DSLR optimum ISO question

"Ron Recer" <ron48 RemoveThis @aol.com> wrote in
news:5eaqhsF37nabhU1@mid.individual.net:

> I shoot RAW images with a 10D and make 8 1/2" x 11" prints. After
> some testing and discussion with others, I shoot almost everything at
> ISO 400. I am sure you can see the difference with larger prints and
> some may notice a difference at 8 1/2" x 11", but I usually don't.

That points up one of the dilemnas with digital cameras; you want clean
ISOs, to get more usable shadows, for instance, but if the midtones and
highlights are too clean, they tend to posterize when converted to 8-bit
display formats. Sometimes blue sky, for instance, looks worse at ISO 100
than 400, because the noise of the 400 dithers the slight variations in the
atmosphere while the ISO 100 makes it block up into posterized patches in
just the red channel. It is my opinion that RAW converters (or image
processing programs) should have an optional feature to add linearly-
distributed noise to areas that are candidates for posterization in 8-bit
display. It should be applied only as necessary, by evaluating small areas
to see how much noise, if any, needs to be added to avoid posterization.

--

<>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<>
John P Sheehy <JPS RemoveThis @no.komm>
><<> <>>< <>>< ><<> <>>< ><<> ><<> <>><
 >> Stay informed about: DSLR optimum ISO question 

John Sheehy wrote:
> "Ron Recer" <ron48 DeleteThis @aol.com> wrote in
> news:5eaqhsF37nabhU1@mid.individual.net:
>
>> I shoot RAW images with a 10D and make 8 1/2" x 11" prints. After
>> some testing and discussion with others, I shoot almost everything at
>> ISO 400. I am sure you can see the difference with larger prints and
>> some may notice a difference at 8 1/2" x 11", but I usually don't.
>
> That points up one of the dilemnas with digital cameras; you want clean
> ISOs, to get more usable shadows, for instance, but if the midtones and
> highlights are too clean, they tend to posterize when converted to 8-bit
> display formats. Sometimes blue sky, for instance, looks worse at ISO 100
> than 400, because the noise of the 400 dithers the slight variations in the
> atmosphere while the ISO 100 makes it block up into posterized patches in
> just the red channel. It is my opinion that RAW converters (or image
> processing programs) should have an optional feature to add linearly-
> distributed noise to areas that are candidates for posterization in 8-bit
> display. It should be applied only as necessary, by evaluating small areas
> to see how much noise, if any, needs to be added to avoid posterization.
>
And it is amazing how how few bits you need it you carefully add noise.
This image using only 6 bits, and yet the sky looks pretty clean and
smooth, I use error diffusion dithering for the image.

http://www.sewcon.com/temp/6bit.tif

If you download it and then open in an editor and check the histogram
it will see just how few levels are used.

Scott
 >> Stay informed about: DSLR optimum ISO question 

Eatmorepies wrote:
> Hello
>
> A friend of mine told me that Canon DSLR sensors were optimised at ISO200
> and there is no noise difference between ISO100 ans ISO200 - I've Googled
> about for technical references on this but failed to find any.
>
> Anyone got a reference?

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

http://www.clarkvision.com/imagedetail/index.html#sensor_analysis

http://www.clarkvision.com/astro/canon-10d-signal-to-noise

There are differences between ISO 100 versus 200 (despite
what some may have posted here), but it depends on many
factors, including brightness in the scene, exposure
time and sensor temperature.

Roger
 >> Stay informed about: DSLR optimum ISO question 

"Eatmorepies" <xyztnday.RemoveThis@lineone.net> wrote in
news:46802fa6_1@mk-nntp-2.news.uk.tiscali.com:

> Hello
>
> A friend of mine told me that Canon DSLR sensors were optimised at
> ISO200 and there is no noise difference between ISO100 ans ISO200 -
> I've Googled about for technical references on this but failed to find
> any.

The total read noise, including read noise at the photosite, through the
final noises in the ADC, are very similar at ISOs 100 and 200 on Canon
DSLRs (and 400 can be very close as well). By ISO 800 and up, the total
read noise tends to increase more rapidly.

Total read noise is only part of the story, though. The other main
component of noise, photon shot noise, has nothing at all to do with the
camera's electronics, and only to do with the amount of photons
collected. A doubling of photons collected means 71% the relative shot
noise, so relative shot noise at ISO 100 is always going to be 71% of
what it is at ISO 200, 50% of what it is at ISO 400, 35.5% of what it is
at ISO 800, 25% of what it is at ISO 1600, etc.

The highlights generally have mainly shot noise, but may have so little
that you don't notice it, at low ISOs. The higher the ISO, the brighter
the tones at which shot noise has a visible effect. Total read noise is
generally small, but is blanketly applied at all tonal ranges, and
therefore affect mainly the deeper shadows and the higher ISOs the most.
With ISOs 100 and 200, there will generally be a small zone in the shadow
ranges where the ISO 200 will have more visible noise. The brighter
areas will have more absolute noise, but relative to signal it will be
small, and barely visible.

Now, some cameras don't have deep enough sensor wells to support ISO 100
correctly, like the Canon 20D and 30D. The highlights clip about 1/4
stop lower than they do with ISO 200, so ISO 200 actually has slightly
more dynamic range with a liberal definition of dynamic range (max signal
divided by the total read noise floor). I don't use ISO 100 on my 20D or
30D. It's broken, AFAIAC. On my 30D, ISO 160 is actually the ISO with
the highest (liberal) dynamic range (2000:1 - that's not rounded off;
that's exactly 2000, to the nearest integer), and that is what I use when
light is ample enough.


--

<>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<>
John P Sheehy <JPS.RemoveThis@no.komm>
><<> <>>< <>>< ><<> <>>< ><<> ><<> <>><
 >> Stay informed about: DSLR optimum ISO question 

I think it is a fair summary to say that every digital camera I've
tested produces the lowest noise at its lowest ISO setting. There can be
and often is little visible increase in noise between the lowest
(usually 100ISO) and the next one (200ISO) in normal use, but that does
not mean that there is no actual increase in noise, just that you won't
typically see it in normal use.

On my site when I test cameras I put RAW files shot at the range of ISO
settings the camera can make. You can download these and look for yourself:
http://www.dimagemaker.com/specials/cameras/camtests.php

Cheers,

Wayne

Wayne J. Cosshall
Publisher, The Digital ImageMaker, http://www.dimagemaker.com/
Blog http://www.digitalimagemakerworld.com/
Publisher, Experimental Digital Photography
http://www.experimentaldigitalphotography.com
Personal art site http://www.cosshall.com/



Eatmorepies wrote:
> Hello
>
> A friend of mine told me that Canon DSLR sensors were optimised at ISO200
> and there is no noise difference between ISO100 ans ISO200 - I've Googled
> about for technical references on this but failed to find any.
>
> Anyone got a reference?
>
> Thanks
>
>
> John
>
>
 >> Stay informed about: DSLR optimum ISO question 

"Wayne J. Cosshall" <wayne RemoveThis @dimagemaker.com> wrote in message
news:468031ba$0$32587$afc38c87@news.optusnet.com.au...
>I think it is a fair summary to say that every digital camera I've tested
>produces the lowest noise at its lowest ISO setting. There can be and often
>is little visible increase in noise between the lowest (usually 100ISO) and
>the next one (200ISO) in normal use, but that does not mean that there is
>no actual increase in noise, just that you won't typically see it in normal
>use.

With exceptions, like the Canon 5D, which has lower image quality at ISO50
than at ISO200.

>
> On my site when I test cameras I put RAW files shot at the range of ISO
> settings the camera can make. You can download these and look for
> yourself:
> http://www.dimagemaker.com/specials/cameras/camtests.php
>
> Cheers,
>
> Wayne
>
> Wayne J. Cosshall
> Publisher, The Digital ImageMaker, http://www.dimagemaker.com/
> Blog http://www.digitalimagemakerworld.com/
> Publisher, Experimental Digital Photography
> http://www.experimentaldigitalphotography.com
> Personal art site http://www.cosshall.com/
>
>
>
> Eatmorepies wrote:
>> Hello
>>
>> A friend of mine told me that Canon DSLR sensors were optimised at ISO200
>> and there is no noise difference between ISO100 ans ISO200 - I've Googled
>> about for technical references on this but failed to find any.
>>
>> Anyone got a reference?
>>
>> Thanks
>>
>>
>> John
 >> Stay informed about: DSLR optimum ISO question 

"John Sheehy" <JPS DeleteThis @no.komm> wrote:
> "Ron Recer" <ron48 DeleteThis @aol.com> wrote:
>
>> I shoot RAW images with a 10D and make 8 1/2" x 11" prints. After
>> some testing and discussion with others, I shoot almost everything at
>> ISO 400. I am sure you can see the difference with larger prints and
>> some may notice a difference at 8 1/2" x 11", but I usually don't.
>
> That points up one of the dilemnas with digital cameras; you want clean
> ISOs, to get more usable shadows, for instance, but if the midtones and
> highlights are too clean, they tend to posterize when converted to 8-bit
> display formats. Sometimes blue sky, for instance, looks worse at ISO 100
> than 400, because the noise of the 400 dithers the slight variations in
> the
> atmosphere while the ISO 100 makes it block up into posterized patches in
> just the red channel.

Do you have any (or know of any) examples of this online? (I've never
noticed it and am curious as to what I'm missing.)

David J. Littleboy
Tokyo, Japan
 >> Stay informed about: DSLR optimum ISO question 

"David J. Littleboy" <davidjl.RemoveThis@gol.com> wrote in
news:f5sj46$brd$2@nnrp.gol.com:

> Do you have any (or know of any) examples of this online? (I've never
> noticed it and am curious as to what I'm missing.)

I am going by memory.

Many people have posted questions on usenet and digital photography web
forums about how they get the noise in the sky, even though they were
shooting at ISO 100, and close inspection shows that the red channel is
blotchy. At higher ISOs, the noise in the sky is usually finer.

When I get out on a blue sky day, I'll try to remember to do a series
through the ISOs, and make a collage of the results on my pbase pages.

--

<>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<>
John P Sheehy <JPS.RemoveThis@no.komm>
><<> <>>< <>>< ><<> <>>< ><<> ><<> <>><
 >> Stay informed about: DSLR optimum ISO question