Luminous Landscape Forum
Equipment & Techniques => Cameras, Lenses and Shooting gear => Topic started by: BernardLanguillier on October 22, 2008, 11:58:56 pm
-
Dear all,
Long ago, a theory was often heard stating that high ISO noise and low ISO DR were 2 faces of the same coin, meaning very tightly related. This would have been related to the size of the photosites, large photosites being able to capture more photons.
The additional photons would enable the reading of a meaningful signal when illumination level is low (meaning when high ISO is used), and would also saturate less quickly resulting in less blown highlights, and therefore more DR.
The MFDB have never met these predictions by far, and now the Sony A900 with its measured DR of 12.6 stop with only average high ISO noise is also clearly not playing by this theory.
What gives?
Cheers,
Bernard
-
The MFDB have never met these predictions by far, and now the Sony A900 with its measured DR of 12.6 stop with only average high ISO noise is also clearly not playing by this theory.
What gives?
Cheers,
Bernard
What I find significant, Bernard, is that dpreview's figure of 12.6 stops for the A900 is a 'best case' scenario which involves extreme adjustments to RAW images in ACR, such as -2.65 EV and a -50 contrast.
I can find no comparable ACR adjustments in relation to other similar cameras that the A900 is compared with. What we have for other cameras are DR figures in relation to the default ACR settings and in relation to 'auto' ACR settings. The figure of 12.6 stops I therefore find misleading. The ACR default settings provide the worst DR figure; the ACR 'auto' settings provide a significantly better figure, and the best DR outcome results from the extreme settings mentioned above which have not been applied to any other camera in the review.
The only DR comparisons I can find in the review relate to in-camera jpegs, and those are quite surprising. Whilst the A900 excels in jpeg mode, having 1.2 stops greater DR than the 5D and 0.8 stops greater DR than the 1Ds3, the 5D is shown as having 0.4 stops greater DR than the D700.
One has to be clear when one is comparing apples with apples. The 12.6 figure for the A900 appears to be an orange.
-
I can find no comparable ACR adjustments in relation to other similar cameras that the A900 is compared with. What we have for other cameras are DR figures in relation to the default ACR settings and in relation to 'auto' ACR settings. The figure of 12.6 stops I therefore find misleading. The ACR default settings provide the worst DR figure; the ACR 'auto' settings provide a significantly better figure, and the best DR outcome results from the extreme settings mentioned above which have not been applied to any other camera in the review.
As usual the default Adobe Camera RAW conversion delivers less dynamic range than JPEG from the camera (the same contrasty tone curve and very little noise reduction in shadows). Simply switching to 'Auto' in the ACR conversion dialog reaps huge rewards, increasing the dynamic range to around 10.5 stops. The very best we could get out of a raw file manually was around 11.3 EV, which is pretty good (though not quite up to the standard set by the Nikon D3 or the Fujifilm S5 Pro).
-
What I find significant, Bernard, is that dpreview's figure of 12.6 stops for the A900 is a 'best case' scenario which involves extreme adjustments to RAW images in ACR, such as -2.65 EV and a -50 contrast.
I can find no comparable ACR adjustments in relation to other similar cameras that the A900 is compared with. What we have for other cameras are DR figures in relation to the default ACR settings and in relation to 'auto' ACR settings. The figure of 12.6 stops I therefore find misleading. The ACR default settings provide the worst DR figure; the ACR 'auto' settings provide a significantly better figure, and the best DR outcome results from the extreme settings mentioned above which have not been applied to any other camera in the review.
In my opinion, any DR test in which the results depend on the RAW developer tool used, and the way to use it (settings), is nonsense.
The ability to capture a certain DR by a camera is a _hardware feature_, and it is clearly hardware limited by: saturation in the highlights, and a given (and there is not a unique criteria for this, but some criteria must be chosen if DR results are to be compared) value for the minimum signal to noise ratio considered valid in the shadows. And all this can (and perhaps should) be measured for each of the individual RGB channels separately.
So the dynamic range of a camera would be the difference in EV between the maximum recordable value in its RAW file, and the value of signal that keeps a given ratio against the std deviation of the noise. Any test involving a particular RAW developer and/or particular settings on a given RAW developer such as ACR for me is crap.
Problems may arise to measure this if one day manufacturers start to do destructive noise reduction on the RAW file, meaning by destructive any processing that reduces noise but also destroys texture and captured detail (that would be the day when RAW is not RAW anymore). In that case, the camera would achieve a better measured DR but in a somewhat unfair way and DR comparisions should be attached to captured detail tests.
And another very important issue that has to be taken into account in DR is the sensor resolution: DR is measured at a pixel level, by calculating SNR on each pixel. If a camera has more Mpx than another one with the same pixel SNR, the highest resolution camera will provide a higher overall image DR (not pixel DR), since SNR will increase when rescaling the image to match the size of the lower resolution sensor.
In general: DR is a hardware feature, and it can never depend on the software used to process the RAW file.
The maximum recordable DR decreases the higher the ISO set. Maximum recordable DR is reached when exposure of the scene is maximum (i.e. ETTR) and the lowest electronic camera ISO is used.
O course I am always talking about shooting in RAW. JPEG is another story which depends on camera's software, and not always the highest DR camera will provide the highest visible DR JPEG files.
BR
-
According to DPR: "Shadow range is more complicated, in our test we stop measuring values below middle gray as soon as the luminance value drops below our defined 'black point' (about 2% luminance) or the signal-to-noise ratio drops below a predefined value (where shadow detail would be swamped by noise), whichever comes first."
In other words, DPR stops measuring DR at 2% regardless of the SNR, and it is notable that in DPR's "best" settings using ACR that the graph ends abruptly at bottom (shadow) end of the spectrum. Probably DPR could have continued measuring DR using their method well past 13 stops by simply continuing to halve the values (1%, .5%, .25%, etc) until they actually reached an unacceptable SNR; after all, they were still recording hundreds of photons when they stopped measuring any further. To illustrate, you can see from these DPR graphs of:
the A900
(http://a.img-dpreview.com/reviews/sonydslra900/Samples/DR/DR-raw.jpg)
from: http://www.dpreview.com/reviews/sonydslra900/page24.asp (http://www.dpreview.com/reviews/sonydslra900/page24.asp)
and the D3
(http://www.dpreview.com/reviews/NikonD3/Samples/DR/DR-RAW.png)
from: http://www.dpreview.com/reviews/NikonD3/page20.asp (http://www.dpreview.com/reviews/NikonD3/page20.asp)
that the bottom of the measurement has been cut off. Measuring DR using a calibrated Stouffer Step Wedge is also problematic as it merely measures gray, which is not the same as usable DR. So much for the DR half of the equation, without which, it is impossible to draw any direct parallels between DR and performance at higher ISOs.
-
QUOTE (DPReview (1Ds MkIII), page 21 @ August 2008)
As usual the default Adobe Camera RAW conversion delivers less dynamic range than JPEG from the camera (the same contrasty tone curve and very little noise reduction in shadows). Simply switching to 'Auto' in the ACR conversion dialog reaps huge rewards, increasing the dynamic range to around 10.5 stops. The very best we could get out of a raw file manually was around 11.3 EV, which is pretty good (though not quite up to the standard set by the Nikon D3 or the Fujifilm S5 Pro).
Jan,
I'm not sure how valid it is to compare test results from the past with current test results. Doesn't dpreview warn against doing this because of slight changes in procedure and methodology?
However, not to nit pick, I would accept that the DR of the A900 looks as though it's the best of any current model of DSLR, at base ISO. Those who complain about increasing pixel count robbing them of increases in DR should be pleased with the performance of the A900.
-
In general: DR is a hardware feature, and it can never depend on the software used to process the RAW file.
Point taken, Guillermo, however we all have to use some type of converter if we shoot in RAW mode. Most people probably use ACR or Lightroom for a variety of reasons and are unlikely to change converters because of some very slight, pixel-peeping advantage that another converter may have in respect of one particular model of camera.
Whenever I've compared ACR with another converter, whether DPP, C1, Breezebrowser, RSP etc, I've found ACR to be either better or very close with regard to recovery of highlight and shadow detail. DCRAW is probably the exception but is not easy to use, is it?
-
Point taken, Guillermo, however we all have to use some type of converter if we shoot in RAW mode. Most people probably use ACR or Lightroom for a variety of reasons and are unlikely to change converters because of some very slight, pixel-peeping advantage that another converter may have in respect of one particular model of camera.
If ACR or any other software does not manage to exploit all the information contained in a given RAW file, it is not fair to show a comparision table saying 'Camera X has a dynamic range of Y' using that software. Dpreview should clearly head their tables with 'If you use ACR and you set these adjustments, you will be able to enjoy with Camera X a dynamic range of Y', and they don't do it. It is not a limitation of the camera, but of the software and the ability of the Dpreview team with the particular settings.
I will give you an example: ACR can obtain more dynamic range in the highlights from Fuji RAF files than the Fuji software. Is that because the Fuji guys are stupid or the ACR guys are very clever? no way. Studying with some depth a RAF file is easy to see that the Fuji software voluntarily discards some highlight information to avoid magenta cast problems that arise when heavily reducing exposure in the Fuji R sensor in strongly exposed shots. ACR allows for this stronger exposure correction, and therefore can suffer from the magenta cast.
If I read a comparision about cameras (not about RAW developers), I want to know what _my Fuji_ can capture, and later will decide which tool to use to fit my needs. But I don't want to see a comparision table subject to a particular software since: this software may not be the one I will use, and/or this software may change tomorrow.
BR
-
If ACR or any other software does not manage to exploit all the information contained in a given RAW file, it is not fair to show a comparision table saying 'Camera X has a dynamic range of Y' using that software. Dpreview should clearly head their tables with 'If you use ACR and you set these adjustments, you will be able to enjoy with Camera X a dynamic range of Y', and they don't do it. It is not a limitation of the camera, but of the software and the ability of the Dpreview team with the particular settings.
To be fair to Dpreview, that's more or less what they do. They produce DR comparisons based on the camera's jpeg output, then they demonstrate what ACR can do with various settings such as 'default', 'auto', and maximum retrieval of highlight and shadow detail after extreme adjustments of EV and contrast.
They also mention that recovery of grayscale detail in wedges is not the same as recovery of full color detail. They haven't extended their testing to include comparisons of ACR's ability to reconstruct the color detail within those extreme DR ranges and how it might vary with different camera models. But there is no doubt that their DR figures of RAW images are a systemic part of the ACR software, just as many lens tests, such as those at Photozone, are tests of both lens and sensor used.
I don't get any sense that Dpreview is trying to hide this fact, nor can I see how another method of determining DR without the use of a specific RAW converter would help the consumer.
-
Jan,
I'm not sure how valid it is to compare test results from the past with current test results. Doesn't dpreview warn against doing this because of slight changes in procedure and methodology?
I was responding to your complaint that they didn't perform similar testing in the past with (somewhat weak) evidence that they did.
But you can't have it both ways.
However, not to nit pick, I would accept that the DR of the A900 looks as though it's the best of any current model of DSLR, at base ISO. Those who complain about increasing pixel count robbing them of increases in DR should be pleased with the performance of the A900.
I don't feel so sure about that.
The methodology seems flawed in various respects, or there is a lack of disclosure regarding the methodology.
I'm with Guillermo on this one.
But there is no doubt that their DR figures of RAW images are a systemic part of the ACR software, just as many lens tests, such as those at Photozone, are tests of both lens and sensor used.
Yes, at least it's fairly consistent, but as you yourself say: you can't compare former tests with current tests.
One of the reasons is of course that ACR may change from version to version...
-
According to DPR: "Shadow range is more complicated, in our test we stop measuring values below middle gray as soon as the luminance value drops below our defined 'black point' (about 2% luminance) or the signal-to-noise ratio drops below a predefined value (where shadow detail would be swamped by noise), whichever comes first."
In other words, DPR stops measuring DR at 2% regardless of the SNR, and it is notable that in DPR's "best" settings using ACR that the graph ends abruptly at bottom (shadow) end of the spectrum. Probably DPR could have continued measuring DR using their method well past 13 stops by simply continuing to halve the values (1%, .5%, .25%, etc) until they actually reached an unacceptable SNR; after all, they were still recording hundreds of photons when they stopped measuring any further.
This shows how robust and convincing is Dpreview's criteria: they consider the low end of the DR when signal drops below a value, a signal that THEY set by chosing a particular tone curve into ACR through the chosen settings, even if there could be a lot of valid information (good enough SNR) below that point that could perfectly be of use to the user.
And even if they chose the SNR criteria, I wonder how they can measure noise in ACR if there is no guarantee whether ACR is applying or not any noise reduction since that software is a black box from the user's point of view.
Doing always the same wrong procedure, can never be considered a valid procedure.
BR
-
In my opinion, any DR test in which the results depend on the RAW developer tool used, and the way to use it (settings), is nonsense.
The ability to capture a certain DR by a camera is a _hardware feature_, and it is clearly hardware limited by: saturation in the highlights, and a given (and there is not a unique criteria for this, but some criteria must be chosen if DR results are to be compared) value for the minimum signal to noise ratio considered valid in the shadows. And all this can (and perhaps should) be measured for each of the individual RGB channels separately.
So the dynamic range of a camera would be the difference in EV between the maximum recordable value in its RAW file, and the value of signal that keeps a given ratio against the std deviation of the noise. Any test involving a particular RAW developer and/or particular settings on a given RAW developer such as ACR for me is crap.
I agree entirely with Guillermo. Achieving greater dynamic range through highlight recovery as done by DPReview involves making up data for the clipped channels, blue and red for daylight exposures with most Bayer array cameras. Shown below is a typical daylight exposure with the Nikon D3 along with a histogram of the raw data. The shot is underexposed, since the green channel is about 2/3 stop from clipping. The red channel is 1 1/3 stops from clipping, and the blue channel is one stop from clipping. With increasing exposure such that the green channel becomes clipped, the other 2 channels would initially still have data and highlight recovery would be possible using data from the non-clipped channels. When the red channel becomes clipped, highlight recovery is no longer possible.
The green channel will have the highest dynamic range, since it receives the most exposure. If you determine DR from a demosaiced image you are combining three different DRs, but what weighting should be used? Probably not an arithemetic average, since the eye is most sensitive to green. Also, in a demosaiced image, the green is contaminated by interpolation from adjacent blue and red pixels. For an excelllent technical article on these matters, see Emil Martinec (http://theory.uchicago.edu/~ejm/pix/20d/tests/noise/noise-p3a.html).
[attachment=9148:BirdsHistogram.jpg]
And another very important issue that has to be taken into account in DR is the sensor resolution: DR is measured at a pixel level, by calculating SNR on each pixel. If a camera has more Mpx than another one with the same pixel SNR, the highest resolution camera will provide a higher overall image DR (not pixel DR), since SNR will increase when rescaling the image to match the size of the lower resolution sensor.
In general: DR is a hardware feature, and it can never depend on the software used to process the RAW file.
The maximum recordable DR decreases the higher the ISO set. Maximum recordable DR is reached when exposure of the scene is maximum (i.e. ETTR) and the lowest electronic camera ISO is used.
If the sensor size is held constant, and the pixel density varied to give different resolutions, the total DR will not change much, since the amount of light collected remains the same for both sensors. As Emil (http://theory.uchicago.edu/~ejm/pix/20d/tests/noise/noise-p3.html#pixelsize) and Guillermo explain, one is trading off dynamic range for resolution. When the image of the smaller pixel camera (with the same total sensor area) is downsized so as to have the same resolution as the larger pixel camera, noise will be decreased by pixel binning. The per pixel noise standard deviation as used by DPReview in determining dynamic range can be misleading.
Of course, the reason for increasing megapixels is to allow a larger print size with acceptable image detail. When the print size with the higher megapixel camera is increased so that pixel binning is no longer done, the higher megapixel camera would cease to have a DR advantage, but it would still have better image detail and most likely finer gained noise, which might appear less objectionable to an observer.
Of course I am always talking about shooting in RAW. JPEG is another story which depends on camera's software, and not always the highest DR camera will provide the highest visible DR JPEG files.
Quite true. The raw file has linear scene referred data, whereas the rendered JPEG is object referred. The rendering process involves DR compression and application of a tone curve--see the white paper by Karl Lange (http://wwwimages.adobe.com/www.adobe.com/products/photoshop/family/prophotographer/pdfs/pscs3_renderprint.pdf) on the Adobe site. Highlight and shadow DR as discussed by DPReview may make sense for JPEG rendered images which have a shoulder, linear segment, and knee, but this concept makes no sense for a raw file which is linear.
Bill
-
When the image of the smaller pixel camera (with the same total sensor area) is downsized so as to have the same resolution as the larger pixel camera, noise will be decreased by pixel binning.
In fact a pure binning strategy by averaging 2x2 pixels into 1 final pixel, which according to Emil's statistics doubles the SNR, is the worst improvement we can expect. When reescaling to 50% the original size with other algorithms like Bicubic, I have done tests on noisy images and the SNR improvement is even higher (the noise histogram gets thinner than with a simple binning averaging). And of course the quality of resizing is also better.
BR
-
In fact a pure binning strategy by averaging 2x2 pixels into 1 final pixel, which according to Emil's statistics doubles the SNR, is the worst improvement we can expect. When reescaling to 50% the original size with other algorithms like Bicubic, I have done tests on noisy images and the SNR improvement is even higher (the noise histogram gets thinner than with a simple binning averaging). And of course the quality of resizing is also better.
BR
Yes, the whole matter becomes rather complex. Another thing to consider is that pixel binning is best done in hardware, not software, as explained in this article (http://www.photomet.com/pm_solutions/library_encyclopedia/index.php) on the Photometrics site. That article discusses 2 by 2 pixel blending. In that case, 4 pixels are binned into one in hardware. The binned superpixel can be read with the same read noise as a single pixel. However, if binning is done after the fact in software, four separate and discrete read noises are involved.
As far as I know, hardware pixel binning is currently done only in scientific grade CCDs. The most advanced 35 mm style digital cameras now use CMOS, but MF digitals still use CCDs. It would be interesting to have a digital 32 MP camera with 2:1 and 4:1 hardware pixel binning.
Bill
-
DPR's tests are often a source of amusement. Apparently they don't understand that it is impossible for a linear medium like RAW to encode more stops of DR than there are bits in the encoding (12 for the A900).
One has to be clear when one is comparing apples with apples. The 12.6 figure for the A900 appears to be an orange.
Or a lemon
-
I agree entirely with Guillermo. Achieving greater dynamic range through highlight recovery as done by DPReview involves making up data for the clipped channels, blue and red for daylight exposures with most Bayer array cameras.
Bill,
ACR is what I use to process my RAW images, and I believe it's what most photographers use. What Dpreview does is what I do, and many others, except I try to avoid pushing exposure to the exteme right because I'm aware that ACR's attempt to reconstruct color data is sometimes unsatisfactory. An obvious example is the shift towards green in skies where the blue channel is clipped.
I don't see that knowing the DR of the sensor helps the photographer since the RAW image is useless without a converter. However, I can see that those who are in the process of developing their own RAW converter would find such information useful. Also, such information may be of academic interest to some, even though of no immediate practical benefit.
For the purpose of comparing the DR of two different cameras, the main criterion should be that the same procedure is applied to both cameras using the same converter. Dpreview applied 3 different groups of settings in ACR (default, auto and extreme) and mentioned the fact that reconstruction of color data would not necessarily be accurate with the extreme settings of -2.65 EV and -50 contrast. What they didn't do was use the same RAW converter and same settings with the other cameras in the review. All we got for the other cameras were jpeg comparisons. I would criticise them for that.
-
Bill,
ACR is what I use to process my RAW images, and I believe it's what most photographers use. What Dpreview does is what I do, and many others, except I try to avoid pushing exposure to the exteme right because I'm aware that ACR's attempt to reconstruct color data is sometimes unsatisfactory. An obvious example is the shift towards green in skies where the blue channel is clipped.
I don't see that knowing the DR of the sensor helps the photographer since the RAW image is useless without a converter. However, I can see that those who are in the process of developing their own RAW converter would find such information useful. Also, such information may be of academic interest to some, even though of no immediate practical benefit.
For the purpose of comparing the DR of two different cameras, the main criterion should be that the same procedure is applied to both cameras using the same converter. Dpreview applied 3 different groups of settings in ACR (default, auto and extreme) and mentioned the fact that reconstruction of color data would not necessarily be accurate with the extreme settings of -2.65 EV and -50 contrast. What they didn't do was use the same RAW converter and same settings with the other cameras in the review. All we got for the other cameras were jpeg comparisons. I would criticise them for that.
It is important to be clear when providing test results, to understand what is being tested. If a purported test of dynamic range is contaminated by the raw converter's highlight recovery algorithm, it is not testing dynamic range; it is bringing in the native white balance of the sensor (which determines when different color channels clip) and a host of other factors. If DPR doesn't know what is going on under the hood in the raw converter, they don't know what combination of camera properties they are measuring.
-
Dear all,
Long ago, a theory was often heard stating that high ISO noise and low ISO DR were 2 faces of the same coin, meaning very tightly related. This would have been related to the size of the photosites, large photosites being able to capture more photons.
The additional photons would enable the reading of a meaningful signal when illumination level is low (meaning when high ISO is used), and would also saturate less quickly resulting in less blown highlights, and therefore more DR.
The MFDB have never met these predictions by far, and now the Sony A900 with its measured DR of 12.6 stop with only average high ISO noise is also clearly not playing by this theory.
What gives?
Cheers,
Bernard
High ISO noise and low ISO DR need not be closely related. The noise floor at low ISO in CMOS DSLR's is controlled by the noise in the ISO amplifier and A/D converter, while the noise floor at high ISO is controlled by the noise of the sensor array. Different components, different noise characteristics, and the one that effects high ISO midtone and shadow noise doesn't enter into the determination of low ISO DR.
For CCD sensor cameras, there tends to be a tighter relationship; data I've seen shows a simple relationship between low and high ISO electronic noise, and so the statement of a coupling between high ISO noise and low ISO DR is more accurate.
While I have singled out electronic read noise for the purpose of illustrating a counterexample, photon noise is independent of the sensor type, and you are right in that the more photons one captures, the less apparent photon noise becomes. Dynamic range can be increased either by capturing more photons, or by lowering electronic read noise. The read noise can be decouple between low and high ISO, but the photon capture halves for each doubling of ISO (more precisely the number of photons captured at the saturation point of the raw data is halved), and so that aspect of DR is directly correlated between high and low ISO.
When considering DR, it is important to realize that it is scale dependent; larger photosites tend to have more DR by capturing more photons, but also eat up more sensor real estate to capture those photons. Photosites of half the pixel pitch (1/4 the area) with the same efficiency per unit area, will have 1/4 the photon capacity per photosite, but when binned will still capture the same number of photons over the whole sensor (while delivering twice the resolution in the bargain). The proper figure of merit is DR per unit area of sensor (or better, DR of an area of sensor that is a fixed percentage of the area of the sensor). When this is done, there is a mild advantage to larger photosites, but not as much as the comparison of DR per photosite would mislead one to believe.
-
For the purpose of comparing the DR of two different cameras, the main criterion should be that the same procedure is applied to both cameras using the same converter.
Nonsense like this is why I ignore your posts, but since Emil responded then I see you have thrown more garbage into the discussion that should be responded to.
As I see it, the issue is that if you really care about increasingly negligible differences in image quality between any two cameras, then you should also care about the entire workflow that results in the final image. What you are arguing is analogous to shopping for a car that has maximum horsepower, but then disregarding the owner's manual and putting regular gas instead of premium into the tank.
-
the main criterion should be that the same procedure is applied to both cameras using the same converter.
Dpreview applied 3 different groups of settings in ACR (default, auto and extreme) and mentioned the fact that reconstruction of color data would not necessarily be accurate with the extreme settings of -2.65 EV and -50 contrast. What they didn't do was use the same RAW converter and same settings with the other cameras in the review. All we got for the other cameras were jpeg comparisons. I would criticise them for that.
Ray, DPreview will never be able to compare cameras using the same tool and parameters unless they repeat all the tests for the old cameras with the last version software everytime a new camera appears in the market, and we know they won't do that. This is a flaw in the method, and you agree.
But there is more, there are also flaws in the concept of what they are doing, such as using highlight recovery to decide the right end of the DR, or setting the low end of the DR according to a threshold signal level, specially when they set ACR to apply a tone curve that will affect this threshold and is out of their control.
I am not an academic expert, what I know is from reading some informed sites and practicing with DCRAW in front of my PC, and I could quickly set up a standard procedure much more rigurous than Dpreview's to yield real and practical DR figures for any camera, now and tomorrow. If they don't do this is simply because DPreview is a end-user oriented site, which is not a bad thing on its own. The problem comes when the way they focus on the end-user to provide him with information, is an incorrect way.
-
Hi!
I was actually thinking about this based on experience while I was photographing with a very good friend on Iceland. My friend was using a Canon 20D and myself a Konica-Minolta 7D. I played around with both my friends pictures and I felt pretty shure I could pull a lot more shadow detail from my pictures than from my friends. It may be be for different causes. I felt especially that the "fill light" was much more effective on KM7D than on the 20D.
Best regards
Erik
What I find significant, Bernard, is that dpreview's figure of 12.6 stops for the A900 is a 'best case' scenario which involves extreme adjustments to RAW images in ACR, such as -2.65 EV and a -50 contrast.
I can find no comparable ACR adjustments in relation to other similar cameras that the A900 is compared with. What we have for other cameras are DR figures in relation to the default ACR settings and in relation to 'auto' ACR settings. The figure of 12.6 stops I therefore find misleading. The ACR default settings provide the worst DR figure; the ACR 'auto' settings provide a significantly better figure, and the best DR outcome results from the extreme settings mentioned above which have not been applied to any other camera in the review.
The only DR comparisons I can find in the review relate to in-camera jpegs, and those are quite surprising. Whilst the A900 excels in jpeg mode, having 1.2 stops greater DR than the 5D and 0.8 stops greater DR than the 1Ds3, the 5D is shown as having 0.4 stops greater DR than the D700.
One has to be clear when one is comparing apples with apples. The 12.6 figure for the A900 appears to be an orange.
-
Actually checking the 12.6 stops claim should be fairly easy since dcraw supports the A900 - as i understand things it is just a matter of taking a black frame and a white frame (i.e. a completely overexposed image), extracting the raw, unmodified values (dcraw -D - although it might make sense to determine DR separately for r, g and b) and determining the ratio of the white frame saturation value and the black frame noise.
In the light of Michaels latest article it might also be interesting to see what DR current top P&S offer in comparison - i am doubtful that they can reach comparable levels but who knows. This would be another test case for Bernard's high iso noise/DR theory - from a completely different side (P&S traditionally offering a very poor high iso performance). The G10 is not supported by dcraw yet but the P6000 is.
-
In the light of Michaels latest article it might also be interesting to see what DR current top P&S offer in comparison - i am doubtful that they can reach comparable levels but who knows.
You mean like this?
http://www.naturescapes.net/phpBB3/viewtop...=a&start=22 (http://www.naturescapes.net/phpBB3/viewtopic.php?f=2&t=141021&st=0&sk=t&sd=a&start=22)
-
In fact a pure binning strategy by averaging 2x2 pixels into 1 final pixel, which according to Emil's statistics doubles the SNR, ...
BR
Emil Martinec has good information, however, unfortunately like many others he also does not consider that there is a difference between noise reduction when a number of images are averaged together and the noise reduction resulting when image is resized. I could have read his pages wrongly. However, reading his pages I could sense it as an underlying assumption as quoted from his website below:
"Bottom line: At the cost of having half the linear resolution, the superpixel made by binning together a 2x2 block of pixels has twice the signal-to-noise ratio,"
"If one downsamples an image properly, one decreases the resolution, and noise decreases in proportion to the linear change in image size.";
and, additionally, Emil does not present any analysis regarding the differential rate of the signal change for SNR formulation, etc. For e.g., best SNR in simple average-based downsampling is related to second derivative of signal intensity.
The problem is that the assumption is that the mean square error used in formulation of SNR is only dependent upon only on noise reduction, where as, it has an additional dependency on the signal degradation due to smoothing, which is frequently ignored (the bias term). In the worst case just consider, that at each pixel position you average all pixels in an image and you end up with a flat, constant image, where each pixel has the same value. That is pathetic SNR.
Perhaps the easiest to way to see that is the mean square error in such estimation is given as:
error = (bias).^2 + (variance of noise)
When you average images of the same scene to get a cleaner image, then though image intensity is varying at each pixel, however, the average is in the time domain and each pixel can be considered to have a true value being estimated, though it is different for each pixel, of course. In this case, since average is an unbiased estimator, the bias is zero, and the variance of the noise decreases with an increase in the number of images being averaged, therefore SNR increases as a measure of the square root of the number of images.
However, when image is resized by averaging neighboring pixels, the average is in the spatial domain. Even in the absence of noise, the signal is not constant in neighboring pixels (unfortunately Emil chose a degenerate example of adding noise to constant image on his website), therefore, the bias term is not zero. Though noise variance goes down by averaging a larger number of pixels, on the other hand the bias increases as a square.
There is an optimal number of pixels to be averaged to get best SNR, however, that varies at each pixel position, implying at each pixel position a different number of pixels must be averaged together to get the same max SNR at each location!. The reason that happens is the differential rate of change of signal and noise determine the optimal number of pixels to be averaged together and these parameters change at each location.
It is easy to figure out when SNR would be less in resizing based upon averaging. Suppose you take an image and average each 2x2 pixels into 1 pixel, then you get an image 1/4 in size of the original image. In this case, spatial pixel averaging results in good noise reduction and many people erroneously conclude that the process could be extended to averaging even larger number of pixels, i.e., 4x4->1, 8x8->1, 16x16->1, etc. But there is a problem, though noise is being reduced, the image is also becoming more and more blurry. And, if you keep on doing that, then you are going to end up with a constant flat image, and terrible SNR.
On the other hand, it does not matter how many images you average in time to get a cleaner image, if it is considered that all those images are of the same static shot where each pixel had a fixed true value, but was corrupted by noise (hopefully zero mean). Averaging a constant number is the same as the constant number, so the original signal is *not* blurred (bias is zero), however, the noise is reduced significantly, and the SNR improves.
-
You mean like this?
http://www.naturescapes.net/phpBB3/viewtop...=a&start=22 (http://www.naturescapes.net/phpBB3/viewtopic.php?f=2&t=141021&st=0&sk=t&sd=a&start=22)
Yes, thanks. This is probably quite good for a sensor of this size.
-
stuff about noise
I think you may be talking about noise reduction, whereas the quoted discussion in the article has in mind the noise itself. Photon noise (the only signal dependent noise of consequence in this context) has variance equal to signal, and when binning pixels, the variance of the binned pixels equals the total signal, which is the sum of the individual signals. This is a basic property of Poisson statistics -- the sum of the variances is the variance of the sum.
The thrust of the argument in the article was to counteract the mistaken impression that many have, that image noise at the pixel level has some sort of absolute significance. Rather, any individual pixel of a group of binned pixels has only a fraction of the binned signal, and therefore lower S/N ratio than the aggregate. This is true regardless of whether the signal has content at high spatial frequencies. The upshot is that two cameras of given sensor size, one with pixel pitch half of the other, will have the same noise at any given spatial scale that both can resolve, even though the one with the finer pixel pitch has higher characteristic noise at its Nyquist frequency than the one with coarser pixel pitch has at its Nyquist frequency.
Your arguments about non-uniform signal have to do with one's ability to estimate how much variance an individual pixel of an individual sample image has from the ensemble average. That is relevant to noise reduction, but it is not relevant to how much noise is present, or how that noise aggregates to coarser scales.
BTW, a "flat, constant image, where each pixel has the same value" is not "pathetic" SNR, it is infinite SNR.
-
Bill,
ACR is what I use to process my RAW images, and I believe it's what most photographers use. What Dpreview does is what I do, and many others, except I try to avoid pushing exposure to the exteme right because I'm aware that ACR's attempt to reconstruct color data is sometimes unsatisfactory. An obvious example is the shift towards green in skies where the blue channel is clipped.
Ray,
I use ACR also. However, if you do not look at the raw file, you do not even know if you have exposed the image properly. For example, in the image which I posted as an example, the raw histogram showed underexposure of the green channel by 2/3 EV. Looking back at the metadata for the image, I see that the camera was on autoexposure and I gave +0.3 EV of exposure compensation to give a reasonable ETTR exposure as judged by the camera histogram. The ACR histogram demonstrates slight clipping of highlights. ACR was set to defaults.
[attachment=9156:BirdsACR_ScrCap.png]
ACR uses a baseline exposure offset of +0.5 EV for the D3 and one should use an exposure compensation of -0.5 EV with ACR in order to get an accurate histogram.
Bill
-
I think you may be talking about noise reduction, whereas the quoted discussion in the article has in mind the noise itself.
Yes, this is true, and I agree with that. And that is the basic point of argument. I realized this, that is why I wrote that the signal blurring/smoothing is not incorporated in typical SNR measures. However, consider for e.g., a valid SNR measure such as PSNR which is dependent upon mean square error (that is why I went into resolving mean square error as bias and variance on estimation noise, which results in the estimator because of the presence of the physical noise), and such measures of SNR will let me include the spatial blur degradation.
Also please note that even with a fixed resize factor of, say original to 1/4 the size of original, I am considering a more general problem than 2x2->1 spatial pixel averaging. Because, we can still have the size reduced by 1/4 by using a larger window size of samples than 2x2 that we consider for averaging. We want to determine what size of window works best for SNR formulations that include signal blur degradation measures resulting from a fixed image size reduction but variable size of window.
BTW, a "flat, constant image, where each pixel has the same value" is not "pathetic" SNR, it is infinite SNR.
Depends upon how you are calculating SNR. Since, I want to include the spatial blur degradation measure then the error term will increase as more and more samples are averaged because of bias the term and that lowers the measure of SNR as it is defined in terms of error.
-
Nonsense like this is why I ignore your posts, but since Emil responded then I see you have thrown more garbage into the discussion that should be responded to.
The reason why I generally don't like responding to you, Tony, is because you come across as a very rude person.
-
Ray, DPreview will never be able to compare cameras using the same tool and parameters unless they repeat all the tests for the old cameras with the last version software everytime a new camera appears in the market, and we know they won't do that. This is a flaw in the method, and you agree.
Guillermo,
I'm not sure whether I would call it a flaw in their methodology. Perhaps laziness would be a better word, or to be kind, a time constraint on the testing procedure. The fact is, dpreview never compare a new model of camera with all the older models of different brands. They are very selective, choosing only similar models with similar pixel count and usually of the same format. But they are not inflexible in this regard. In the current review of the A900 they have sensibly included the cameras from which many readers will be upgrading, should they buy an A900.
What I criticise them for is not taking the extra time to repeat the DR tests of RAW images from the other cameras included in the review. If they had the time, they might also examine how ACR handles the reconstruction of color data in clipped channels and how that varies with camera model.
A standardised DR test procedure which excludes or bypasses the most commonly used RAW converter would not necessarily be a better guide for the photographer or camera buyer than Dpreview's current DR testing procedure. We could have a situation where camera 'A' has a slightly higher DR than camera 'B', according to your hypothetical objective testing procedure, but in practice camera 'B' produces slightly better highlight and shadow detail than camera 'A' when its RAW images are processed in ACR, the converter which most people use.
-
The reason why I generally don't like responding to you, Tony, is because you come across as a very rude person.
Generally you might as well not respond to me, because I have you on my "Ignore User" list and am not paying attention to what you write since it is almost always misinformed and centered on your own biases and needs. However, your posts seep over to other's replies when they quote you and you have a tendency to dominate many of the threads you participate in (for instance, in the A900 Update thread started by Nick Rains as a follow-up to his A900 review, where you posted fully a third of the replies). I think you should consider reading more and writing less, you might actually learn something that way; you could ask questions instead of making assertions along the lines of if a camera doesn't work as well with ACR as another camera does, then that camera is necessarily not as good as the camera that works better with ACR.
-
DPR's tests are often a source of amusement. Apparently they don't understand that it is impossible for a linear medium like RAW to encode more stops of DR than there are bits in the encoding (12 for the A900).
Or a lemon
That's a good point. I'd forgotten that the A900 doesn't have 14 bit encoding like some of its competitors. One might also wonder how a camera with 12 bit encoding can deliver greater DR than another camera, such as the 1Ds3, which has 14 bit encoding.
Nevertheless, the fact that 12.6 stops might be a bit of an exaggeration is not significant in paractical terms, provided the DR of the other cameras with which the A900 is compared, is exaggerated to the same degree.
There is always a subjective element to the meaningfulness of DR figures relating to captured images, as opposed to the engineering DR specification of the sensor.
I once tested the DR of my 5D at 11.6 stops using Jonathan Wienke's DR target which he posted on the forum. A long argument followed between Jonathan and John Sheehy regarding the meaningfulness of the faint detail in the shadows. From my perspective it doesn't really matter whether the DR of the 5D is actually 11.6 stops, 10.3 stops or 9.5 stops. Where there's a subjective element, one can argue till the cows come home without resolution. The essential point is, how does the DR compare (in real world prints and images on monitor) with other cameras which I might be thinking of buying, and other cameras which I already use. Is it greater or less, and by a degree which is worth bothering with?
The A900 appears to be a camera which has a slight edge in DR and resolution at base ISO, compared with the 1Ds3. How significant that edge is, I simply don't know. I've not seen any 100% crops showing the amount of greater detail in either shadows or highlights that the claimed addtional DR should provide.
-
Ray,
I use ACR also. However, if you do not look at the raw file, you do not even know if you have exposed the image properly. For example, in the image which I posted as an example, the raw histogram showed underexposure of the green channel by 2/3 EV. Looking back at the metadata for the image, I see that the camera was on autoexposure and I gave +0.3 EV of exposure compensation to give a reasonable ETTR exposure as judged by the camera histogram. The ACR histogram demonstrates slight clipping of highlights. ACR was set to defaults.
[attachment=9156:BirdsACR_ScrCap.png]
ACR uses a baseline exposure offset of +0.5 EV for the D3 and one should use an exposure compensation of -0.5 EV with ACR in order to get an accurate histogram.
Bill
Bill,
Achieving a full ETTR without clipping any of the channels is a separate issue, surely. Whatever the DR of the camera, those problems remain. If one is using ACR to compare the DR of various cameras, then obviously one must ensure that the images being compared have the same amount of detail at some reference point, such as highlights, after all the stops have been pulled out to get the best result in ACR.
In my experience, there comes a point where no further detail can be extracted as one moves the EV slider into negative territory, for example. That point might be -2.65 EV for one camera in relation to one particular scene and -2 EV in relation to another camera and/or another scene.
The point that such a process is not completely objective and standardised is understood. What procedure do you have in mind that would be more helpful for people who use ACR to process their RAW images?
-
High ISO noise and low ISO DR need not be closely related. The noise floor at low ISO in CMOS DSLR's is controlled by the noise in the ISO amplifier and A/D converter, while the noise floor at high ISO is controlled by the noise of the sensor array. Different components, different noise characteristics, and the one that effects high ISO midtone and shadow noise doesn't enter into the determination of low ISO DR.
Emil's explanation clarifies the basis for the findings Roger Clark's (http://www.clarkvision.com/imagedetail/digital.sensor.performance.summary/#dynamic_range) figure 5, where he shows that the DR of large and small pixel cameras may not differ much at base ISO, whereas the large pixel cameras have a marked advantage at high ISOs. Previously, I had trouble in understanding that figure. Roger's have a wealth of information and are well worth a look. His concept of unity gain still seems a bit confusing.
Bill
-
Bill,
Achieving a full ETTR without clipping any of the channels is a separate issue, surely. Whatever the DR of the camera, those problems remain. If one is using ACR to compare the DR of various cameras, then obviously one must ensure that the images being compared have the same amount of detail at some reference point, such as highlights, after all the stops have been pulled out to get the best result in ACR.
In my experience, there comes a point where no further detail can be extracted as one moves the EV slider into negative territory, for example. That point might be -2.65 EV for one camera in relation to one particular scene and -2 EV in relation to another camera and/or another scene.
The point that such a process is not completely objective and standardised is understood. What procedure do you have in mind that would be more helpful for people who use ACR to process their RAW images?
The main point is that to expose to the right, one needs to determine at what exposure channel clipping begins. Most photographers use the camera histogram or the blinking highlights. However, on many cameras, these are a bit conservative and may indicate clipping before it actually occurs. The best way to determine the behavior of these tools for your camera is to bracket exposures and compare the results of the exposure tool to the contents of the raw file. The best tool that I have found for this process is Rawnalize (http://www.cryptobola.com/PhotoBola/Rawnalyze.htm). You can then determine how much headroom you have when the blinking highlights or clipped histogram appear on the camera display. Similar considerations apply to the histogram in ACR. The exposure offset that ACR uses for your camera may be determined by looking at a DNG conversion of the raw file. The offset of +0.5 EV for the D3 can cause the appearance of overexposure when none exists. Some ISO standards allow 0.5 EV of headroom for the highlights.
One must be aware that the camera histogram is derived from the JPEG preview of the raw file and is affected to some extent by the camera picture control settings. The camera black and white histogram is usually a luminance histogram and is mainly sensitive to green and may not show red or blue clipping. The camera RGB histogram (if available) will show clipping of the color channels after white balance has been applied, but clipping may be indicated for the red and blue channels when these channels are actually intact in the raw file. Again, one must look at the raw file.
Bill
-
Emil's explanation clarifies the basis for the findings Roger Clark's (http://www.clarkvision.com/imagedetail/digital.sensor.performance.summary/#dynamic_range) figure 5, where he shows that the DR of large and small pixel cameras may not differ much at base ISO, whereas the large pixel cameras have a marked advantage at high ISOs. Previously, I had trouble in understanding that figure. Roger's have a wealth of information and are well worth a look. His concept of unity gain still seems a bit confusing.
Bill
Actually, what you are seeing in this graph are several things. First, the fact that both CCD cameras (Canon S70 and Nikon D200) have DR that drops nearly a stop with each stop increase in ISO; this is because the read noise floor in photoelectrons is roughly independent of ISO for typical CCD designs, while the increased amplification loses a stop of headroom for each doubling of ISO. Second, the intial flatness of the DR curve for CMOS is due to the different sources of the noise floor at low and high ISO with current designs. The fact that the amplifier is the main source of the noise at low ISO limits the DR of current CMOS implementations; a cleaner amplifier/ADC would extend the low ISO DR of Canon/Nikon CMOS cameras by as much as two stops, as I showed in a thread I started a couple of months back. Third, DR of pixels is correlated to pixel size. Dynamic range is scale dependent; the proper scaling multiplies pixel DR by the square root of the number of megapixels to obtain a figure of merit that can be compared equitably between different formats, and pixel counts within a given format. In this sense, Roger's graph is misleading; if the D3 and 1Ds3 were plotted together in such a graph, the large pixel D3 would come out substantially ahead, but when one properly compares the aforementioned DR figure of merit, the two are quite close. The S70 is a poor performer because it has a tiny sensor; that necessitates small pixels in order to have decent resolution. But Roger would have you believe that large sensors with small pixels have as little DR as the S70, where the truth is that the performance differences are much smaller (the big pixels do have an advantage but not nearly as much as indicated in Roger's graph).
-
The main point is that to expose to the right, one needs to determine at what exposure channel clipping begins. Most photographers use the camera histogram or the blinking highlights. However, on many cameras, these are a bit conservative and may indicate clipping before it actually occurs. The best way to determine the behavior of these tools for your camera is to bracket exposures and compare the results of the exposure tool to the contents of the raw file. The best tool that I have found for this process is Rawnalize (http://www.cryptobola.com/PhotoBola/Rawnalyze.htm). You can then determine how much headroom you have when the blinking highlights or clipped histogram appear on the camera display. Similar considerations apply to the histogram in ACR. The exposure offset that ACR uses for your camera may be determined by looking at a DNG conversion of the raw file. The offset of +0.5 EV for the D3 can cause the appearance of overexposure when none exists. Some ISO standards allow 0.5 EV of headroom for the highlights.
One must be aware that the camera histogram is derived from the JPEG preview of the raw file and is affected to some extent by the camera picture control settings. The camera black and white histogram is usually a luminance histogram and is mainly sensitive to green and may not show red or blue clipping. The camera RGB histogram (if available) will show clipping of the color channels after white balance has been applied, but clipping may be indicated for the red and blue channels when these channels are actually intact in the raw file. Again, one must look at the raw file.
Bill
Bill,
I understand what you are saying. These are issues which all users have to sort out one way or another if they shoot RAW and are concerned about maximising the DR capability of their camera. I've had the jpeg picture styles in my 5D set at minimum contrast, saturation and sharpness for at least the past couple of years now, ever since I realised the camera's histogram and blinking review was giving me a false reading regarding exposure in RAW mode, which I always use.
The fact is, if I want to compare the DR of my own cameras using ACR as my converter, I don't have a problem. I just recently compared the DR of my new 50D with that of my old 5D. I was curious as to whether or not the cropped format really did have a shutter speed advantage resulting from its greater DoF at any given aperture and the option of using a wider aperture. I found that it didn't. Images from the 5D at F7.1 and ISO 250-320 were just as clean as images from the 50D at F4 and ISO 100, or as close as matters. I bracketed all shots in 1/3rd stop increments, compared ETTR images with equal highlight detail, then examined the shadows for detail and noise. (Same scene, same FoV and same lighting conditions, of course).
Is such a comparison invalid? If so, why? Does it matter if the pairs of images I was comparing, after applying a negative EV adjustment in ACR, -50 contrast and linear tone curve etc, had slight clipping of highlights according to Rawnalyze (for example), despite the fact that they didn't look clipped in ACR? I can't see that it does matter, provided that both images appear to contain the same amount of highlight detail after the same ACR adjustments.
-
Is such a comparison invalid? If so, why? Does it matter if the pairs of images I was comparing, after applying a negative EV adjustment in ACR, -50 contrast and linear tone curve etc, had slight clipping of highlights according to Rawnalyze (for example), despite the fact that they didn't look clipped in ACR? I can't see that it does matter, provided that both images appear to contain the same amount of highlight detail after the same ACR adjustments.
A contrast curve affects the quarter and three quarter tones but does not change the black point or the white point as shown below.
[attachment=9164:contrastCurve.gif]
With ACR, the default tone curve clips the black point rather strongly. If you set the black point to zero in the default tone curve and leave the other settings unchanged, the resulting tone curve does not differ that much from the linear tone curve in either the blacks or whites. However, the quarter tones are brighter and the slope of the curve is steeper near the midtones with the default curve. The camera histogram may behave slightly differently, but the tone curve on my D200 or D2 does not significantly affect the appearance of the highlights in the histogram, so I normally leave the camera tone curve to normal to get a better preview of the image. Others use low contrast.
[attachment=9162:NikonGraph.png]
If you bracket the images and use negative exposure correction to determine when additional highlight detail is no longer recovered, you may achieve the same end result. However, without looking at the raw files, you don't know when clipping has occurred and has been at least partially corrected by highlight recovery. In most instances, I prefer not to clip the highlights with the D3 since this camera has fairly good shadow detail. However, if the dynamic range of the scene is large and the camera histogram demonstrates clipping in both the shadows and highlights, some bracketing may be advisable and one could consider using mulitple exposures and HDR if the image is static.
Bill
-
Actually, what you are seeing in this graph are several things.
Thanks, very informative.
Bill
-
If you bracket the images and use negative exposure correction to determine when additional highlight detail is no longer recovered, you may achieve the same end result. However, without looking at the raw files, you don't know when clipping has occurred and has been at least partially corrected by highlight recovery. In most instances, I prefer not to clip the highlights with the D3 since this camera has fairly good shadow detail. However, if the dynamic range of the scene is large and the camera histogram demonstrates clipping in both the shadows and highlights, some bracketing may be advisable and one could consider using mulitple exposures and HDR if the image is static.
Bill
The question I'm asking is, does it really matter if, technically, highlighted areas of an image are clipped, but don't look clipped because ACR has done some clever reconstruction work? Does such a situation invalidate any DR comparisons?
Images are all about appearance. If ACR is my default converter, then the characteristics of all my images will be influenced to some degree by the characteristics of ACR. If we have a situation where camera 'A' appears to have a greater DR than camera 'B' in ACR, but the same camera 'B' appears to have a greater DR than the same camera 'A' when DPP is used as the converter (for example), then we have a problem. If this is the sort of thing that actually happens amongst different converters in a significant way, above the extreme pixel-peeping level, then that might invalidate any DR comparisons based on results from one particular converter.
Whenever I've compared different converters, I've always found that ACR is at least the equal of the other converters, and better than some, with regard to shadow detail and highlight recovery.
It's understood if you are going to compare shadow detail in images, then you should have contrast settings at a minimum in ACR, including 'black' at zero and the tone curve at 'linear'.
-
That's a good point. I'd forgotten that the A900 doesn't have 14 bit encoding like some of its competitors. One might also wonder how a camera with 12 bit encoding can deliver greater DR than another camera, such as the 1Ds3, which has 14 bit encoding.
Nevertheless, the fact that 12.6 stops might be a bit of an exaggeration is not significant in paractical terms, provided the DR of the other cameras with which the A900 is compared, is exaggerated to the same degree.
Hi Ray, the problem is not that they give a 12-bit camera a higher DR than a 14-bit camera. In fact, that's perfectly possible since DR is not bitdepth-limited but noise-limited, and the 14-bit camera could have a worse SNR than the 12-bit.
The problem is that a 12-bit linear encoding can _NEVER_ encode (I am not talking about noise here, just about encoding) more than 12 f-stops of DR. That's why anyone looking at a table where a 12-bit camera is given more than 12 f-stops of DR should not think of any exageration, but simply of wrong information.
Ideally the following levels would be encoded in a 12-bit RAW file where the whole range 0..4095 were used:
0EV: 2048 levels, 2048..4095
-1EV: 1024 levels, 1024..2047
-2EV: 512 levels, 512..1023
-3EV: 256 levels, 256..511
-4EV: 128 levels, 128..255
-5EV: 64 levels, 64..127
-6EV: 32 levels, 32..63
-7EV: 16 levels, 16..31
-8EV: 8 levels, 8..15
-9EV: 4 levels, 4..7
-10EV: 2 levels, 2..3
-11EV: 1 level, 1
There is no room for more than 12 f-stops, and in any case the lowest f-stops would be poorly represented (1 level, 2 levels, 4 levels,...)
BR
-
The problem is that a 12-bit linear encoding can _NEVER_ encode (I am not talking about noise here, just about encoding) more than 12 f-stops of DR. That's why anyone looking at a table where a 12-bit camera is given more than 12 f-stops of DR should not think of any exageration, but simply of wrong information.
What if the data is not encoded linearly?
-
What if the data is not encoded linearly?
Then it is possible (like in the M8), but the A900 uses a linear encoding as far as I know.
-
The problem is that a 12-bit linear encoding can _NEVER_ encode (I am not talking about noise here, just about encoding) more than 12 f-stops of DR. That's why anyone looking at a table where a 12-bit camera is given more than 12 f-stops of DR should not think of any exageration, but simply of wrong information.
Well, there's nothing new there, regarding wrong information. It happens all the time . I imagine not only the lowest stop would be poorly represented but the highest stop also.
Since useful DR is to some extent a subjective factor, then I suggest the precise figure doesn't matter. Even if you use a program like Rawnalize to ensure that absolutely nothing is clipped in the highlights, you still have to determine what level of image degradation in the shadows is acceptable and constitutes another stop of DR.
What I find significant and interesting is not that the DR of the A900 is 12.6 stop or 11.6 stops or whatever, but that it is claimed to be 0.8 stops better than the 1Ds3 at base ISO. It is the accuracy of this figure which should be questioned. The DR test of 1ds3 RAW images would have been carried out a few months ago with a previous version of ACR, possibly by a different person and possibly with different adjustments in ACR in the attempt to get the best outcome.
I'm suggesting that it's this procedure of not carrying out simultaneous tests which represents a flaw in the methodology. Nevertheless, it would be better if we could have an international standard for DR measurement, like the ISO standard for exposure.
I'm thinking of buying a Panasonic 50" plasma display which boasts a contrast ratio of 1,000,000:1. I wonder what method was used to calculate that .
-
I'm thinking of buying a Panasonic 50" plasma display which boasts a contrast ratio of 1,000,000:1. I wonder what method was used to calculate that .
That's nearly 20 f-stops and sounds quite exagerated. But you can check it with your camera. Just make 2 shots, one with a completely white display (i.e. maximum luminosity) and second with a black display (i.e. minimum luminosity, which in plasma and LCD is never pure black) in a completely dark room, making sure to obtain a good exposure in both cases. By comparing the resulting RAW values, and after correcting them by the difference in the exposure settings, you can trust your sensor's linearity to get the figure of contrast.
I did something similar on my windows, which are of that kind of glass that is seen as a mirror when looking from outside so I can walk naked without any neighbour noticing me . And I found out they eliminate 2/3 of the incident light!!. I deeply regret now having paid for such kind of windows since my place is not the best lighted one could expect.
-
My observations are:
1. the A900 raw is nonlinear (due to the lossy compression); perhaps it *can* create lossless raw files as well?
2. The noise at ISO 200 is about 0.4 EV better than the 1DsMkIII. However, it is possible that the A900 applies some smoothing to the raw data. The images I have DO show some degradation of very fine noise compared to the 1DsMkIII despite the higher resolution of the A900, but I would not state firmly, that a noise reduction took place. I guess Emil could make a Furier analysis in order to clarify this.
-
My observations are:
1. the A900 raw is nonlinear (due to the lossy compression); perhaps it *can* create lossless raw files as well?
2. The noise at ISO 200 is about 0.4 EV better than the 1DsMkIII. However, it is possible that the A900 applies some smoothing to the raw data. The images I have DO show some degradation of very fine noise compared to the 1DsMkIII despite the higher resolution of the A900, but I would not state firmly, that a noise reduction took place. I guess Emil could make a Furier analysis in order to clarify this.
It would depend on the nature of the compression. I wasn't aware that the A900 used lossy compression; are you saying the raw files don't use lossless compression algorithms? If so, what are they doing? That would certainly affect the analysis of raw data.
I can believe that the A900 is better than the 1Ds3 at low ISO; the D300 is better than the 40D at low ISO. This may be due to the column-parallel ADC's on the Sony CMOS sensor, which allows slower data rates and correspondingly lower noise. At high ISO, though, the amplifier/ADC noise is irrelevant, and the photosite readout noise dominates; there Canon is the leader in low noise designs.
-
I wasn't aware that the A900 used lossy compression; are you saying the raw files don't use lossless compression algorithms? If so, what are they doing?
Both the A700 and A900 offer the selection betwen 12bit, lossless data and a lossy, non-linear version. I did not know this from the A900 when I wrote my previous message; I just found one A900 lossless file. I have not implemented the decompression, I don't know the algorythm. I am converting the lossy Sony files in DNG for analysis.
That would certainly affect the analysis of raw data
I'm afraid it would; see the attached histograms. However, the lossless files (if you find some) should be suitable. The one I found is with ISO 1000, and it looks like manipulated for BOTH smoothing and to achieve this ISO from something else (see the attachment), thus it is probably not suitable for a Furier analysis.