How does the D3 achieve such high ISO?

Started Feb 17, 2008 | Discussions
PIXmantra Senior Member • Posts: 1,637
Well, this does not seem to be true...

Marianne Oelund wrote:

Honor wrote:

The Canon 1Dm3 sensor has better high ISO than the Nikon D3. period.

I own both cameras, and have tested both extensively. I do not rely
on other's test results, which are too easy to misinterpret.

...But the same thing could be said about your tests or conclusions (which I do observe, BTW).

Fact #1: Quantum efficiency is significantly higher on the D3,
yielding greatly reduced shot noise.

...It is higher but not as efficiently coupled with its actual Full Well capacity. As a result, you can achieve higher dynamic range with the 1D3, and a better or more-balanced distribution of sensitivity and performance that includes ISO50.

Fact #2: Read noise levels are nearly the same on the two cameras,
across the ISO range.

Well, the 1D3 seems to have LOWER levels of read-out noise, indeed. This is exactly what allows it to attain to a slightly larger dynamic range, at ISO100, than anything the D3 can attain at any ISO... And that is with substantially SMALLER, 7.2 micron sensels! :-)))

Fact #3: The D3 has in-camera NR which significantly improves IQ of
its JPEG output.

...Oh yes, this is undeniable... but that does not change the underlying or underpinnings.

The result: The D3 is superior across the board, in its RAW output,
and its advantages regarding JPEG output are even greater.

The D3 shows its 0.5EV advantage in the ISO2000-ISO4000 range... but it actually seems to offer DISADVANTAGES at any other lower step of the scale. That is thanks to 1D3 usable (and superior) ISO50-ISO200 performance and dynamic range, not to mention superior per-pixel quality and sharpness, most notably in .JPGs images (the 1D3 are clearly more detailed and sharp, with less sharpening effort), all the way up to ISO3200.

If you insist on continuing to believe the Mk III is the equal of the
D3, you are entitled to your opinion

...Well, R. Clark's finding do not seem to be very aligned to yours (notice the Perceived Image Quality model that brings noise and sharpness together):

http://www.clarkvision.com/imagedetail/digital.sensor.performance.summary/index.html

At the end, the result boils down to:

1. NIKON D3 is almost exclusively tuned for High-ISO shooting (but NOT up to ISo25+ is hyped. It actually maxes out at ISO4000-6400). Actual, real advantage in THAT range is about +0.5EV.

2. EOS 1D MarkIII is tuned to be used accross pretty much everything, from low-light captures up to ISO3200 to day-light scorching scenes or fill-in-flash needs at ISO50-ISO100 due to less sync. speed required, which is UNATAINABLE on D3.

  1. 2 for me and my needs, please. I know you will be very happy with #1, but that would be at the expense of other precious things, that is.

Enjoy!

but I see you as no wiser than

those who continue to insist that the earth is flat.

As far as the 1DSm3 it should have same ISO as the 1dS2 since Canon
stated the photodiode pitch in a pixel has been kept the same even
though the pixel area is smaller. Only the light insensitive portion
of the pixel was stripped out.

This same argument was made regarding the 1D Mk III vs. the 1D Mk II.
The truth is that the Mk III ended up with a smaller full well
capacity.

Here is proof of the superior ISO from the 1D3 over the Nikon D3.

If you think that is proof, you obviously had your mind made up
before you even read it.

-- hide signature --

'What we do in life, echoes in eternity...'

blackhawk13 Veteran Member • Posts: 4,742
Re: Glass is a huge issue, L glass wuv!

SDRebel wrote:

Having a few Canon lenses myself, that is a major consideration when
thinking of switching.

However, switching isn't my issue. Rather, I just want Canon to
provide the ability to set the higher ISO available with the D3 -- if
it could be done with a firmware upgrade.

Whether extra ISO settings can be provided via firmware is one of the
questions I've been trying to get answered.

-- hide signature --

The control styles and the RAW format are big concerns as well. I never lost (or corrupted)a single Canon RAW file while using ZB or DPP. Can Nikon make to same boost?

Compared to the MK-3, the D3's advantages are small or in some cases nonexistent.

The D3 is a big leap for Nikon, although it took a retaredly long time, sloppy of them. It's far from a knockout punch though, and at this point, not longer concerns me.... and I wuv my L glass!!!!

-tyrants and cowards hide behind censorship-

Marianne Oelund Veteran Member • Posts: 7,788
Let's normalize things to the same ISO

PIXmantra wrote:

Marianne Oelund wrote:

I own both cameras, and have tested both extensively. I do not rely
on other's test results, which are too easy to misinterpret.

...But the same thing could be said about your tests or conclusions
(which I do observe, BTW).

Indeed. However, I would openly invite any whom are interested and locally available, to perform tests with my cameras.

Fact #1: Quantum efficiency is significantly higher on the D3,
yielding greatly reduced shot noise.

...It is higher but not as efficiently coupled with its actual Full
Well capacity. As a result, you can achieve higher dynamic range with
the 1D3, and a better or more-balanced distribution of sensitivity
and performance that includes ISO50.

What you are actually doing now (and Roger Clark fell into the same error in compiling his Figure 4), is to compare the Mk III at ISO 50, with the D3 at ISO 200! All that this proves, is that the D3 DR at ISO 200 falls short of the Mk III at ISO 50 - should anyone be surprised? Ask yourself: How does the Mk III at ISO 200 compare to itself at ISO 50?
Let's put the comparison on the same basis: ISO 200.
D3 full-scale e- count = 65,600 (Roger's number)
D3 read noise = 6 e- (my number, more conservative than Roger's)
Mk III full-scale e- count = 22,000
Mk III read noise = 4 e-
D3 sensor DR = 65,600/6 = 11,000 = 13.4 stops
Mk III sensor DR = 22,000/4 = 5,500 = 12.4 stops

The most important point here, though, is that huge 3:1 ratio of electron counts, which holds as we ascend the ISO scale. We'll see below, how significant this is to noise levels.

Fact #2: Read noise levels are nearly the same on the two cameras,
across the ISO range.

Well, the 1D3 seems to have LOWER levels of read-out noise, indeed.
This is exactly what allows it to attain to a slightly larger dynamic
range, at ISO100, than anything the D3 can attain at any ISO... And
that is with substantially SMALLER, 7.2 micron sensels! :-)))

Again, we must use the same basis (the D3 does not actually have ISO 100). Expressing read noise in absolute electrons fails to correctly compare the noise level at the A/D converter outputs, when the full-scale electron counts are so different (3:1 in favor of D3). What is the noise expressed in DN, which is what we actually see in the image?
For example, at a moderate sensitivity of ISO 800:
D3 full-scale e- count = 16,400
D3 read noise = 6 e-
Mk III full-scale e- count = 5,500
Mk III read noise = 4 e- (using an optimistic value; actual is a bit higher)
D3 read noise = 6/16,400 * 16,383 DN = 6 DN
Mk III read noise = 4/5,500 * 16,383 DN = 12 DN

Thus the D3's read noise component at the A/D converter output is only half that of the Mk III (note my "Fact #2" statement previously was based on Roger's older data for the Mk III of 2.1 e-, which he has since updated).

The result: The D3 is superior across the board, in its RAW output,
and its advantages regarding JPEG output are even greater.

The D3 shows its 0.5EV advantage in the ISO2000-ISO4000 range... but
it actually seems to offer DISADVANTAGES at any other lower step of
the scale. That is thanks to 1D3 usable (and superior) ISO50-ISO200
performance and dynamic range, not to mention superior per-pixel
quality and sharpness, most notably in .JPGs images (the 1D3 are
clearly more detailed and sharp, with less sharpening effort), all
the way up to ISO3200.

As already addressed above, the Mk III is inferior at ISO 200. Comparing lower ISO values is not possible, as the D3 doesn't offer them.

The Mk III has very aggressive JPEG sharpening, but this acts to its detriment in regard to noise, if one is not careful. I find that to achieve comparable JPEG sharpness, the Mk III sharpening must be set to "0", with the D3 at "3" or "4". With those settings, detail capture is equivalent, yet the D3 still shows lower noise levels, even with NR turned off. With the D3's NR turned on, it's simply no contest.

...Well, R. Clark's finding do not seem to be very aligned to yours
(notice the Perceived Image Quality model that brings noise and
sharpness together):

I'm not convinced of the accuracy of Roger's AIQ model, but putting my doubts aside for a moment, he gives the D3 a rating of 69, to the Mk III's 57. Is not the higher value the more desirable one?

At the end, the result boils down to:

From ISO 200 and up, the D3 significantly outperforms the Mk III, in both DR and noise levels, if one is careful to use data for corresponding ISO settings as the basis for comparison. To do otherwise results in a serious miscompare.

OP SDRebel Senior Member • Posts: 1,489
In 2 months all pros will have D3?

I'm not quite sure what to conclude from your remarks.

Is it your opinion that a professional photographer should, in general, now be using the D3 rather than the 1Ds3?

If so, there is not only a $4000 savings but the opportunity to out shoot all the Canon loyalists.

My sense of the matter is that reviewers still favor the 1Ds3; however, Nikon does offer high ISO opportunities that are occasionally useful and so far can't be matched by the 1Ds3.

No doubt if I were just beginning to acquire equipment and didn't have numerous Canon lenses, I would carefully consider the D3. After all, $4000 buys several great lenses. But, I don't know what my decision would be, the issue of high ISO (which I would like to have) aside.

ejmartin Veteran Member • Posts: 6,274
1D3 full well?

Marianne, where are you getting your figure of 22K electrons at saturation for the 1D3 at ISO 200? Mine has 71K electrons at saturation for ISO 100 and about 38K at ISO 200. 22K seems awfully low, more like the value for the 1D2.

As for the change in Clark's data, he had measured the gain at ISO 50, not realizing that ISO 50 is really ISO 100 overexposed by a stop. So values for gain above ISO 50 are double what he intially reported, and read noise in electrons consequently also doubled for all ISO's above 50. (There was clearly a problem, since if you took the original figures seriously, the 1D3 had the worst photon collection efficiency of any Canon DSLR since before the 10D).
--
emil
--

http://theory.uchicago.edu/~ejm/pix/20d/

Honor Regular Member • Posts: 266
Re: Do your own testing, please

Marianne Oelund wrote:

Honor wrote:

The Canon 1Dm3 sensor has better high ISO than the Nikon D3. period.

I own both cameras, and have tested both extensively. I do not rely
on other's test results, which are too easy to misinterpret.

You credibility is about the same as a cow in a pasture who tells me it saw an ufo last night. You have published zero to substantiate your fallacies yet you accuse other people of tainting their results when they at least have made the effort to publish their findings. This foremosts tells me you could well be a Nikon troll. While I am ready to believe you own a D3, I highly doubt you own a 1dmk3.

Here is what you hit:
credibility 0%
arrogance 80%

I did not make my mind on that article alone but while it may be a limited scope test it is as untainted as it gets, same lense on both cameras no post-processing.

The 1dmk3 has superior IQ compared to the D3 in all usable ISO.

The D3 noise reduction is actually removing details and that is not a path I will condone.

Marianne Oelund Veteran Member • Posts: 7,788
Re: 1D3 full well?

ejmartin wrote:

Marianne, where are you getting your figure of 22K electrons at
saturation for the 1D3 at ISO 200? Mine has 71K electrons at
saturation for ISO 100 and about 38K at ISO 200. 22K seems awfully
low, more like the value for the 1D2.

Can you reconcile your values to Roger Clark's 12-bit unity gain ISO value of 1000 for the Mk III? That would imply 20,500 electrons full-scale at ISO 200.

As for the change in Clark's data, he had measured the gain at ISO
50, not realizing that ISO 50 is really ISO 100 overexposed by a
stop. So values for gain above ISO 50 are double what he intially
reported, and read noise in electrons consequently also doubled for
all ISO's above 50. (There was clearly a problem, since if you took
the original figures seriously, the 1D3 had the worst photon
collection efficiency of any Canon DSLR since before the 10D).

Thanks for the explanation - I was wondering. Has this also caused an error in his unity-gain ISO calculation?

If the Mk III unity-gain ISO is actually closer to 2000, then the D3 and Mk III would be back to rough parity regarding read noise, with the D3 retaining an advantage in shot noise.

ejmartin Veteran Member • Posts: 6,274
Re: 1D3 full well?

Marianne Oelund wrote:

ejmartin wrote:

Marianne, where are you getting your figure of 22K electrons at
saturation for the 1D3 at ISO 200? Mine has 71K electrons at
saturation for ISO 100 and about 38K at ISO 200. 22K seems awfully
low, more like the value for the 1D2.

Can you reconcile your values to Roger Clark's 12-bit unity gain ISO
value of 1000 for the Mk III? That would imply 20,500 electrons
full-scale at ISO 200.

That has as its source the same gain measurement. "Unity gain ISO" is simply the gain times the ISO at which it is measured, normalized to 12-bit ADU's in Clark's conventions. So Clark's original 2.46 electron/14-bit ADU at ISO 100, is about 1000 for the 12-bit normalized unity gain ISO. Now with 4.8 electrons/14-bit ADU at ISO 100, his unity gain ISO is around 1900.

That translates into about 37K electrons at ISO 200 raw saturation.

As for the change in Clark's data, he had measured the gain at ISO
50, not realizing that ISO 50 is really ISO 100 overexposed by a
stop. So values for gain above ISO 50 are double what he intially
reported, and read noise in electrons consequently also doubled for
all ISO's above 50. (There was clearly a problem, since if you took
the original figures seriously, the 1D3 had the worst photon
collection efficiency of any Canon DSLR since before the 10D).

Thanks for the explanation - I was wondering. Has this also caused
an error in his unity-gain ISO calculation?
If the Mk III unity-gain ISO is actually closer to 2000, then the D3
and Mk III would be back to rough parity regarding read noise, with
the D3 retaining an advantage in shot noise.

Indeed, that's how I see it. The D3 has an advantage of about 60% in per-pixel photon collection efficiency over the 1D3, though that becomes about 15% when referred to a per-area basis. The 1Ds3 and 1D3 have comparable efficiencies on a per-area basis.
--
emil
--

http://theory.uchicago.edu/~ejm/pix/20d/

Marianne Oelund Veteran Member • Posts: 7,788
Re: Do your own testing, please

Honor wrote:

You credibility is about the same as a cow in a pasture who tells me
it saw an ufo last night. You have published zero to substantiate
your fallacies yet you accuse other people of tainting their results
when they at least have made the effort to publish their findings.
This foremosts tells me you could well be a Nikon troll. While I am
ready to believe you own a D3, I highly doubt you own a 1dmk3.

You will be hard pressed to find anyone, anywhere, with more experience using the Mk III in precisely the situations at issue here. Since July, over 275,000 frames of low-light action. I know the IQ of this camera very well, indeed. I am not the one spreading fallacies here.

Here is what you hit:
credibility 0%
arrogance 80%

Your opinion of my credibility is immaterial; at least I don't believe in talking cows. The numbers speak for themselves - you should be arguing with them, not me.

I did not make my mind on that article alone but while it may be a
limited scope test it is as untainted as it gets, same lense on both
cameras no post-processing.

The 1dmk3 has superior IQ compared to the D3 in all usable ISO.

I invite you to visit and prove this with my cameras.

The D3 noise reduction is actually removing details and that is not a
path I will condone.

We can leave NR off, if you prefer.

D3 pusher Regular Member • Posts: 250
Re: Do your own testing, please

hmm .. I think youshould read her other threads

D3 pusher Regular Member • Posts: 250
Re: Very simple: LARGE photo-sensitive area and MASSIVE N.R....

I think your over reaching by a mile....... read your own posts. Go buy the camera and do your own experiments rather then borrowing images that you had no control over . Marianne Oelund at least has the cameras and can tell you first hand what her impression is whether subjective or not. This whole thread reminds me of school days when we used to get Motortrend and talk numbers with cars and argue whats better???

sandy b
sandy b Veteran Member • Posts: 9,676
These seem usable
 sandy b's gear list:sandy b's gear list
Nikon D750 Nikon D7500 Nikon Z6 Nikon Z50 Nikon AF-S Teleconverter TC-14E II +12 more
Marianne Oelund Veteran Member • Posts: 7,788
You can do high-ISO boost yourself

SDRebel wrote:

Given what I paid for my 1Ds3, I'd like to think that it outperforms
every camera on the market in all respects.

It is not possible for any one camera to be the best in all respects. You have the highest-resolution 35mm-format dSLR currently available; it is not reasonable to expect it to also be the best high-sensitivity camera. That is an unavoidable tradeoff.

Anyway, my own initial testing of ISO 3200 leads me to believe that
for night urban work of people on the streets, etc., that the results
are perfectly usable. I'd just like Canon to give me the opportunity
to set the camera at the 6400, 12800, and 25600 choices offered by
the D3 and thereby allow me to decide if the results are acceptable
and usable.

Once you are above unity-gain ISO, you can simulate higher sensitivities simply by underexposing, then adding exposure compensation back in during post. You will lose nothing in doing this, compared to what your camera would give you if Canon had designed in the higher ISO settings.

Don't hesitate to experiment with this. I'm sure you'll find the results interesting; you might even want to share some of them with us.

ohyva Veteran Member • Posts: 6,342
Re: Very simple: LARGE photo-sensitive area and MASSIVE N.R....

PIXmantra wrote:

...The massive N.R. piece applies to .JPGs, at the expense of overall
sharpness and image quality (as all samples, images and conversions
that I have seen/worked-with from the D3 indicate).

I'd not say massive, as I see D3 to be quite good.

No exact info of the Nikon sensor, but I'd say it's quite safe to expect that to use the new Sony CMOS technology. There you have two issues I see to affect the noise performance.

1) In Sony presentation material, you can see there is a NR block in the digital domain after ADC which explain the low(er) noise but also the slight softness it seems to be there is in all the images. No info though if this NR block can be disabled to get "true RAW data" from the sensor.

2) The sensitivity of the Sony sensor is made higher and thus the base ISO value 200. This should give upto 1 stop advantage at ISO200 and above. Negative side is then the slight reduction of DR at ISO100 (and of course no ISO50).

Generally, IMHO Sony has made a good job with their new sensor technology and got quite near if not par to what Canon has.

As for true RAW data, I have seen plenty of ISO6400 that do not
really impress me at all (that is, YES, from a pure technical
point-of-view, but NO in light of the "hype").

Here is just one of them:

http://www.pbase.com/feharmat/image/89206710/original

And, just to give you an idea of what happens with D3's on-board .JPG
cooking, check this example on how that technology actually does
wonders on MUCH less-capable cams (I had access to both .JPG and RAW
files on this example, camera RICOH GRD2, from-cam-.JPG on the left,
and my best LR conversion, on the right):

http://www.pbase.com/feharmat/image/89617778/original
http://www.pbase.com/feharmat/image/89617779/original

As you can see, the Ricoh GRD2 is using an on-board NR technology
that yields in REMARKABLY SIMILAR-LOOKING files and output as the D3
does with its .JPGs.

Enjoy!

F.H.

-- hide signature --

'What we do in life, echoes in eternity...'

OP SDRebel Senior Member • Posts: 1,489
Underexpose using ISO 3200?

Are you saying that I should set ISO to 3200, then using exposure compensation underexpose step-by-step to achieve a result whereby I can effectively push the 1Ds3 ISO to the levels of the D3?

OP SDRebel Senior Member • Posts: 1,489
These results seem to confirm reviews

Thanks for the cite to this thread.

I wish someone had been there with a 1Ds3 so we could compare the results it could achieve under the same lighting conditions.

OP SDRebel Senior Member • Posts: 1,489
I want high ISO

It isn't the end of the world; however, I want Canon to provide me the high ISO option the D3 achieves.

If that requires a tradeoff, then I may find the price of high ISO isn't worth paying, e.g., in lower quality files at other ISO settings.

But, if higher ISO can easily be provided by Canon with no significant tradeoff, then I would like to see a firmware upgrade, if that is possible.

I follow your point, yet I see it as engaging in a discussion similar to one I remember of whether or not carburators would inevitably be replaced by fuel injection.

The first thing I needed to know in that discussion was what fuel injection did that carburators didn't, if anything. Or, was fuel injection just marketing hype? Well, the answer has been in for a long time now; but in the beginning it generated quite a bit of discussion and required a learning curve.

So, again, how has Nikon accomplished the high ISO provided by the D3? Can Canon provide it in the 1Ds3 with a firmware upgrade or will it have to await a new sensor design? Is there a tradeoff that makes high ISO ultimately undesirable for most shooting conditions, given even the Nikon technology?

I still am not certain that I've read any definitive answers to my initial questions.

Many thanks for your thoughts -- I enjoy learning as well as I enjoy shooting with my 1Ds3 with the technology is presently offers even if that is limited for the moment to ISO 3200.

ohyva Veteran Member • Posts: 6,342
Re: That's the trick

SDRebel wrote:

Are you saying that I should set ISO to 3200, then using exposure
compensation underexpose step-by-step to achieve a result whereby I
can effectively push the 1Ds3 ISO to the levels of the D3?

commonly used - and already in the film days.

And that's the way I have understood the current dSLRs do when monig to Hi_xx settings. Would though better have some higher truly supported ISO setting as those are done using the analog amplifiers in the chip and thus slightly lower noise than those pushed Hi_xx "equivalents".

ohyva Veteran Member • Posts: 6,342
Re: 1D3 full well?

Marianne Oelund wrote:

ejmartin wrote:

Marianne, where are you getting your figure of 22K electrons at
saturation for the 1D3 at ISO 200? Mine has 71K electrons at
saturation for ISO 100 and about 38K at ISO 200. 22K seems awfully
low, more like the value for the 1D2.

Can you reconcile your values to Roger Clark's 12-bit unity gain ISO
value of 1000 for the Mk III? That would imply 20,500 electrons
full-scale at ISO 200.

First it's good to remember the full-well capacity in Canon cams are reached at ISO100 while in D3 it's at ISO200. So comparison at ISO200 is apples against oranges. What's then more important, low noise in general or low noise at certain expo-time&aperture settings, depends then on the aimed use of the gear. I persponally prefer speed for sport photos, so the new Nikon would be tempting if I'd not be so committed with my current Canon gear.

Second I'm still sceptic about these well capacity "measurements". No direct measurement method AFIAK (unless you probe into the chip which I think has not been done (and would probably to some extend interfere with the masurement). If done indirectly based on noise figs, then I'd really love to know how all the different noise sources are separated. And how to ensure there is no in-sensor digital domain NR applied - which is reported by Sony to be one of the methods how they have reduced the noise levels in their new CMOS sensor technology.

As for the change in Clark's data, he had measured the gain at ISO
50, not realizing that ISO 50 is really ISO 100 overexposed by a
stop. So values for gain above ISO 50 are double what he intially
reported, and read noise in electrons consequently also doubled for
all ISO's above 50. (There was clearly a problem, since if you took
the original figures seriously, the 1D3 had the worst photon
collection efficiency of any Canon DSLR since before the 10D).

Thanks for the explanation - I was wondering. Has this also caused
an error in his unity-gain ISO calculation?
If the Mk III unity-gain ISO is actually closer to 2000, then the D3
and Mk III would be back to rough parity regarding read noise, with
the D3 retaining an advantage in shot noise.

FretNoMore Veteran Member • Posts: 7,729
Re: trade-off

Marianne Oelund wrote:

...
It is not possible for any one camera to be the best in all respects.
You have the highest-resolution 35mm-format dSLR currently available;
it is not reasonable to expect it to also be the best
high-sensitivity camera. That is an unavoidable tradeoff.
...

You're right here of course, and I think Nikon probably made the better choice for this type of camera. It's geared towards PJ and sports photographers and for their publishing needs and shooting conditions I think most of them will gladly give up some at the low-ISO end to get better high-ISO performance. I'm not going to switch systems to get it - I don't doubt Canon will have an answer as technology inevitably improves and now that there is competition in this segment.

Keyboard shortcuts:
FForum MMy threads