More D3s high iso samples

[BernardLanguillier]

http://www.imaging-resource.com/PRODS/D3S/D3SA7.HTM

Cheers,
Bernard

[Daniel Browning]

Those out-of-camera JPEGs do not show the quality of the D3s raw files very well at all. Iliah Borg took those Imaging Resource raw files and did some quality raw conversions:

Camera JPEG:
http://208.43.255.194/PRODS/D3S/FULLRES/Yc...drig_102400.JPG

RPP:
http://www.pochtar.com/Ycircus_vidrig_102400.jpeg

Camera JPEG:
http://208.43.255.194/PRODS/D3S/FULLRES/Yc...proud_10000.JPG

RPP (2-stop nonlinear EC):
http://www.pochtar.com/Ycircus_proud_10000.jpeg

RPP (3-stop):
http://www.pochtar.com/Ycircus_proud_10000_3EV.jpeg

[Ray]

Maybe not bad, but not good either. I'm reminded of some of the underexposed test shots I took with my 5D a couple of years ago. Here's one that I consider to be an ISO 51,200 (ISO 3200 underexposed 4 stops). It's seems about roughly as noisy as the D3s RPP conversion of the ISO 102,400 shot, which would give the D3s a full stop better high-ISO performance than the 5D. The D3s shot appears a bit sharper, but that might be due to the that fact my shot was hand-held at just 1/25th sec.

The screen grab of the ACR window shows a +3 EV adjustment which might give the impression the shot is only 3 stops underexposed. However, my reasoning is a full ETTR exposure normally requires a -1 EV adjustment in ACR to get the histogram looking as it does in this shot. There are no values anywhere near 255 and the second image showing the levels histogram after conversion, indicates there's plenty of highlight headroom, so I'm convinced this shot is at least 4 stops underexposed at ISO 3200, but I'll listen to reason if anyone wants to persuade me otherwise   .

[attachment=17530:Canon_5D...R_window.jpg]  [attachment=17531:Canon_5D...onverted.jpg]

I know it's difficult and inconclusive to compare image quality using completely differenct scenes, but I prefer the 5D image in this case, probably because I'm closer to the subject. The D3s shot levels histogram shows a lot of black clipping. The slight clipping of the highlights is due to specral highlights which are not present in my shot.

[attachment=17532:RPP_conv..._102_400.jpg]

[Ray]

Is that lined wallpaper in the 5D shot to the subjects left and right?

No, it's the inside of a shop. I was just passing buy with 5D around my neck, at night, noticed this attractive lady sitting in front of her shop, and snapped her at F8. I was in high-ISO experimental mode (mood).

[Daniel Browning]

Maybe not bad, but not good either. I'm reminded of some of the underexposed test shots I took with my 5D a couple of years ago. Here's one that I consider to be an ISO 51,200 (ISO 3200 underexposed 4 stops).

It's not possible to compare ISO settings between different models this way. For one, the 5D meters quite a bit brighter than the D3 when both are at the same ISO setting (i.e. what the 5D calls ISO 3200 is less than 3200 on the D3). For two, different QE and analog gain would put the same signal at a different brightness level. Third, the color balance of the light may be more even in one than the other. Even the difference between 2900K and 3200K can make a huge impact on noise. Furthermore, Adobe applies hidden exposure compensation on cameras.

[Panopeeper]

Maybe not bad, but not good either. I'm reminded of some of the underexposed test shots I took with my 5D a couple of years ago. Here's one that I consider to be an ISO 51,200 (ISO 3200 underexposed 4 stops)

Ray, I don't know which number system you are using, but you capture of ACR display shows THREE stops increase; that makes 3.25 EV (BaselineExposure is +0.25), not FOUR.

Anyway, Rawnalyze shows that the 5D is a far cry from the D3S (is anyone surprized by that?). I attach two crops, made by Rawnalyze, i.e. no true rendering, not even WBd, at 50%; of course no noise reduction. The first one is the ISO 102400 shot from the D3S; the second one is from your 5D, shot with ISO 3200, displayed with +4 EV, i.e. this corresponds to ISO 51200.

[Ray]

It's not possible to compare ISO settings between different models this way. For one, the 5D meters quite a bit brighter than the D3 when both are at the same ISO setting (i.e. what the 5D calls ISO 3200 is less than 3200 on the D3). For two, different QE and analog gain would put the same signal at a different brightness level. Third, the color balance of the light may be more even in one than the other. Even the difference between 2900K and 3200K can make a huge impact on noise. Furthermore, Adobe applies hidden exposure compensation on cameras.

 
Exactly! But this is all we have at present; pretty awful and noisy shots from the Ds3 with claims that they look at least a stop, maybe 2 stops better than the D3.

The sensitivity differences between the D3 and 5D is not as great as you seem to think. According  DXOMark (what a wonderful site for the rational and objective amongst us   ) the D3 ISO 3200 is actually ISO 2512, and the 5D ISO 3200 is actually ISO 2710. That's much less than 1/4 of a stop difference, isn't it?

[Ray]

Ray, I don't know which number system you are using, but you capture of ACR display shows THREE stops increase; that makes 3.25 EV (BaselineExposure is +0.25), not FOUR.

Gabor,
The number system I'm using is, (1) a full ETTR requires approximately a -1 EC adjustment in ACR, (2) an underexposure of 1 stop requires no EC adjustment in ACR and no highlight recovery, (3) an underexposure of 2 stops requires a +1 EC adjustment in ACR. If you think this is incorrect, please explain why.

Anyway, Rawnalyze shows that the 5D is a far cry from the D3S (is anyone surprized by that?). I attach two crops, made by Rawnalyze, i.e. no true rendering, not even WBd, at 50%; of course no noise reduction. The first one is the ISO 102400 shot from the D3S; the second one is from your 5D, shot with ISO 3200, displayed with +4 EV, i.e. this corresponds to ISO 51200.

The D3 is a far cry from the 5D, but not at ultra-high ISOs. As regards DR, the big improvement occurs at ISOs lower than 1600. Here's the DXOMark comparison for the two cameras showing a 1/2 stop advantage to the D3 at ISO 3200.

My own tests indicate that above ISO 3200, that difference diminishes slightly, so at a simulated ISO 51,200 or 102,400 we have perhaps a 1/3rd stop difference between the two cameras.

However, it is true that the 5D has a tendency to show banding in the deepest shadows, even at ISO 100.

[attachment=17540:DXOMark_...mparison.jpg]

[Daniel Browning]

The sensitivity differences between the D3 and 5D is not as great as you seem to think. According  DXOMark (what a wonderful site for the rational and objective amongst us   ) the D3 ISO 3200 is actually ISO 2512, and the 5D ISO 3200 is actually ISO 2710. That's much less than 1/4 of a stop difference, isn't it?

I was wrong. Thank you for the correction.

I guess I've had my 5D2 long enough that the 5D classic is nothing but a distant memory.

[Panopeeper]

(1) a full ETTR requires approximately a -1 EC adjustment in ACR, (2) an underexposure of 1 stop requires no EC adjustment in ACR and no highlight recovery, (3) an underexposure of 2 stops requires a +1 EC adjustment in ACR

This is a strange way of determining the "underexposure" of a shot. Should we call it the Rayposure?

If you think this is incorrect, please explain why

It is not incorrect, it is absurd, for the required (why would it be required, btw?) exposure correction depends on the other adjustments (for  example Brightness = +50 is +1 exposure adjustment, except for the highlights), and on the illumination and on the colors of the scenery's highlights.

The D3 is a far cry from the 5D, but not at ultra-high ISOs

I thought the topic is the D3S.

My own tests indicate

Your own tests indicate nothing useful. The rendered ISO 102400 shot of D3S is crap. I don't know how you created it, but the non-rendered Rawnalyze display is better than that; thus the comparison is worthless. If you are interested on the outcome of the comparison, you should do it with a Nikon Capture X rendered image, compared to the DPP output of your 5D shot.

[Ray]

This is a strange way of determining the "underexposure" of a shot. Should we call it the Rayposure?

Never mind if it's strange, what's incorrect about it?

It is not incorrect, it is absurd, for the required (why would it be required, btw?) exposure correction depends on the other adjustments (for  example Brightness = +50 is +1 exposure adjustment, except for the highlights), and on the illumination and on the colors of the scenery's highlights.

As you can see in the screen grab of the histogram for my simulated ISO 51,200 shot, in the ACR window, all adjustments are at zero apart from EC which is +3. I rarely find that a brightness of +50 is right for an image. Why should I use that as a standard? I'm trying to assess exposure as accurately as possible in the program I use for all image conversions. My reasoning is, if the shot of the lady were fully exposed at ISO 3200, I would expect to get a histogram looking as it does with a -1 EC adjustment. Brightness and contrast can be adjusted to taste in Photoshop and as you can see from the 'levels' histogram after the conversion, there's plent of scope for brightness adustment and other processing.

I thought the topic is the D3S.

It is, but the closest camera to it, and the one for which we have test results, is the D3. Considering the D3 has no more than a 1/2 stop improvement over the 5D at ultra-high ISO, I'd be surprised if the D3s offers more than an another 1/2 stop imrovement over the D3. But I have no objection to being surprised  .

Your own tests indicate nothing useful. The rendered ISO 102400 shot of D3S is crap. I don't know how you created it, but the non-rendered Rawnalyze display is better than that; thus the comparison is worthless. If you are interested on the outcome of the comparison, you should do it with a Nikon Capture X rendered image, compared to the DPP output of your 5D shot.

I didn't render it. I believe it's a rendering by Iliah Borg using RPP (linked in the second post in this thread). If you think it's crap, better contact Iliah Borg.

[Ray]

Gabor,
Here's my simulated ISO 51,200 shot with +3 EC, +50 brightness and +25 contrast. From an ETTR perspective in ACR this definitely looks like a 4 stop underexposure to me. Can't see anything absurd here, except the rather garish rendition is not to my taste.

[attachment=17549:5D_50_br...R_window.jpg]