Digital Camera Sensor Performane Summary updated Feb 3

In article <cYEpj.1450$XI.77@text.news.virginmedia.com>, David J Taylor
<david-taylor.DeleteThis@blueyonder.neither-this-bit.nor-this-bit.co.uk> wrote:

> Roger N. Clark (change username to rnclark) wrote:
> > If you haven't visited this month, see:
> >
> > http://www.clarkvision.com/imagedetail/digital.sensor.performance.summary
> >
> > I've added data for several newer cameras, added a new
> > section on Apparent Image Quality (AIQ) along with models
> > for constant megapixels and constant format size. What formerly
> > looks like a scatter plot now is taking shape in that the
> > pixel size and sensor format are showing trends. The models
> > should now allow prediction of performance when new cameras
> > come out. The new AIQ plot is Figure 9.
>
> Most helpful, Roger, thanks. The D300 does look very good!

agreed, it's an excellent chart! one request - it would be nice to see
a dashed line for the 1.5x cameras (nikon, pentax, sony).
 >> Stay informed about: Digital Camera Sensor Performane Summary updated Feb 3 

nospam wrote:
> In article <cYEpj.1450$XI.77@text.news.virginmedia.com>, David J
> Taylor
> <david-taylor DeleteThis @blueyonder.neither-this-bit.nor-this-bit.co.uk>
> wrote:
>
>> Roger N. Clark (change username to rnclark) wrote:
>>> If you haven't visited this month, see:
>>>
>>> http://www.clarkvision.com/imagedetail/digital.sensor.performance.summary
>>>
>>> I've added data for several newer cameras, added a new
>>> section on Apparent Image Quality (AIQ) along with models
>>> for constant megapixels and constant format size. What formerly
>>> looks like a scatter plot now is taking shape in that the
>>> pixel size and sensor format are showing trends. The models
>>> should now allow prediction of performance when new cameras
>>> come out. The new AIQ plot is Figure 9.
>>
>> Most helpful, Roger, thanks. The D300 does look very good!
>
> agreed, it's an excellent chart! one request - it would be nice to
> see a dashed line for the 1.5x cameras (nikon, pentax, sony).

I'd like to see a cost per Apparent Image Quality increment chart, by
camera model.

--
Frank ess
 >> Stay informed about: Digital Camera Sensor Performane Summary updated Feb 3 

nospam wrote:
[]
>> Roger N. Clark (change username to rnclark) wrote:
>>> If you haven't visited this month, see:
>>>
>>> http://www.clarkvision.com/imagedetail/digital.sensor.performance.summary
[]
> agreed, it's an excellent chart! one request - it would be nice to
> see a dashed line for the 1.5x cameras (nikon, pentax, sony).

Do you mean in figure 9?

David
 >> Stay informed about: Digital Camera Sensor Performane Summary updated Feb 3 

In article <CTGpj.1501$XI.1484@text.news.virginmedia.com>, David J
Taylor <david-taylor.DeleteThis@blueyonder.neither-this-bit.nor-this-bit.co.uk>
wrote:

> > agreed, it's an excellent chart! one request - it would be nice to
> > see a dashed line for the 1.5x cameras (nikon, pentax, sony).
>
> Do you mean in figure 9?

yep. although it's obvious where the line would be, it would be
helpful for it to be there.
 >> Stay informed about: Digital Camera Sensor Performane Summary updated Feb 3 

Roger N. Clark (change username to rnclark) wrote:
> If you haven't visited this month, see:
>
> http://www.clarkvision.com/imagedetail/digital.sensor.performance.summary
>
> I've added data for several newer cameras, added a new
> section on Apparent Image Quality (AIQ) along with models
> for constant megapixels and constant format size. What formerly
> looks like a scatter plot now is taking shape in that the
> pixel size and sensor format are showing trends. The models
> should now allow prediction of performance when new cameras
> come out. The new AIQ plot is Figure 9.

Thanks, that'll take a while to absorb. The MTF is for a particular
f/stop? The curved lines seem to suggest a theoretical benefit of
increased MP count up to 60MP for a full frame sensor and a sweet spot
of 30MP, right? Why does the 1DmIII have three diamonds? Is fine grained
35mm film that bad at IQ?
 >> Stay informed about: Digital Camera Sensor Performane Summary updated Feb 3 

Paul Furman wrote:
> Roger N. Clark (change username to rnclark) wrote:
>> If you haven't visited this month, see:
>>
>> http://www.clarkvision.com/imagedetail/digital.sensor.performance.summary
>>
>> I've added data for several newer cameras, added a new
>> section on Apparent Image Quality (AIQ) along with models
>> for constant megapixels and constant format size. What formerly
>> looks like a scatter plot now is taking shape in that the
>> pixel size and sensor format are showing trends. The models
>> should now allow prediction of performance when new cameras
>> come out. The new AIQ plot is Figure 9.
>
> Thanks, that'll take a while to absorb. The MTF is for a particular
> f/stop?

Yes, 50% MTF response for 2 indicated f/stops.
The curves and key are labeled. The two limits used are f/8 and f/4.

> The curved lines seem to suggest a theoretical benefit of
> increased MP count up to 60MP for a full frame sensor and a sweet spot
> of 30MP, right?

Yes, that is what the model says. It would sure be interesting
to see images from such a camera. The model also indicates
that the 1.3x crop sensors will peak at about 12-megapixels
(assuming f/8 diffraction limited lens performance).

> Why does the 1DmIII have three diamonds?

It has one (a little arrow points to the middle one).
The other two are other sensors. Like the one to the
right is a Kodak KAI-16000 CCD.

> Is fine grained
> 35mm film that bad at IQ?

Yes, by the AIQ definition. Film resolves detail pretty
well, but the noise, through film grain, is pretty bad.
Other pages on my site show measurements of image noise on
scanned film.

Roger
 >> Stay informed about: Digital Camera Sensor Performane Summary updated Feb 3 

[A complimentary Cc of this posting was sent to
Roger N. Clark (change username to rnclark)
<username.TakeThisOut@qwest.net>], who wrote in article <47A6B0B2.8040005.TakeThisOut@qwest.net>:
> If you haven't visited this month, see:
>
> http://www.clarkvision.com/imagedetail/digital.sensor.performance.summary
>
> I've added data for several newer cameras

Thanks.

What is very surprizing is how large is the improvement due to switch
from 12-bit ADC to 14-bit: not *exactly* zero, but only 11%!

Camera Full well Noise at ISO50 Dynamic range Sensel
--------------------------------------------------------------------
Canon 1DMII CMOS 79,900* 30.6 2611 8.2um
Canon 1DMIII CMOS 70,500 24.4 2889 7.2um

Most probably the "improvement" in noise is just due to lower area of
a sensel, thus lower capacitance, thus larger voltage generated by
each electron. So it looks like 14-bitness is 93% hoax: the ADC
behaved as an ideal 10-bit ADC before, now it behaves as 10.15-bit...

Puzzled,
Ilya
 >> Stay informed about: Digital Camera Sensor Performane Summary updated Feb 3 

Ilya Zakharevich wrote:
> [A complimentary Cc of this posting was sent to
> Roger N. Clark (change username to rnclark)
> <username DeleteThis @qwest.net>], who wrote in article <47A6B0B2.8040005 DeleteThis @qwest.net>:
>> If you haven't visited this month, see:
>>
>> http://www.clarkvision.com/imagedetail/digital.sensor.performance.summary
>>
>> I've added data for several newer cameras
>
> Thanks.
>
> What is very surprizing is how large is the improvement due to switch
> from 12-bit ADC to 14-bit: not *exactly* zero, but only 11%!
>
> Camera Full well Noise at ISO50 Dynamic range Sensel
> --------------------------------------------------------------------
> Canon 1DMII CMOS 79,900* 30.6 2611 8.2um
> Canon 1DMIII CMOS 70,500 24.4 2889 7.2um
>
> Most probably the "improvement" in noise is just due to lower area of
> a sensel, thus lower capacitance, thus larger voltage generated by
> each electron. So it looks like 14-bitness is 93% hoax: the ADC
> behaved as an ideal 10-bit ADC before, now it behaves as 10.15-bit...
>
> Puzzled,
> Ilya

Try looking at more of the picture. The smaller pixel
has less photons, so if the 1DIII used the same 12-bit ADC
as in the 1DII, the higher read noise and lower full well
capacity would result in a dynamic range 11% SMALLER than
the 1DII, so overall, the 1DIII 14-bit ADC and read noise
has improved about 22% at low ISOs. But that too is not the whole
story. Look at higher ISOs, and the 1DIII read noise drops to
2.1 electrons compared to the 1DII 3.9 electrons, about
an 86% improvement, which helps high ISO imaging. That and
greatly improved fixed pattern noise makes the new generations
of DSLRs much better at high ISO imaging (both Canon and Nikon).

Read noise is not related to pixel size, see Figure 3.

"Hope this Helps"
Roger
 >> Stay informed about: Digital Camera Sensor Performane Summary updated Feb 3 

Ilya Zakharevich <nospam-abuse.DeleteThis@ilyaz.org> wrote:
>[A complimentary Cc of this posting was sent to
>Roger N. Clark (change username to rnclark)
><username.DeleteThis@qwest.net>], who wrote in article <47A6B0B2.8040005.DeleteThis@qwest.net>:
>> If you haven't visited this month, see:
>>
>> http://www.clarkvision.com/imagedetail/digital.sensor.performance.summary
>>
>> I've added data for several newer cameras
>
>Thanks.
>
>What is very surprizing is how large is the improvement due to switch
>from 12-bit ADC to 14-bit: not *exactly* zero, but only 11%!
>
> Camera Full well Noise at ISO50 Dynamic range Sensel
> --------------------------------------------------------------------
> Canon 1DMII CMOS 79,900* 30.6 2611 8.2um
> Canon 1DMIII CMOS 70,500 24.4 2889 7.2um
>
>Most probably the "improvement" in noise is just due to lower area of
>a sensel, thus lower capacitance, thus larger voltage generated by
>each electron. So it looks like 14-bitness is 93% hoax: the ADC
>behaved as an ideal 10-bit ADC before, now it behaves as 10.15-bit...

Changing from a 12 bit ADC to a 14 bit ADC does *not* change the
dynamic range of the sensor.

--
Floyd L. Davidson <http://www.apaflo.com/floyd_davidson>
Ukpeagvik (Barrow, Alaska) floyd.DeleteThis@apaflo.com
 >> Stay informed about: Digital Camera Sensor Performane Summary updated Feb 3 

[A complimentary Cc of this posting was sent to
Roger N. Clark (change username to rnclark)
<username RemoveThis @qwest.net>], who wrote in article <47AA92F0.2010006 RemoveThis @qwest.net>:
> > Camera Full well Noise at ISO50 Dynamic range Sensel
> > --------------------------------------------------------------------
> > Canon 1DMII CMOS 79,900* 30.6 2611 8.2um
> > Canon 1DMIII CMOS 70,500 24.4 2889 7.2um
> >
> > Most probably the "improvement" in noise is just due to lower area of
> > a sensel, thus lower capacitance, thus larger voltage generated by
> > each electron. So it looks like 14-bitness is 93% hoax: the ADC
> > behaved as an ideal 10-bit ADC before, now it behaves as 10.15-bit...

> Try looking at more of the picture. The smaller pixel
> has less photons, so if the 1DIII used the same 12-bit ADC
> as in the 1DII, the higher read noise and lower full well
> capacity would result in a dynamic range 11% SMALLER than
> the 1DII

It would help if you FINALLY get some clue about the relation of pixel
size to capacitance to voltage.

Photons do not come into the context of this table at all. Full well
is measured in electrons; likewise for the noise. An electron
added to a smaller sensel will generate higher voltage than when added
to a larger sensel.

> so overall, the 1DIII 14-bit ADC and read noise
> has improved about 22% at low ISOs.

As I said, with CORRECT arithmetic there is no improvement.

> Look at higher ISOs, and the 1DIII read noise drops to
> 2.1 electrons compared to the 1DII 3.9 electrons, about
> an 86% improvement, which helps high ISO imaging.

You know that THIS has nothing to do with high bit count. In the
regime of the full well of about 4K electrons, 2 electron noise is
"approximately 9-bits".

Hope this helps,
Ilya
 >> Stay informed about: Digital Camera Sensor Performane Summary updated Feb 3 

Ilya Zakharevich wrote:
> [A complimentary Cc of this posting was sent to
> Roger N. Clark (change username to rnclark)
> <username RemoveThis @qwest.net>], who wrote in article <47AA92F0.2010006 RemoveThis @qwest.net>:
>>> Camera Full well Noise at ISO50 Dynamic range Sensel
>>> --------------------------------------------------------------------
>>> Canon 1DMII CMOS 79,900* 30.6 2611 8.2um
>>> Canon 1DMIII CMOS 70,500 24.4 2889 7.2um
>>>
>>> Most probably the "improvement" in noise is just due to lower area of
>>> a sensel, thus lower capacitance, thus larger voltage generated by
>>> each electron. So it looks like 14-bitness is 93% hoax: the ADC
>>> behaved as an ideal 10-bit ADC before, now it behaves as 10.15-bit...
>
>> Try looking at more of the picture. The smaller pixel
>> has less photons, so if the 1DIII used the same 12-bit ADC
>> as in the 1DII, the higher read noise and lower full well
>> capacity would result in a dynamic range 11% SMALLER than
>> the 1DII
>
> It would help if you FINALLY get some clue about the relation of pixel
> size to capacitance to voltage.
>
> Photons do not come into the context of this table at all. Full well
> is measured in electrons; likewise for the noise. An electron
> added to a smaller sensel will generate higher voltage than when added
> to a larger sensel.
>
>> so overall, the 1DIII 14-bit ADC and read noise
>> has improved about 22% at low ISOs.
>
> As I said, with CORRECT arithmetic there is no improvement.
>
>> Look at higher ISOs, and the 1DIII read noise drops to
>> 2.1 electrons compared to the 1DII 3.9 electrons, about
>> an 86% improvement, which helps high ISO imaging.
>
> You know that THIS has nothing to do with high bit count. In the
> regime of the full well of about 4K electrons, 2 electron noise is
> "approximately 9-bits".
>
> Hope this helps,
> Ilya
>
>
BS
 >> Stay informed about: Digital Camera Sensor Performane Summary updated Feb 3 

[A complimentary Cc of this posting was NOT [per weedlist] sent to
Floyd L. Davidson
<floyd.RemoveThis@apaflo.com>], who wrote in article <878x1xxqyb.fld.RemoveThis@apaflo.com>:

> Changing from a 12 bit ADC to a 14 bit ADC does *not* change the
> dynamic range of the sensor.

Thanks, but no thanks. I'm not interested in definition games.

What it COULD change was the dynamic range of the whole sensor/ADC etc
pipeline. But it did not.

Hope this helps,
Ilya
 >> Stay informed about: Digital Camera Sensor Performane Summary updated Feb 3 

On Thu, 07 Feb 2008 19:16:33 -0700, "Roger N. Clark (change username to
rnclark)" <username.TakeThisOut@qwest.net> wrote in <47ABBB81.4050806.TakeThisOut@qwest.net>:

>Ilya Zakharevich wrote:

>> It would help if you FINALLY get some clue about the relation of pixel
>> size to capacitance to voltage.
>>
>> Photons do not come into the context of this table at all. Full well
>> is measured in electrons; likewise for the noise. An electron
>> added to a smaller sensel will generate higher voltage than when added
>> to a larger sensel.

>> As I said, with CORRECT arithmetic there is no improvement.

>> You know that THIS has nothing to do with high bit count. In the
>> regime of the full well of about 4K electrons, 2 electron noise is
>> "approximately 9-bits".

>BS

That may be, but I see quite a few unsupported assumptions in your work,
so as interesting as it may be, I can't accept it at face value.

--
Best regards,
John Navas
Panasonic DMC-FZ8 (and several others)
 >> Stay informed about: Digital Camera Sensor Performane Summary updated Feb 3