How does the D3 achieve such high ISO?

Started Feb 17, 2008 | Discussions
Osiris Ani Senior Member • Posts: 1,126
Entitlement

What is it that you think Canon can suddenly make better than the 1Ds3?

It's a case of unreasonable expectations due to an unwarranted sense of entitlement. He's probably also still hoping that someone will tell him that, yes, Canon will enable those ISO settings in his 1Ds III via a firmware update.

...Real Soon Now™.

-- hide signature --

'Passion will make you crazy, but is there any other way to live?' - Kara Saun

 Osiris Ani's gear list:Osiris Ani's gear list
Canon EOS 5D Mark II Canon EF 50mm F1.4 USM Canon EF 24-70mm f/2.8L USM Canon EF 28-135mm F3.5-5.6 IS USM Canon EF 100mm f/2.8 Macro USM +2 more
Scott Larson Veteran Member • Posts: 7,033
Re: Easy: large CMOS sensor, large pixels and good electronics

bullet1 wrote:

With the updated electronics, larger CMOS sensor and larger pixels,
it is not that big a surprise that the D3 can achieve another stop
high ISO advantage over the 1D3 and 5D. Its other ISO speeds, 12800
and 25600, are really there for marketing reasons.

There are several D3 images in this thread that were taken beyond ISO 6400 that I think are usable for many purposes. Maybe not for yours, but certainly for others:

http://forums.dpreview.com/forums/read.asp?forum=1021&message=26647291

I wish Canon let us adjust the ISO between ISO 3200 and 6400. It's as if they added "H" for marketing purposes.

 Scott Larson's gear list:Scott Larson's gear list
Sony RX10 IV Canon EOS 5D Canon EOS 5D Mark II Canon EOS-1D Mark IV Canon EOS-1D X +16 more
ejmartin Veteran Member • Posts: 6,274
Re: Who doesn't get it

DSPographer wrote:

Sensor noise for Canon and and Nikon D-SLR cameras comes from the
photosite pre-amplifier not from dark leakage (dark leakage is
negligible for exposures faster than 1s). This means that sensor
noise is independant of photosite area so bigger photosites are more
sensitive.

With respect to read noise, I was making an empirical observation.

Aas an extereme example, the Panasonic FZ50 has 2µ pixels and a read noise of 2.3 electrons at ISO 100. The Canon 1D3 has 7.2µ pixels and a read noise of 23 electrons at ISO 100. It takes 13 FZ50 pixels to make the same collection area as the 1D3; combining their noise as independent fluctuations would yield an aggregate read noise of less than 9 electrons, less than half that of the 1D3. The 40D has read noise of 17 electrons at ISO 100. Scaling by the pixel size, 7.2/5.7, to refer this noise to a per area basis, this is slightly less than the 1D3's read noise on a per area basis.

I haven't looked at the data for a wide variety of cameras, but there is somewhat of a trend for low ISO read noise to vary with pixel size. For high ISO, I agree that there seems to be something of a wall at about 4 electrons of read noise at ISO 1600 that seems to be independent of pixel size for the last couple of generations of 1-series cameras.

So there will be a deteriment to smaller pixels in shadow noise for high ISO; for midtones and highlights where photon noise dominates, there is no penalty within a wide range of pixel sizes, as the example of the 40D vs 1D3 shows -- they collect the same amount of light per unit area of sensor, while the 40D has 60% higher pixel density. Thus photon noise at fixed spatial frequency (lines per mm) is the same for these two cameras.

-- hide signature --
Hans Kruse Veteran Member • Posts: 3,432
Re: Easy: large CMOS sensor, large pixels and good electronics

Scott Larson wrote:

I wish Canon let us adjust the ISO between ISO 3200 and 6400. It's as
if they added "H" for marketing purposes.

That's easy: Shoot raw and underexpose and you will automatically have > iso 3200. Compared to a camera setting you don't loose much since 14 bits should be enough to represent the information available in the raw file.

 Hans Kruse's gear list:Hans Kruse's gear list
Canon EOS 5DS
bullet1 Veteran Member • Posts: 7,326
ISO12800 on the 5D!

I have shot in RAW (of course) with the 5D at ISO3200, 2 stops underexposed to get the decent shutter speed due to the dim lighting. Capture One Professional was able to produce very decent results. Some require a bit Neat Image but the effective ISO speed was ISO12800 on a 5D!

All of the shots in the gallery referenced below were shot 1 to 2 stops underexposed, some more like 2.5 under. The lighting was so bad. I was not able to shoot from a closer distance with a faster prime on a tripod. All shots were handheld with the 5D and the 70-200 f2.8 IS at 2.8 and ISO3200. The light was so dim that you could not even read very well.

http://www.nelsonchenphotography.com/gallery/4258695_NT78n#249362250
--
Nelson Chen
http://pbase.com/nelsonc
http://NelsonChenPhotography.com/
100% RAW shooter with Capture One Pro

2007 Colorado Renaissance Festival photo gallery:
http://www.pbase.com/nelsonc/2007_colorado_renaissance_festival&page=all

 bullet1's gear list:bullet1's gear list
Panasonic FZ1000 Canon PowerShot G5 X Canon EOS 6D Canon EOS M5 Sony a7 III +22 more
ChanStudio New Member • Posts: 17
Re: WRONG Canon 1D3 has better ISO than the D3

Check out the beloved test between Canon's L Prime 14mm against nikon 14-24mm f2.8G AFS. The "L" is nothing more than marketing hype.

blackhawk13 wrote:

Honor wrote:

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

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

Period.

ISO 3200 1DM3

http://www.clublexus.com/forums/attachment.php?attachmentid=116418&stc=1&d=1199070427

ISO 3200 1D3

http://www.clublexus.com/forums/attachment.php?attachmentid=116415&stc=1&d=1199070052

http://www.clublexus.com/forums/showthread.php?t=326811

ChanStudio New Member • Posts: 17
Re: Nikon is the best in noise reduction. That's all

TheVoIP wrote:

SDRebel wrote:

The jury seems to be in on the Nikon D3's successful implementation
of significantly higher and usable ISO than I can achieve with my
1Ds3.

Has Nikon made a breakthrough in sensor design? Or, is the Nikon ISO
range a result of a breakthrough in post processing?

Thanks to Nikon bad past experinces with CCDs that are terrible with
noise. Thier engineers have no choice but to researce heavily on
noise reduction post processing in order to get 'accpetable' image.

When they have similar CMOS sensor as Canon, thier advantages of
having previous noise reductions algorithm are rewarded, with the so
call, low noise at ISO6400.

The fact is that.... However, to me excessive noise reduction destroy
image quality. I wonder why in the first place people buying great
lenses that insist on sharpness, where they are willing to scarify
that for low noise, which can be done on computer software as well?

Anyway, the success of D3 works for most people because they were not
from photography background and do not have the capability to judge
image quality using thier own eyes. Thus the cheat from Nikon works
for them.

I am not saying that D3 suck... I am saying it is not better than 2
years old EOS 5D nor 1D mark III in actual image quality. I am still
with Canon because it clearly provide me with some of the best lenses
for the job; EF85 f1.2 L II, EF135 f2 L and EF100-400 IS L that Nikon
do not have.

I tell you what. Why don't you show me your work and I show you my work and see how your statement stack up against? By the way, I have no favor of either brand.

DSPographer Senior Member • Posts: 2,483
Re: Who doesn't get it

Why quote read noise at minimum ISO? That just tells you about the maximum photosite level minus the A-D system DR. The read noise at high iso tells you about the sensor sensitivity. State of the art cameras have read noise of about 2.5 electrons independant of photosite size.

 DSPographer's gear list:DSPographer's gear list
Canon EOS 5D Mark II Canon EOS M5 Canon EF 24mm f/2.8 Canon EF 50mm F1.8 II Canon EF 200mm f/2.8L II USM +4 more
OP SDRebel Senior Member • Posts: 1,489
New Plateau

I don't know whether or not Canon will alter its schedule for introducing upgrades to current models such as the 1Ds3.

However, given Nikon's accomplishment and the Sony announcement, it seems possible that there will be cameras available by next year that raise the bar substantially.

I have no idea why the 1Ds3 commands the price that it does. For me the price wasn't relevant. Perhaps the Sony and whatever else arrives at 25 megs, if these new cameras can match the ISO range of the D3, will be priced similarly to the 1Ds3, as opposed to the $4000 tag on the D3.

What I don't have the slightest doubt about is that if Nikon can produce the ISO range of the D3 coupled with a 25 meg sensor, Nikon will hit the market with that camera body ASAP.

Will Canon respond? Your guess may be far better than mine. My guess is that Canon will be of the opinion that it has to maintain the lead and so Canon will respond regardless of the impact on normal schedules for product cycles.

Time will tell. And if the money isn't important in comparison to what the new products can offer, well we all will benefit if Canon reaches a new plateau before anyone else because of our investment, at least speaking for myself, in Canon lenses.

Meanwhile, I'm having a great time taking photographs with my 1Ds3.

Scott Larson Veteran Member • Posts: 7,033
Re: Easy: large CMOS sensor, large pixels and good electronics

Hans Kruse wrote:

That's easy: Shoot raw and underexpose and you will automatically
have > iso 3200.

Easy for you. Not easy for me to shoot raw at the basketball games.

 Scott Larson's gear list:Scott Larson's gear list
Sony RX10 IV Canon EOS 5D Canon EOS 5D Mark II Canon EOS-1D Mark IV Canon EOS-1D X +16 more
ohyva Veteran Member • Posts: 6,342
Re: 1D3 full well?

ejmartin wrote:

Really? Googling "dark current noise", the first listing
http://www.roperscientific.de/tdark.html

I just wonder why then so many research activities in silicon sensor structures to minimize the dark current. Like this one WO patent http://www.wipo.int/pctdb/en/wo.jsp?wo=2005083791

Any isolation should not make any effect to thermal noise as far as I can remember to my very much past phisics lessons.

ejmartin Veteran Member • Posts: 6,274
Re: Who doesn't get it

DSPographer wrote:

Why quote read noise at minimum ISO? That just tells you about the
maximum photosite level minus the A-D system DR. The read noise at
high iso tells you about the sensor sensitivity. State of the art
cameras have read noise of about 2.5 electrons independant of
photosite size.

Because so many people seem to think that small photosites are inherently noisy with low DR, not realizing that in comparing a small photosite camera to a large photosite camera by looking at 100% is not comparing the two on an equal footing.

So one point I wanted to make is that at low ISO, and comparing the image quality (noise and DR) at a fixed spatial scale, there is essentially no difference between the two, at any luminance level, and the small photosite camera has the additional advantage of higher resolution.

For high ISO applications, I agree that the one advantage of larger photosites is in high ISO read noise, yielding lower noise in shadows; photon noise, and hence noise in midtones and highlights, is still comparable between the small and large photosite cameras, just as it was at low ISO.

-- hide signature --
bullet1 Veteran Member • Posts: 7,326
Re: Easy: large CMOS sensor, large pixels and good electronics

Scott Larson wrote:

Hans Kruse wrote:

That's easy: Shoot raw and underexpose and you will automatically
have > iso 3200.

Easy for you. Not easy for me to shoot raw at the basketball games.

Is the buffer not large enough or the RAW processing slowing you down?

If you are shooting pro basketball games with ceiling/catwalk mounted strobes, you really cannot shoot faster than the strobe refresh time. It would be like 1 to 2 frames per second. The buffer is a non-issue. If you are shooting with existing light, the buffer still should not be a problem unless your style is machine gun all the time. I tend to time my shots at the right moment and get 2 or 3 in succession. The buffer is never a problem with shooting 100% RAW. I don't ever shoot RAW+JPEG which I consider a real waste of time and space.

If the RAW process is slowing you down, you don't have the right RAW processor. Capture One Pro can load, adjust and process a large number of files in a hurry. For those type of events, I never use PhotoShop. I would load all the RAW files in, typically a few hundred to 1000. Adjusting WB on one and apply to all simiilar ones. Next just pick out ones you want to keep, process them and throw them into the processing queue. By the time you have gone through the files, the ones you want are processed and ready to be burned to a CD or posted to the web. I can process RAW files faster than JPEGs with the tools I have.
--
Nelson Chen
http://pbase.com/nelsonc
http://NelsonChenPhotography.com/
100% RAW shooter with Capture One Pro

2007 Colorado Renaissance Festival photo gallery:
http://www.pbase.com/nelsonc/2007_colorado_renaissance_festival&page=all

 bullet1's gear list:bullet1's gear list
Panasonic FZ1000 Canon PowerShot G5 X Canon EOS 6D Canon EOS M5 Sony a7 III +22 more
Marianne Oelund Veteran Member • Posts: 7,788
The mathematician's vs. the engineer's approach

ohyva wrote:

ejmartin wrote:

Really? Googling "dark current noise", the first listing
http://www.roperscientific.de/tdark.html

I just wonder why then so many research activities in silicon sensor
structures to minimize the dark current. Like this one WO patent
http://www.wipo.int/pctdb/en/wo.jsp?wo=2005083791

Any isolation should not make any effect to thermal noise as far as I
can remember to my very much past phisics lessons.

I can assure you that the dislocations which that patent addresses are without question of thermal origin. Do you honestly believe that component of dark current would persist at absolute zero?

To the main point of this post:

Mathematicians expend considerable effort studying single-point anomalies such as singularities, path-dependent limits, etc., which have no practical real-world counterpart (aside from black holes). Engineers recognize what is most significant, pertinent and practical, and limit their dealings to such aspects.

What you are doing is devoting a disproportionate effort to effects which, if significant, would already have been included in the noise models. The models work very well, to experimental accuracy. That has been repeatedly demonstrated, and if you insist on pursuing negligible effects, assume that you will be on your own with them; you can certainly count on no support here.

Finally, it is beyond annoying to have one who has asked a question, then turn around and put up a series of arguments against the answer provided. Emil has been more than patient with you - far more than you deserve.

MirkoK Contributing Member • Posts: 741
Re: How does the D3 achieve such high ISO?

our "beloved" KR made a compairsion of the D3 and 5D, at same (high) iso levels, the 5D remained to be more detailful!

D3 pusher wrote:

Loss of quality as compared to what might I ask??

-- hide signature --
MirkoK Contributing Member • Posts: 741
Re: I want high ISO

well i didnt read this post completely because i have enough of reading the same stuff over and over, but i will repeat myself once more.

nikon redesigned so called microlenses, which seemed to be also one part of causing noise so.....NO canon cannot provide a firmware upgrade to reach the level of low noise as the D3.......the 1dsIII is "probably" using the same microlenses technology like the 5D, and has a MUCH higher pixel density and the 2nd to smallest pixel size of all cameras, i think the new 450D has even smaller.

i feel like in kindergarten, its been said a thounsand time and yet there are the "whyyyy"s.....

but yes canon can provide an firmware upgrade to extend to iso 6400 and maybe 12800.....but do we want to see those results? no!

cheers

-- hide signature --
Kyle Jones
Kyle Jones Senior Member • Posts: 2,320
Darned impressive results. (nt)
Hans Kruse Veteran Member • Posts: 3,432
Re: I want high ISO

MirkoK wrote:

nikon redesigned so called microlenses, which seemed to be also one
part of causing noise so.....NO canon cannot provide a firmware
upgrade to reach the level of low noise as the D3.......the 1dsIII is
"probably" using the same microlenses technology like the 5D, and has
a MUCH higher pixel density and the 2nd to smallest pixel size of all
cameras, i think the new 450D has even smaller.

See the Canon white papers on micro lens design and you will see that it was optimized in the last generation over what was done for the 5D.

http://www.google.dk/search?hl=da&q=canon+1ds+white+paper&meta=

 Hans Kruse's gear list:Hans Kruse's gear list
Canon EOS 5DS
ohyva Veteran Member • Posts: 6,342
Re: The mathematician's vs. the engineer's approach

Marianne Oelund wrote:

ohyva wrote:

ejmartin wrote:

Really? Googling "dark current noise", the first listing
http://www.roperscientific.de/tdark.html

I just wonder why then so many research activities in silicon sensor
structures to minimize the dark current. Like this one WO patent
http://www.wipo.int/pctdb/en/wo.jsp?wo=2005083791

Any isolation should not make any effect to thermal noise as far as I
can remember to my very much past phisics lessons.

I can assure you that the dislocations which that patent addresses
are without question of thermal origin. Do you honestly believe that
component of dark current would persist at absolute zero?

Absolute zero point - no as there is (in theory) no movement whatsoever.

Is this strictly proportional to temperature - I do not know, as I do no longer remember all the mechanisms and formulas related to different current/charge leakages.

Term "thermal noise" AFIAK and can remember has a special meaning in physics, so not all thermal dependent phenomena fall in this category.

To the main point of this post:

Mathematicians expend considerable effort studying single-point
anomalies such as singularities, path-dependent limits, etc., which
have no practical real-world counterpart (aside from black holes).
Engineers recognize what is most significant, pertinent and
practical, and limit their dealings to such aspects.
What you are doing is devoting a disproportionate effort to effects
which, if significant, would already have been included in the noise
models.

And engineers, in real life, fall regularly into problems when trying to oversimplify the models. I can assure I know and have learned this in a painfull way in my 20 years of professional work in electrical engineering.

The models work very well, to experimental accuracy. That
has been repeatedly demonstrated, and if you insist on pursuing
negligible effects, assume that you will be on your own with them;
you can certainly count on no support here.

So did the Newton Laws before Treory or Relativity was introduced, now recently challenged with theories of Quantum Physics. All these in the dómains and limitations they have.

Finally, it is beyond annoying to have one who has asked a question,
then turn around and put up a series of arguments against the answer
provided. Emil has been more than patient with you - far more than
you deserve.

Patience is a virtue.

And Marianne, when you make the sort of statement below, I guess we can expect you have many the analysis beyond any reasonable doubt. So complaining other people to be too strict in details is a bit indiscreet IMHO.

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

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

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

papparazzi Senior Member • Posts: 1,997
Re: Ha, ha

PerL wrote:

Anyway, the success of D3 works for most people because they were not
from photography background and do not have the capability to judge
image quality using thier own eyes. Thus the cheat from Nikon works
for them.

So most people who buys a D3 for 5000 dollars doesnt have any
photographic background? Sounds probable.

go back to your hole TROLL
--

http://www.pbase.com/jdf

 papparazzi's gear list:papparazzi's gear list
Canon EOS 6D Canon EOS 5D Canon EOS 5D Mark II Canon EOS 7D Nikon D600 +10 more
Keyboard shortcuts:
FForum MMy threads