like the 60MP:
Sony provides specs for many of its smaller "digicam" sensors but not for its SLR sensors,
and Canon and Panasonic provide none, AFAIK.
With film the customers are photographers, so we are provided with all the specs.
Yes, true. But I was thinking only of MF at the time I posted, and what we don't have AFAIK is the specification for the sensor when it is implemented in a particular back (P-45 or HD-39, for example). I was assuming that the firmware, channel ampliifiers, signal processing path and so forth would vary between manufacturers, so that the response (and perhaps DR) of a Kodak KAV 39000 sensor might be different in a Phase, Hasselblad or Leaf back. But this is just speculation on my part.
There is an awful lot of highly technical stuff in this thread which I don't pretend to understand (and I never understood film chemistry either, I just used it) - fortunately one doesn't need to get that deep in order to take photographs. But others do need to understand it to make cameras, and write software, and it is a very good thing that there are people here who also have a grip on the technicalities so that manufacturers of cameras and software can be properly scrutinised and held to account.
From my own very set-of-the-pants subjective viewpoint - having shot film in all its forms since 1958, and for many years having done all my own B/W processing and E6 processing and B/W printing in 35mm, MF, and large format - I would say that my Hasselblad CFV-39 digital back has more DR than colour transparency film, about the same DR as B/W negative film, and perhaps a little less DR than colour negative film. Unfortunately I do not have a Nikon D3X to compare it with, but if anyone is offering . . .
John
-
Hi,
Lloyd Chambers got a Leica S2 for testing yesterday (or so). He compares to the D3X. Unfortunately a pay site, but I'm interested in his writing so I find worth it.
Anyway, he says that although the AF on the Leica is correct under optimal conditions it doesn't work at all for his shooting. He has at least one sample showing advantage of S2 over D3X in sharpness but also DOF limitation.
Best regards
Erik
That's why I added a graph to illustrate.
Yes, it all adds up, although a good sharpening method can sharpen detail more than noise. In fact sharpening the output of an OLPF image sharpens quite well, without the more horrible low frequency aliasing-artifacts, while an unfiltered image can hardly be sharpened without exaggerating the aliasing artifacts. The shaded areas in my graph represent aliasing territory which folds back to the lower spatial frequencies. But then again, this only applies for in-focus high spatial frequencies. A larger sensor also means less DOF (or more diffraction when trying to get more DOF), and that's an effective MTF killer that could also require sharpening to compensate.
Things are not simple, we have to choose our compromises.
Cheers,
Bart
-
... what we don't have AFAIK is the specification for the sensor when it is implemented in a particular back (P-45 or HD-39, for example). I was assuming that the firmware, channel ampliifiers, signal processing path and so forth would vary between manufacturers, ...
True, but none of those is likely to produce a variation vaguely close to the alleged five or six stops, so sensor spec. sheets and third party test results for 35mm format DSLRs are easily accurate enough to answer the question that motivated this thread ...
... in the negative. None of the secondary factors raised in this debate looks likely to come close to supporting that five or six stop claim scientifically defensible. Probably the biggest factor causing a correction to per photosite DR values is the higher pixel count and related dithering when images are viewed at equal (apparent) size ... but that is only good for about a factor of at most sqrt(60/22)=1.65 in S/N ratio, about 0.7 of a stop, for the extreme case of 60MP vs 22MP.
Based on the impressive discussion in this thread so far, the best conclusion is that any genuinely perceived differences are related to
- sensor resolution
- lens resolution, and
- other lens quality factors like less damage to shadow regions by flare.
The latter two are likely to be exaggerated if comparisons are based on 35mm with zoom lenses vs MF with primes.
I have read that flare typically limits the DR of the image reaching the focal plane to something of order of 12 stops through lightening of shadow regions, though of course that depends on the pattern of lighting in the subject.
-
Hi Bill,
As I said earlier, DR is only a (small) factor in the explanation of the perceived dynamic range. One other factor of importance, IMHO, is differences in MTF due to sensor array dimensions. A larger sensor array will result in a larger magnification factor of a given FOV on the sensor. A larger magnification will mean that a different part of the MTF response curve is used, a part with potentially a significant higher response. That higher response will allow to discriminate microcontrast that may otherwise result in loss of discrimination due to veiling glare or other optical contributions to the system MTF.
Here's what I mean:
In his preliminary testing of the Leica S2, Digilloyd reports that the DR of the Leica (actually a tweener camera and not a full MFDB) seemed less than that of the D3x and that the latter camera handles high contrast better.
Regards,
Bill
-
I have read that flare typically limits the DR of the image reaching the focal plane to something of order of 12 stops through lightening of shadow regions, though of course that depends on the pattern of lighting in the subject.
Indeed, see also http://luminous-landscape.com/forum/index.php?s=&showtopic=42158&view=findpost&p=352656 (http://luminous-landscape.com/forum/index.php?s=&showtopic=42158&view=findpost&p=352656).
Cheers,
Bart
-
But microcontrast is not DR. Why would an improvement in MTF affect DR?
-
As I said earlier, DR is only a (small) factor in the explanation of the perceived dynamic range. One other factor of importance, IMHO, is differences in MTF due to sensor array dimensions. A larger sensor array will result in a larger magnification factor of a given FOV on the sensor. A larger magnification will mean that a different part of the MTF response curve is used, a part with potentially a significant higher response. That higher response will allow to discriminate microcontrast that may otherwise result in loss of discrimination due to veiling glare or other optical contributions to the system MTF.
I'm not convinced that such factors are particularly significant to the perception of DR, Bart. It sounds as though you are making excuses for someone who doesn't know what DR is, or doesn't know what he's doing. On the other hand, perhaps it's me who doesn't understand the significance of DR.
I'd like to refer to those extraordinary results from DXO Mark that describe the D3X as having 1.33 stops greater DR than the P65+ (at the pixel level) and 0.68 stops greater DR when both images are downsampled to 8"x12" size for printing.
From the perspective of the practical photographer, this is what I think the results mean. Correct me if I'm wrong.
If I shoot a scene with a high SBR (subject brightness range) using the P65+ and the D3X, using the same focal length of lens with each camera but underexposing the D3X shot by 1 & 1/3rd stops, then both images should be perceived as having the same dynamic range.
Now it's true that the P65 shot will have a wider FoV because it has a bigger sensor and we used the same FL of lens in each case, but we're examining image quality at 100% on screen, not admiring the composition. As long as the full brightness range of the scene is represented in both shots, the different FoV is irrelevant to our purposes. The pixel pitch of the D3X is very close to that of the P65, so the 100% crops that are examined will be about the same size, which makes the comparison easier.
It may be the case that the P65+ shot will be perceived as being slightly crisper due to its lack of an AA filter, but such subtle effects should not change the perception of DR. Even if there's a mismatch of lenses and we use the most expensive Digitar lens with the P65 but an average Nikkor lens with the D3X, the better resolution in the P65 image would be apparent across the entire image and not confined to the shadows. The brain should be able to make allowances, realise that a better quality lens has been used for one of the images, and that the DR of both images is still the same.
In the comparison of images downsampled to 8"x12" size, I understand that equal FoV images would be the source, which perhaps creates a slight problem due to the different aspect ratio of the two cameras. Ignoring this slight mismatch, the DXO results are telling me when the D3X shot (underexposed just 2/3rds of a stop in this instance) and the P65+ shot fully exposed, are downsampled to 8"x12" at 300dpi, thus discarding some resolution from both images, but discarding more resolution from the P65+ image, both images will appear to have the same DR, and presumably the same resolution.
Have I got this right?
A high dynamic range capability in a camera allows for greater exposure latitude. I used to prefer shooting with negative film, rather than transparencies for that reason. I could underexpose one stop and still get acceptable shadows in situations where the subject had a wide brightness range. With slide film one had to be more precise and sometimes risk overexposure. It was widely accepted that B&W film had the highest DR, color negetive film next, and slide film a couple of stops lower.
The DR differences between these types of film was pretty obvious and didn't require endless debate. If camera A has a significantly high DR than camera B, then the shadows in camera A shots will be cleaner, more detailed and of higher quality.
If the DR differences are in the order of 4-6 stops, the cleaner and more detailed shadows from camera A would be so obvious they would smack you in the face. If the DR difference is in the order of 1/4 to 1/2 a stop, then. even with the best intentions, other factors may skew the result and it may not be perfectly clear which camera has the better DR.
Of course, you do have a point about the potential of the MTF of a lens to affect the perception of a camera's DR. This affect would be most obvious in the corners of the image where the MTF response of most lenses is worst. One could use this characteristic of lenses to demonstrate that an APS-C (cropped-format) camera has a higher DR than a FF 35mm.
This is how you would do it. Take a good 35mm format lens that has a reasonably flat MTF response to the corners and virtually no fall-off when used on the APS-C format. Take another FF lens of equivalent focal length for use with the 35mm format camera, but make sure it's a lens with significant MTF fall-off at the corners, as well as severe vignetting. Such a lens could still be impressively sharp in the centre.
Select a scene with a very high brightness range, but arrange the composition so that the shadows are all in the corners of the frame in both shots. Make sure you expose correctly, aiming for an accurate ETTR by bracketing in 1/3rd stop increments.
Display 100% crops of the highlights and mid-tones of both image, demonstrating the superior tonality and resolution of the FF 35mm format and the fact that both shots have received a correct ETTR. Then display 100% crops of the shadows in the corners, lightening the shadows so that one can see clearly the better detail and lower noise in the APS-C shot.
This is an extreme example. I don't think many photographers would be hoodwinked by those results. However, the choice of F stop when comparing different format cameras can certainly skew the results against the smaller format, but I would think that resolution softness can be distinguished from noise.
If you want to exaggerate the resolution advantage of the DB in comparison with the FF 35mm format, then you would use wider apertures for a shallow DoF. For example, if we start with the premise that most lenses are sharpest at F5.6, then we choose F5.6 as the aperture used with the MF format.
In order to equalize DoF in both shots, we then have to use F4 (or even wider) with FF 35mm. But a 35mm lens at F4 is generally not at its sharpest. (Are we testing lenses here, or are we supposed to be testing camera sensor performance?)
If one is aware of such factors, then any serious comparison should include shots of the same scene taken at other apertures. In addition to pairing F5.6 with F4, we need to see F8 on the MFDB and F5.6 on the FF 35mm, or F11 on the MFDB and F8 on the FF 35mm.
I believe it's the case that MFDBs will generally produce noticeably sharper results than FF 35mm DSLRs when a shallow DoF is sought, even when the pixel count is the same. The smaller pixel pitch of the smaller format requires a higher MTF from the lens, not a lower MTF. The MTF at F4 is generally lower than it is at F5.6. However, the MTF is generally higher at F8 than at F11.
-
But microcontrast is not DR. Why would an improvement in MTF affect DR?
Hi Emil,
It wouldn't, because they are different things. However, it might explain what is incorrectly perceived as increased DR (even though it's actually a cumulation of different things). The partial explanation by a potentially boosted MTF does not only boost microcontrast, all spatial frequencies get a better start, but that is going to be reduced by several optical components/effects in actual practice.
Cheers,
Bart
-
Emil,
If you consider your image here: http://theory.uchicago.edu/~ejm/pix/20d/te...3noise-8bit.gif (http://theory.uchicago.edu/~ejm/pix/20d/tests/noise/grad-DOI-1d3noise-8bit.gif)
How do you think it would be effected by increasing MTF? Text would be darker and have higher edge contrast?
If we increase sharpening, to compensate for OLP filtering how would that affect the image? If we do that indiscriminately I guess noise would go up?
Best regards
Erik
But microcontrast is not DR. Why would an improvement in MTF affect DR?
-
I'm not convinced that such factors are particularly significant to the perception of DR, Bart. It sounds as though you are making excuses for someone who doesn't know what DR is, or doesn't know what he's doing. On the other hand, perhaps it's me who doesn't understand the significance of DR.
I'd like to refer to those extraordinary results from DXO Mark that describe the D3X as having 1.33 stops greater DR than the P65+ (at the pixel level) and 0.68 stops greater DR when both images are downsampled to 8"x12" size for printing.
Hi Ray,
I hope you're not misreading my intentions. My motivation is to understand why some people perceive a difference in image quality. The mere fact that they attribute it to a factor that demonstrably cannot explain their perception is of lesser importance (not everybody has the same technological insight).
Based on the measurements that I can personally verify (by rigorous testing), I have no problem accepting the DxOmark data as generally reliable (within reasonable per-sample variation limitations). As such, the DR alone cannot explain the perceived differences (unless those differences are grossly overstated). So, assuming the observations of 'some' difference in (presumably) shadow detail performance and/or highlight recovery capability are real (i.e. not just post-purchase-justification), the assumption of other contributing (because they could add up) factors playing a role seems valid.
From the perspective of the practical photographer, this is what I think the results mean. Correct me if I'm wrong.
If I shoot a scene with a high SBR (subject brightness range) using the P65+ and the D3X, using the same focal length of lens with each camera but underexposing the D3X shot by 1 & 1/3rd stops, then both images should be perceived as having the same dynamic range.
Well, assuming the same lens (not just the focal length, because design plays a role and we want to eliminate as many variables as possible), and the same sensor design (there are potentially major differences between CCD and CMOS designs, depending on the actual implementation and supporting electronics), one would expect similar results. However, there are already several variables. Whether they are relevant would need to be determined. Rigorous testing is not easy, one preferably only varies a single parameter to derive some level of relevance.
Now it's true that the P65 shot will have a wider FoV because it has a bigger sensor and we used the same FL of lens in each case, but we're examining image quality at 100% on screen, not admiring the composition. As long as the full brightness range of the scene is represented in both shots, the different FoV is irrelevant to our purposes. The pixel pitch of the D3X is very close to that of the P65, so the 100% crops that are examined will be about the same size, which makes the comparison easier.
Well, I'm not convinced that the difference in magnification factor does not lead a casual observer to make a wrong assumption. Human vision is not very good at objective quantification, we need direct and comparable scale samples side by side for a somewhat valid opinion.
It may be the case that the P65+ shot will be perceived as being slightly crisper due to its lack of an AA filter, but such subtle effects should not change the perception of DR.
Well, there you have another subjective factor clouding the judgement. I'm not yet convinced that the effects are as subtle, QED.
Even if there's a mismatch of lenses and we use the most expensive Digitar lens with the P65 but an average Nikkor lens with the D3X, the better resolution in the P65 image would be apparent across the entire image and not confined to the shadows. The brain should be able to make allowances, realise that a better quality lens has been used for one of the images, and that the DR of both images is still the same.
One can only hope, but I'm not optimistic.
In the comparison of images downsampled to 8"x12" size, I understand that equal FoV images would be the source, which perhaps creates a slight problem due to the different aspect ratio of the two cameras.
I tried to accomodate for that by using an aspect ratio indifferent metric in my graph by using the picture height (LW/PH) criterion, which disregards certain differences in aspect ratio, i.e. width. It was a deliberate part of my suggestion (always trying to eliminate variables).
Ignoring this slight mismatch, the DXO results are telling me when the D3X shot (underexposed just 2/3rds of a stop in this instance) and the P65+ shot fully exposed, are downsampled to 8"x12" at 300dpi, thus discarding some resolution from both images, but discarding more resolution from the P65+ image, both images will appear to have the same DR, and presumably the same resolution.
Have I got this right?
More or less, yup. There are considerations of the actual downsampling algorithm used, but I'm feeling generous tonight ;-) .
A high dynamic range capability in a camera allows for greater exposure latitude. I used to prefer shooting with negative film, rather than transparencies for that reason. I could underexpose one stop and still get acceptable shadows in situations where the subject had a wide brightness range. With slide film one had to be more precise and sometimes risk overexposure. It was widely accepted that B&W film had the highest DR, color negetive film next, and slide film a couple of stops lower.
The DR differences between these types of film was pretty obvious and didn't require endless debate. If camera A has a significantly high DR than camera B, then the shadows in camera A shots will be cleaner, more detailed and of higher quality.
Well, that depends on one's background. The debates, although more in the scientific community (where calibration of one's detectors plays a role), and the (somewhat) subjective preferences (yet an interesting stastistical basis) amongst professional photographers, were there.
If the DR differences are in the order of 4-6 stops, the cleaner and more detailed shadows from camera A would be so obvious they would smack you in the face. If the DR difference is in the order of 1/4 to 1/2 a stop, then. even with the best intentions, other factors may skew the result and it may not be perfectly clear which camera has the better DR.
Which is what we are, at least I am, trying to dissect. Where does the disconnect (between observed and quantifiable) come from?
Of course, you do have a point about the potential of the MTF of a lens to affect the perception of a camera's DR. This affect would be most obvious in the corners of the image where the MTF response of most lenses is worst. One could use this characteristic of lenses to demonstrate that an APS-C (cropped-format) camera has a higher DR than a FF 35mm.
Yes, I've read "How to Lie with Statistics" ;-(
Cheers,
Bart
-
Which is what we are, at least I am, trying to dissect. Where does the disconnect (between observed and quantifiable) come from?
Hi Bart,
I don't think I've misconstrued your intentions here. Best of luck! . This issue is particularly paradoxical when one takes into consideration that comparison that Michael made, at A3+ print size, using a P45+ and the 15mp Powershot G10. If one looks at the DXO scores comparing those sensors at the normalised print size of 8"x12", the P45+ is way ahead with 1/4 the noise at 18% grey, 4.7 bits better color sensitivity and 2 bits better tonal range, yet at double the size of the DXO normalized print, experienced photographers were not able to distinguish any significant qualitative difference between the two prints, except the shallower DoF in one of the prints which was an indication that that was the print made from the P45 image.
If Michael had equalized the DoF in both shots, using F18 or F22 with the P45+ instead of F11, we might have had the even more amazing situation whereby the G10 print was confused as orininating from the P45 camera because it appeared slightly sharper.
It's because such differences can go unnoticed in a 'real world' print (as opposed to a print that is specifically designed to highlight such differences, such as a resolution line chart which highlights resolution differences), that I'm very skeptical that relatively small differences in 'effective' lens MTF performance, due to sensor size & pixel pitch, could have any significant bearing on the perception of DR.
When I first visited the DXO Mark site, I was a bit confused about the apparent conflict between SNR and dynamic range. We tend to associate a good SNR figure with a high DR, yet there are many examples of sensors having a worse DR but a better SNR on the DXO website. I presume this situation arises because DXO only measure SNR at 18% grey, which is a fairly light shade of grey, in the mid-tone range. I presume, if DXO were to measure SNR at significantly darker shades of grey, the SNR figures would correspond more closely with the DR figures, ie. a higher DR would tend to be associated with a higher SNR at the darker shade of grey. Is this correct?
Since DXO Mark don't provide results at larger, normalised print sizes, such as 23"x31" for the P65+ at 300 dpi, I can only speculate how the DR of an interpolated D3X image would compare, at this size.
Since the D3X image has a 1.33 stop DR advantage at the pixel level, and a 2/3rds stop advantage at the downsampled size of 8"x12", I think it would be reasonable to guess that at any larger, interpolated print size, the DR of the D3X print would not be worse than that of the P65+, or at least not perceptibly worse, because the resolution advantage of the P65+ image, in those darker shades of grey in the 11th and 12th stops, probably doesn't exist. It's been obliterated by noise.
If the resolution advantage in the darker shadows hasn't been recorded by the P65 sensor in the first instance, then it cannot magically materialize, however big the print size. Nor, I would imagine, could any slight advantage from an effectively higher MTF from the lens provide that additional detail in areas of the image that are swamped in noise.
However, I would of course expect to see higher resolution in the 23"x31" print from the P65, in the range from the lower mid-tones to the highlights. I would expect to see sharper eylashes and a slightly smoother, creamier complexion of the model's face. But I wouldn't expect to see a more detailed chihuahua, hiding in the dark, under the couch.
-
Since the D3X image has a 1.33 stop DR advantage at the pixel level, and a 2/3rds stop advantage at the downsampled size of 8"x12",
The downsampling algorithm has an effect on SNR improvement. For e.g., is it binning (box filter), or sinc, or lanczos, or something else. In general the implicit assumption has been that pixels are simply binned/averaged, derived from all those discussions on smaller/larger sensor SNR behavior. However, in practise, one would use binning seldomly for downsampling as more sophisticated methods are used. The good news is that it is possible to derive an analytical, closed form, solution to SNR improvement in a generalized downsampling operation that can cater to different methods, say lanczos downsampling.
But the point is not to go into detail of such a solution, but just to emphasize that quoting a figure such as '2/3rds stop advantage at the downsampled size of 8"x12"' without mentioning how the image was downsampled is not full piece of information.
Joofa
-
Ray,
If, if, if
Lloyd Chambers havs an excellent review of the S2 on his DAP site. It's a pay site, in my view worth every penny. Lloyd published a couple of DNG images from the Leica S2 and the Nikon D3X. unfortunately on quite restrictive terms. I have played a bit with those images and came up with a few interesting observations. The key observation is that although the S2 images have without any doubt higher resolution the Nikon images are smoother and have much less artifacts. So I essentially prefer the Nikon images but the look of the images is much dependent on the processing parameters.
Best regards
Erik
Hi Bart,
I don't think I've misconstrued your intentions here. Best of luck! . This issue is particularly paradoxical when one takes into consideration that comparison that Michael made, at A3+ print size, using a P45+ and the 15mp Powershot G10. If one looks at the DXO scores comparing those sensors at the normalised print size of 8"x12", the P45+ is way ahead with 1/4 the noise at 18% grey, 4.7 bits better color sensitivity and 2 bits better tonal range, yet at double the size of the DXO normalized print, experienced photographers were not able to distinguish any significant qualitative difference between the two prints, except the shallower DoF in one of the prints which was an indication that that was the print made from the P45 image.
If Michael had equalized the DoF in both shots, using F18 or F22 with the P45+ instead of F11, we might have had the even more amazing situation whereby the G10 print was confused as orininating from the P45 camera because it appeared slightly sharper.
It's because such differences can go unnoticed in a 'real world' print (as opposed to a print that is specifically designed to highlight such differences, such as a resolution line chart which highlights resolution differences), that I'm very skeptical that relatively small differences in 'effective' lens MTF performance, due to sensor size & pixel pitch, could have any significant bearing on the perception of DR.
When I first visited the DXO Mark site, I was a bit confused about the apparent conflict between SNR and dynamic range. We tend to associate a good SNR figure with a high DR, yet there are many examples of sensors having a worse DR but a better SNR on the DXO website. I presume this situation arises because DXO only measure SNR at 18% grey, which is a fairly light shade of grey, in the mid-tone range. I presume, if DXO were to measure SNR at significantly darker shades of grey, the SNR figures would correspond more closely with the DR figures, ie. a higher DR would tend to be associated with a higher SNR at the darker shade of grey. Is this correct?
Since DXO Mark don't provide results at larger, normalised print sizes, such as 23"x31" for the P65+ at 300 dpi, I can only speculate how the DR of an interpolated D3X image would compare, at this size.
Since the D3X image has a 1.33 stop DR advantage at the pixel level, and a 2/3rds stop advantage at the downsampled size of 8"x12", I think it would be reasonable to guess that at any larger, interpolated print size, the DR of the D3X print would not be worse than that of the P65+, or at least not perceptibly worse, because the resolution advantage of the P65+ image, in those darker shades of grey in the 11th and 12th stops, probably doesn't exist. It's been obliterated by noise.
If the resolution advantage in the darker shadows hasn't been recorded by the P65 sensor in the first instance, then it cannot magically materialize, however big the print size. Nor, I would imagine, could any slight advantage from an effectively higher MTF from the lens provide that additional detail in areas of the image that are swamped in noise.
However, I would of course expect to see higher resolution in the 23"x31" print from the P65, in the range from the lower mid-tones to the highlights. I would expect to see sharper eylashes and a slightly smoother, creamier complexion of the model's face. But I wouldn't expect to see a more detailed chihuahua, hiding in the dark, under the couch.
-
The downsampling algorithm has an effect on SNR improvement. For e.g., is it binning (box filter), or sinc, or lanczos, or something else. In general the implicit assumption has been that pixels are simply binned/averaged, derived from all those discussions on smaller/larger sensor SNR behavior. However, in practise, one would use binning seldomly for downsampling as more sophisticated methods are used. The good news is that it is possible to derive an analytical, closed form, solution to SNR improvement in a generalized downsampling operation that can cater to different methods, say lanczos downsampling.
But the point is not to go into detail of such a solution, but just to emphasize that quoting a figure such as '2/3rds stop advantage at the downsampled size of 8"x12"' without mentioning how the image was downsampled is not full piece of information.
Joofa
Joofa,
Yes, I know that different methods of downsampling and interpolation can affect the result. But what can I do about it? I'm not aware of anyone having the full information on any topic or issue under the sun. There are always some unanswered questions, approximations, assumptions or uncertainties.
As regards a dynamic range comparison between the D3X and P65+, the DXOmark results are the only 'substantiated' results I'm aware that exist. I'm therefore not able to quote any other results whether they may be more accurate, complete or incomplete.
I understand that DXOmark results are based on an analysis of the RAW file from each camera they test, thus avoiding endless dispute about which RAW converter may be best for which camera. I don't know how DXO derive the 8x12" normalised image. Perhaps those DR figures relating to the downsampled images are merely mathematical calculations based on the original analysis of the sensor at the pixel level.
I'm also aware from comparisons of my own images that downsampling an image tends to reduce noise, so the DXOmark results for their 8mp 8"x12" normalised images tend to be credible for me.
-
Ray,
If, if, if
Lloyd Chambers havs an excellent review of the S2 on his DAP site. It's a pay site, in my view worth every penny. Lloyd published a couple of DNG images from the Leica S2 and the Nikon D3X. unfortunately on quite restrictive terms. I have played a bit with those images and came up with a few interesting observations. The key observation is that although the S2 images have without any doubt higher resolution the Nikon images are smoother and have much less artifacts. So I essentially prefer the Nikon images but the look of the images is much dependent on the processing parameters.
Best regards
Erik
Erik,
I've noticed the other interesting thread on the Lloyd Chambers' review of the Leica S2, but I'm not subscribed to that pay site. I'm not sure I need to see in-depth reviews of equipment I'm never likely to buy. At the moment I feel as though I have enough photographic equipment but not enough time to use it thoroughly and experiment with its potential. Nevertheless your report on your comparison of the DNG images you downloaded is interesting.
As regards equipment purchases, the next logical step for me would be a 5D2. My highest resolution camera is still the 15mp 50D. If I hadn't previously purchased a Nikkor 14-24/2.8, which led to the purchase of a D700, I would probably by now have a 5D2 and one of the new 17 or 24 TSE lenses (or perhaps both).
On my recent river cruise in Europe and Russia, I frequently carried around my neck both the D700 with 14-24, and D50 with 17-55. I must have got asked about 50 times, often by complete strangers, why I was carrying two cameras.
I did my best to explain that I effectively get a high quality F2.8 zoom from 14-88mm and that no such single F2.8 zoom lens of similar quality and range exists anywhere at any price.
When I've got time, I'd like to do a thorough comparison of these two camera and lens combinations, comparing the Nikkor at 24mm with the Canon at 17mm (which is effectively 27mm on FF). My impression so far is that the advantages of the greater DoF of the cropped format, combined with the advantages of the image stabilisation of the Canon EF-S 17-55/2.8, sometimes result in better image quality, or at least not worse image quality.
For example, shooting a fresco in a church where flash and tripod is not allowed, I might use the D700 at ISO 1600 and F2.8 but can use the 50D at ISO 800 or even 400 at F2.8 because of IS, which the Nikkor lens doesn't have. When photographing something more 3-dimensional in low light, when the DoF at F2.8 is too shallow on full frame, I might need to use F5 at ISO 1600 with the D700, as opposed to F2.8 at ISO 200 with the 50D, on the basis that IS provides a 2-stop shutter speed advantage and the smaller format provides a 1.6 stop DoF advantage. In such circumstances, there's no doubt that the 50D at ISO 200 has lower noise and higher DR than the D700 at ISO 1600. On the other hand, there's no doubt that the Nikkor 14-24 at 24mm and F5 is sharper than the Canon 17-55 at 17mm and F2.8.
Cheers!
-
Joofa,
Yes, I know that different methods of downsampling and interpolation can affect the result. But what can I do about it? I'm not aware of anyone having the full information on any topic or issue under the sun. There are always some unanswered questions, approximations, assumptions or uncertainties.
I'm also aware from comparisons of my own images that downsampling an image tends to reduce noise, so the DXOmark results for their 8mp 8"x12" normalised images tend to be credible for me.
Hi Ray,
I know what you are saying and I agree with it. SNR would typically improve in downsampling operations. The reason is that in this case the signal has a correlation structure where as the noise by assumption doesn't.
I just wanted to point out that according to the link that Bart provided a few messages up DXO also assumes a "binning-like, fit-to-size" operation in their "Print" DR comparison. Where as, in reality, hardly any serious resampling operation would use binning. (That is another topic that in the first place I personally think that even the usual binning SNR improvement is not fully correct as done by DXO, and others, but that is a much debated topic by now).
Regards,
Joofa
-
Now, many experienced observers clearly see a 4-6 stop advantage with MFDBs over DSLRs. I don't have any issue with that, but I cannot understand where it is coming from.
It is coming from their lack of knowledge about what DR is.
Typical useful per-pixel DR in DSLR's is 8-9 stops (SNR criteria of 12dB (http://www.guillermoluijk.com/article/digitalp02/ruido_0db_12db.jpg), which can be considered a good reference for photographers). A back with 6 extra stops over that, i.e. 14-15 stops of DR, could virtually capture 99% of HDR scenes in real life with a single shot and without noise problems in the deep shadows. Obviously this is not the case.
This kind of scene is one of the most challenging for any digital camera:
(http://img683.imageshack.us/img683/6107/bracketing.jpg)
It has about 15 stops of DR:
(http://img708.imageshack.us/img708/5562/hist.gif)
Has anyone with a back tried to shoot in a dark room with a sunshine outdoor window, and was able both to preserve the highlights in the window and display no visible or very acceptable noise in the deep shadows inside? if a back does this then I'll believe it has 6 stops of advantage over my 350D's DR.
A 15 stops DR camera is very far from what can be found in the market today so the 4-6 stops advantage claim is nonsense.
Regards
-
It is coming from their lack of knowledge about what DR is.
Typical useful per-pixel DR in DSLR's is 8-9 stops (SNR criteria of 12dB...) ... A back with 6 extra stops over that, i.e. 14-15 stops of DR ...
Thanks. And it is even more obvious, since the MF sensor and back makers have given as a hard 12-stop upper limit (engineering SNR, criteria of 0dB) via their spec. sheets. The only way for high end DSLR's to be six stops worse would to start require them being down at about 6 stops "usable" DR. How high a SNR threshold would be needed to justify claiming that low a DR for DSLR's?
-
Typical useful per-pixel DR in DSLR's is 8-9 stops (SNR criteria of 12dB (http://www.guillermoluijk.com/article/digitalp02/ruido_0db_12db.jpg), which can be considered a good reference for photographers).
Hi Guillermo,
The 'typical' 12 dB criterion is a bit arbitrary IMHO, for some it's too noisy already, and for others there is a lot that can still be done with decent noise reduction. Besides, there are already differences in Raw converter behavior with all noise reduction zeroed out. Obviously, it is also about a single Raw file, not e.g. a 'Zero-noise'ed or HDR file.
Just trying to understand what you are saying, but as I understand it it is more of an example than a statement, correct?
Cheers,
Bart
-
Just trying to understand what you are saying, but as I understand it it is more of an example than a statement, correct?
Just wanted to point that a 4-6 stops advantage in DR is too huge to be true, no matter what SNR criteria we use (roughly 6 extra stops means 64 times or 36dB higher SNR in the deep shadows). There is not such difference even between a compact camera and a FF DSLR, where relative differences in sensor size are much larger.
Yes 12dB is an arbitrary criteria but surely will satisfy many more photographic applications and users than 0dB.
The only way for high end DSLR's to be six stops worse would to start require them being down at about 6 stops "usable" DR. How high a SNR threshold would be needed to justify claiming that low a DR for DSLR's?
Without trying to be very precise, I did some SNR measurements for the Canon 5D, 7D and 5D2. Looking at this per-pixel SNR plots:
(http://www.guillermoluijk.com/tutorial/noisedr/curvassnr.gif)
To justify a DR of 6 stops, on a Canon 5D2 at ISO100 we must rise up to a SNR=27dB!
BTW according to that plot, the per-pixel DR(0dB) for the Canon 5D2 would be 11 stops (DxO says 11,16 stops).
Regards
-
...
-
Hi,
I just looked at a couple of well exposed files from Leica S2 and Nikon D3X courtesy of Lloyd Chambers, when I increased exposure in Lightroom to +3.6 and added fill light 35% the Leica S2 image was clearly worse than the D3X, so I couldn't see any increase in DR in the dark end. Regarding highlights I don't know.
Best regards
Erik
Thanks. And it is even more obvious, since the MF sensor and back makers have given as a hard 12-stop upper limit (engineering SNR, criteria of 0dB) via their spec. sheets. The only way for high end DSLR's to be six stops worse would to start require them being down at about 6 stops "usable" DR. How high a SNR threshold would be needed to justify claiming that low a DR for DSLR's?