Pages: 1 ... 6 7 [8] 9 10   Go Down

Author Topic: Nikon D7000 Dynamic Range  (Read 126688 times)

JR

  • Newbie
  • *
  • Offline Offline
  • Posts: 24
Re: 1% gray is 4-stops below normal midtones, in the deep shadows
« Reply #140 on: December 10, 2010, 06:47:04 am »


Thank you for that link. I have read some of their "insight" articles but have not seen that one. Will read it through.

- John


I invite you to look at the following extract from their 'insights'  articles at
http://www.dxomark.com/index.php/en/Our-publications/DxOMark-Insights/Noise-characterization/Summary

Logged

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: 1% gray is 4-stops below normal midtones, in the deep shadows
« Reply #141 on: December 10, 2010, 11:10:57 am »

Hi,

Det är i Österrike nära Dachstein tror jag.

Hälsningar
Erik


http://www.silberkar.com/

Where is this, Erik? Var är detta?

Logged
Erik Kaffehr
 

BJL

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6600
my comment was on what 1% diffuse reflectance looks like
« Reply #142 on: December 10, 2010, 01:48:13 pm »

BJL,
What an excellent justification for not needing a better camera than a Point & Shoot, as regards SNR and dynamic range.
However, some of us, notwithstanding the reflectivity of charcoal, wish for lower noise and better detail in the shadows than a P&S or even an Olympus 4/3rds can provide.
And once again, your respond to my factual posting not by refuting my claims but by making personal, insulting insinuations about camera choice an such that have nothing to do with what I said. I was talking about how dark something of 1% gray (1% diffuse reflectance) looks, or how one is likely to want it to look in a displayed image, nothing more. Answer: very, very dark.

I will confine my replies to calm, factual responses like Erik's.
Logged

BJL

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6600
Re: Nikon D7000 Dynamic Range
« Reply #143 on: December 10, 2010, 02:54:58 pm »

Except that the white snowman may be in the sun, melting away, while that piece of charcoal may be in a deep shadow.
Indeed. I was addressing the issue of handling subjects of various albedos under similar levels of incident lighting when optimum low ISO speed can be used, and the basic conclusion is that the DR needs for that situation are modest. That is, the principal need for wide DR at base ISO speed is in scenes with high subject brightness range (SBR) due to wanting to place the midtones in one, relatively dimly lit, part of the scene while avoiding blown-out highlights in other far more brightly lit parts of the scene. For this what the camera needs is lots of "highlight headroom" above the light levels being received from the adequately exposed though dimly lit subject matter. Evidence from sources like DXo indicate that with the current state of the art of DSLR sensors, the way forward in handling high SBR at base ISO speed has become a matter of improving photo-electron counting headroom capacity, not further reduction of dark noise. (It is a different story with light-starved "high ISO" situations like fast action photography, where one might have to accept exposures so low that the final images have visible shot noise, and then dark noise can make a visible contribution.)
 
To rephrase my argument from earlier, to control shot noise in the dimly lit midtones, the camera needs to gather enough light from that dimly lit region, which requires detection of about 500 or more photons from the "18% gray" subject matter within that dimly lit region, and so 100 photons or more even from charcoal (4%) in that same dim light, and thus over 10e- of shot noise from anything in that dim lighting level. Thus even the darkest (meaning least reflective) parts in that dimly lit region will have shot noise that overwhelms the 2 or 3 e- of dark noise that the best current DSLR sensors reportedly have. (The only alternative is exposure levels and photo-electron counts so low that one has highly visible shot noise in the midtones, at which case there is little point worrying about visible dark noise in the charcoal.) Any part of the scene sending out less light than charcoal in the most dimly lit regions of interest can surely be rendered as black, and will be on any print once the midtones are appropriately placed.

Thus, once exposure is adequate to control shot noise in these more dimly lit parts of the scene, the remaining challenge is entirely in the realm of "highlight headroom": being able to handle the far higher photon counts coming from other, far more brightly lit parts of the scene, like sunlit clouds or snow visible through the open window of a heavily shadowed church interior, without ND filters, and with a single exposure rather than relying on HDR blending or such. The challenge going forward for photographing such scenes of unusually high subject brightness range seems to be mostly or entirely in photon counting headroom, and progress there will likely have little or nothing to do with further reduction of dark noise. Adding one stop of DR by just halving the dark noise level of the D7000's sensor to about 1e- would be useless in this situation, or any low ISO speed situation I can imagine.

By the way, I believe that the far smaller photosites of good "compact digicam" sensors have had dark noise levels of 2e- or less for some years, so their DR limitations at base ISO speed have already been dominated by the limited electron well capacity of their small photosites.
Logged

BJL

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6600
Re: 1% gray is 4-stops below normal midtones, in the deep shadows
« Reply #144 on: December 10, 2010, 03:06:23 pm »

The image below illustrates real world outdoor contrast range. It was taken using a Konica Minolta 7D in 2005. The image was shot in raw and processed in Lightroom. Better processing would certainly be possible.

Best regards
Erik
By the way, that is a nice shot, and good handling of a situation quite typical of the ones where I most often struggle with high subject brightness range. (Aside: I prefer to use established photographic terms like SBR rather than electrical engineering terms like DR that are currently fashionable in digital photography.) And with scenes like that, getting enough light from the shaded regions to avoid noise is not my main problem, it is blowing out the sky!
Logged

JR

  • Newbie
  • *
  • Offline Offline
  • Posts: 24
Re: 1% gray is 4-stops below normal midtones, in the deep shadows
« Reply #145 on: December 11, 2010, 08:32:08 am »


Det er et fint bilde. Jeg tenkte at du kanskje hadde tatt deg en tur over grensen til Norge. Vi har jo mye fjell og daler her.


Hi,

Det är i Österrike nära Dachstein tror jag.

Hälsningar
Erik


http://www.silberkar.com/

Logged

JR

  • Newbie
  • *
  • Offline Offline
  • Posts: 24
Re: 1% gray is 4-stops below normal midtones, in the deep shadows
« Reply #146 on: December 11, 2010, 08:38:32 am »


I think we have to ask DxoMark if we want to know exactly where the deep shadows begin. It could be that you are right, that 0.1% gray is the answer. I think only Dxomark knows the answer. Another question is of course how do they define "deep shadows"?

In general 1% gray is considered very deep shadows, as BJL wrote, but I read the link you provided to their "Insight" articles and the logarithmic scale can only be interpreted by Dxomark themselves. Interesting reading.

What is clear however is that the D7000 has excellent SNR also at 0.1% gray. Actually a bit better than the D700. And that is quite an achievement.




...DXOMark's own explanations of the meaning of their results...

...the SNR also has three regimes...

Logged

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: 1% gray is 4-stops below normal midtones, in the deep shadows
« Reply #147 on: December 11, 2010, 09:22:47 am »

Nej, tyvärr!

Kanske 2011 ?

MVH
Erik

Det er et fint bilde. Jeg tenkte at du kanskje hadde tatt deg en tur over grensen til Norge. Vi har jo mye fjell og daler her.


Logged
Erik Kaffehr
 

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
Re: 1% gray is 4-stops below normal midtones, in the deep shadows
« Reply #148 on: December 11, 2010, 12:05:04 pm »

I think we have to ask DxoMark if we want to know exactly where the deep shadows begin. It could be that you are right, that 0.1% gray is the answer. I think only Dxomark knows the answer. Another question is of course how do they define "deep shadows"?

In general 1% gray is considered very deep shadows, as BJL wrote, but I read the link you provided to their "Insight" articles and the logarithmic scale can only be interpreted by Dxomark themselves. Interesting reading.

What is clear however is that the D7000 has excellent SNR also at 0.1% gray. Actually a bit better than the D700. And that is quite an achievement.

I don't think you don't need to ask anyone where deep shadows begin. Use your eyes. I've never found any difficulty in determining what is in shade, what's not in shade and what is in deep shade.

There seems to be confusion here between the range of input signals (from the scene being photographed) and the range of output signals as displayed on print or monitor.

A 1% grey patch is certainly a dark shade of grey and might be quite appropriate for the display of deep shadows on a print. However, in a high contrast, real-world scene where a white cloud is 100% grey, 1% on a logarithmic scale is not particularly dark.

At the 1% point on the DXO graph of the D7000, the signal has dropped only 15dB from its full-well capacity. The dynamic range of real-world scenes in sunny conditions can be as high as 100dB, even higher.

You think just 15dB down, at the 85dB point of a very bright and high-contrast scene, we're into deep shadows. Pull the other leg.  ;D

Even if we take a less extreme example, say a scene of 60dB dynamic range, do you think we're into deep shade at the 45dB point on the scale?
Logged

JR

  • Newbie
  • *
  • Offline Offline
  • Posts: 24
Re: Nikon D7000 Dynamic Range
« Reply #149 on: December 11, 2010, 01:12:53 pm »



G´Day Ray,

I think we are talking about two different things. I am talking about the D7000 sensor, Dxomark and where to place the deep shadows.  :)  Not a real world situation with a 100dB high contrast scene.

At 0.1% grey ( logarithmic scale ) the SNR is ca 16dB for the D7000. The 50D´s SNR is just above 6dB. That is very low. What do you think? Would you consider 6dB deep shadows? According to this link http://www.dxomark.com/index.php/en/Learn-more/DxOMark-database/Measurements/Noise  they say: the lowest gray luminance makes sense only if it is not drowned by noise, thus this lower boundary is defined as the gray luminance for which the SNR is larger than 1. 6dB= 1EV  We are close to hitting the floor.

Scene dependent DR and SNR is another subject and of course I agree you have to trust your eyes when defining the shadow region as every scene can be different. As long as I keep my monitor calibrated and my reading glasses clean I usually do.  :)

- John
Logged

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
Re: Nikon D7000 Dynamic Range
« Reply #150 on: December 11, 2010, 06:39:21 pm »


G´Day Ray,

I think we are talking about two different things. I am talking about the D7000 sensor, Dxomark and where to place the deep shadows.  :)  Not a real world situation with a 100dB high contrast scene.

At 0.1% grey ( logarithmic scale ) the SNR is ca 16dB for the D7000. The 50D´s SNR is just above 6dB. That is very low. What do you think? Would you consider 6dB deep shadows? According to this link http://www.dxomark.com/index.php/en/Learn-more/DxOMark-database/Measurements/Noise  they say: the lowest gray luminance makes sense only if it is not drowned by noise, thus this lower boundary is defined as the gray luminance for which the SNR is larger than 1. 6dB= 1EV  We are close to hitting the floor.

Scene dependent DR and SNR is another subject and of course I agree you have to trust your eyes when defining the shadow region as every scene can be different. As long as I keep my monitor calibrated and my reading glasses clean I usually do.  :)

- John


John,
Current cameras are not capable of simultaneously recording meaningful detail in deep shadows and in bright highlights. It's why people use fill-flash where possible, or merge multiple exposures to HDR, or render such deep shadows black on the print.

We're both looking at the same graph, the Full SNR graph for the D7000 for example, and what I see on the x-axis is a representation of an input signal. What I see on the y-axis is the camera's response to that input signal.

At an input of 1% luminance on the log scale, the SNR of the camera has fallen by 15-18dB. At 100% luminance its DR is close to 14 EV. At an input of 1% grey its DR is still a very healthy 8 EV.

Unfortunately, those last couple of stops of DR are a bit noisy and useless, so 11 or 12 EV of DR for the D7000 would be a more reasonable estimate. At a 1% input level, the very limited D7000, which is incapable of capturing the full dynamic range of a very bright scene with deep shadows, has still got 5 or 6 stops of useable DR.

Yeah!! Yeah! Please give me a camera that still has 5 to 6 useable stops of DR when I record those deep shadows in a rainforest and the white rocks by the waterfall glistening in the sun.

Surely it's apparent from those DXO graphs, if an input signal of 1% grey really is representative of a deep shadow, then any P&S camera will suffice. As I mentioned before, even the Canon G12 has a healthy SNR of 19dB at a 1% input signal.

Perhaps part of the problem is that some of you guys who live in Northern Europe don't realise how bright the sun is in Australia. Perhaps most of your photos are taken on dull, cloudy days.  ;D

By the way, the shadows should always be placed at a point which does not result in a blow-out of the highlights, if the highlights are important. I really don't care much for landscapes with clouds that show no detail and skies that appear more cyan than blue.
« Last Edit: December 11, 2010, 07:22:16 pm by Ray »
Logged

bjanes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3387
Re: Nikon D7000 Dynamic Range
« Reply #151 on: December 11, 2010, 08:06:53 pm »

By the way, I believe that the far smaller photosites of good "compact digicam" sensors have had dark noise levels of 2e- or less for some years, so their DR limitations at base ISO speed have already been dominated by the limited electron well capacity of their small photosites.

That is true, but a read noise of 2e- with a small pixel has much more significance than a similar read noise for a large pixel when the electron count is converted to a 12 or 14 bit data number because of gain (electrons/data number). See Table 3 of Roger Clark's essay.

Regards,

Bill
Logged

JR

  • Newbie
  • *
  • Offline Offline
  • Posts: 24
Re: Nikon D7000 Dynamic Range
« Reply #152 on: December 12, 2010, 07:51:31 am »


...some of you guys who live in Northern Europe don't realise how bright the sun is in Australia. Perhaps most of your photos are taken on dull, cloudy days.  ;D


Sunshine? What is that?  :D

A friend of mine is an Aussie. He used to do a lot of bush-hiking when he was younger and he has told me how extremely bright it can be. And hot. Really hot. Not Europe-hot but down-under hot  :)

Dull, cloudy days are not a photographic problem in the northern hemisphere, but a long dark winter & short days with only a couple of hours of twilight is.

- John
Logged

JR

  • Newbie
  • *
  • Offline Offline
  • Posts: 24
Re: Nikon D7000 Dynamic Range
« Reply #153 on: December 12, 2010, 08:11:14 am »


Bill,

Conclusion from Roger Clark:
Current good quality sensors in digital cameras are photon noise limited. This means there is no possible improvement in performance for the high signal region (bright things in an image) except to increase quantum efficiency of the devices and/or the fractional active area for which the sensor converts photons to electrons (called the fill factor). As both of these properties are reasonably high already, there is limited room for improvement. And even if these properties were improved, there would still be a big difference between large and small pixels. Larger pixels have higher signal-to-noise ratio at all levels, but especially at low signal levels. The obvious improvement still possible would be to reduce the read noise, but that would likely improve large sensors also, thus large sensors with large pixels will always have an advantage.

What he is saying is interesting. We have not seen much of an improvement in capturing highlights. HTP, Active D-Lightning etc are only in-camera manipulations. The Fuji S5 Pro is the only camera using new technology in this area.

But we have seen  improvements in read-noise at low signal levels. The new Sony made sensor we are discussing here is an example. It will be interesting to see if this will benefit future full-frame sensors from Nikon or hopefully Canon?

- John


See Roger Clark's essay.

Regards,

Bill
Logged

bjanes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3387
Re: Nikon D7000 Dynamic Range
« Reply #154 on: December 12, 2010, 08:46:57 am »

Conclusion from Roger Clark:

And even if these properties were improved, there would still be a big difference between large and small pixels. Larger pixels have higher signal-to-noise ratio at all levels, but especially at low signal levels. The obvious improvement still possible would be to reduce the read noise, but that would likely improve large sensors also, thus large sensors with large pixels will always have an advantage. [/i]

But we have seen  improvements in read-noise at low signal levels. The new Sony made sensor we are discussing here is an example. It will be interesting to see if this will benefit future full-frame sensors from Nikon or hopefully Canon?

John,

I don't follow Canon that closely, but Nikon has already made considerable progress in improving the read noise for their full frame sensors. Look at the following DR plot for the D7000, D3x and D3 as determined by DXO. I choose to use the screen mode to look at per pixel performance. The D3 curve flattens to the left due to excessive read noise, whereas the D3x exhibits less flattening, and the D7000 practically none. For a tutorial on reading the DXO DR plot, see Emil Martinec.

Regards,

Bill
« Last Edit: December 12, 2010, 09:12:28 am by bjanes »
Logged

JR

  • Newbie
  • *
  • Offline Offline
  • Posts: 24
Re: Nikon D7000 Dynamic Range
« Reply #155 on: December 12, 2010, 09:11:03 am »


Bill,

I have not read Emil´s tutorial  yet but I will.

As a Canon user I have to agree that Nikon have made huge improvements in regards to lower the read noise. I don´t think Canon´s sensor performs bad, but I do think they have fallen behind a bit. If they don´t keep up with Nikon/Sony with their next releases in 2011 I think they will fall further behind in this respect.

- John

John,

I don't follow Canon that closely, but Nikon has already made considerable improvement in improving the read noise for their full frame sensors. Look at the following DR plot for the D7000, D3x and D3 as determined by DXO. I choose to use the screen mode to look at per pixel performance. The D3 curve flattens to the left due to excessive read noise, whereas the D3x exhibits less flattening, and the D7000 practically none. For a tutorial on reading the DXO DR plot, see Emil Martinec.

Regards,

Bill
Logged

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
Re: Nikon D7000 Dynamic Range
« Reply #156 on: December 12, 2010, 07:26:48 pm »


At 0.1% grey ( logarithmic scale ) the SNR is ca 16dB for the D7000. The 50D´s SNR is just above 6dB. That is very low. What do you think? Would you consider 6dB deep shadows?


John,
I think this point needs further clarification. What is deep shadow is what the eye perceives as deep shadow, bearing in mind that the aperture of the eye, the pupil, is constantly changing according to the changing levels of reflected light within the scene, as the eye peruses the scene.

If the eye had a fixed aperture like a camera does at the moment the shot is taken, we'd be in deep trouble. We'd be blinded by even moderately bright light, and even moderately dark shadows would appear impenetrably black.

So the answer is, a signal level which is only 6dB above the noise floor of the camera can be representative of a deep shadow in a scene, especially if the scene is of low contrast and low dynamic range. On the other hand, if the scene is of high dynamic range, a shadow which the eye perceives as normal shade where the light is more than sufficient to read a book easily, may appear in the camera's sensor, and consequently in the processed image, as a deep, noisy shadow with a 6dB SNR or less, which has to be rendered black.

Now, I know it's the case that the sun was last seen in Great Britain in the afternoon of the 23rd of August 1955 (well, let's not exaggerate. I think it might have been seen on a few occasions  since then), and I was informed recently by a tourist guide that St Petersburg gets on average about 65 sunny days per year, so I considered myself lucky that the sun was shining for 2 out of the 4 days I was there.

So, for the benefit all the sunshine-deprived people who might never have seen a 'high dynamic range' scene, I searched through some shots I'd taken with my 5D of the temples at Siem Riep in Cambodia, back in 2006, looking for an example with shade which was not at all deep, in which the light was perfectly adequate for reading a book, but which the camera simply cannot render as the eye saw it, because of the camera's limited dynamic range.

Here's that shot which is totally useless because of noisy shadows. The sky and clouds are okay though.

First the ACR Window demonstrating that the shot is a reasonable ETTR, followed by a crop of the sky showing good detail in the clouds.
Logged

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
Re: Nikon D7000 Dynamic Range
« Reply #157 on: December 12, 2010, 07:43:42 pm »

Let's now have a look at some shadows, not deep shadows, but merely areas in the shade under a tree.

Phwoar! What subtle detail and texture in that laterite stone. Can you see it?  ;D

I certainly could at the time, when I was there.
Logged

BJL

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6600
where deep shadows begin? depends on desired "tonal placement" in the print
« Reply #158 on: December 12, 2010, 10:37:12 pm »

I think we have to ask DxoMark if we want to know exactly where the deep shadows begin.
I think the answer depends on what in the Zone System is called your choice of "placement": at what brightness level you intend to print (or otherwise present) subject matter that is at a certain brightness level in the scene. Sometimes, there could be a shaded subject that is only 1% as bright as another, sunlit, part of the scene and you wish to place that shaded subject in the print as only a stop or so below middle gray ... at the cost of greatly reducing the relative brightness level at which the sunlit parts are placed on the print. Maybe ranges even more extreme than 100:1 can occur between, say, (a) shaded leaves or rocks that one wants to rendered light enough to show details, so not much more than one stop darker than middle gray, and (b) sunlit clouds or snow that one does not want rendered as sterile blank white, which on a print means not more than three stops lighter than middle gray. But slides and computer screens can go far brighter with the highlights than light reflected of a paper print.
Logged

thierrylegros396

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1947
Re: Nikon D7000 Dynamic Range
« Reply #159 on: December 13, 2010, 03:10:41 am »

To futher improve DR, may we imagine ISO25 based sensors ?!

Thierry
Logged
Pages: 1 ... 6 7 [8] 9 10   Go Up