Sign in with
Sign up | Sign in
Your question

Optimal size of photosensor

Last response: in Digital Camera
Share
Anonymous
February 10, 2005 6:02:12 AM

Archived from groups: rec.photo.digital (More info?)

For a fixed number of pixels, larger sensors give greater sensitivity. I
read somewhere that due to "other factors" the optimal sensor size was
not 35 mm format (24 m x 36mm) but rather something about 2/3 this size.
Does anyone know of a web site that discusses this "optimal sensor size"
issue? Thanks.
Anonymous
February 10, 2005 6:02:13 AM

Archived from groups: rec.photo.digital (More info?)

Mel Comisarow wrote:

> For a fixed number of pixels, larger sensors give greater sensitivity. I
> read somewhere that due to "other factors" the optimal sensor size was
> not 35 mm format (24 m x 36mm) but rather something about 2/3 this size.
> Does anyone know of a web site that discusses this "optimal sensor size"
> issue? Thanks.

Signal-to-noise, the larger sensor wins:

Digital Cameras: Does Pixel Size Matter?
Factors in Choosing a Digital Camera:
http://clarkvision.com/imagedetail/does.pixel.size.matt...

The question about whether more pixels in the same area,
is a different one and is not so clear. But if you
mean like 8 megapixels in 35mm format versus 8 megapixels
in 2/3 the 35mm size, then the larger sensor wins in general.
But if your lens was the resolution and you want telephoto
reach then the smaller sensor might make sense. But given
lenses covering the same field of view, the larger sensor
wins, just like with film: large format cameras produce better
images than medium format, which produce better images than
35mm. (The are always exceptions for some specific restricted
cases.)

Roger
Anonymous
February 10, 2005 8:07:30 AM

Archived from groups: rec.photo.digital (More info?)

Mel Comisarow <melcom@shaw.ca> wrote in news:melcom-6B311F.19021209022005
@news.vc.shawcable.net:

> For a fixed number of pixels, larger sensors give greater sensitivity.

Less noise, that is.

> I
> read somewhere that due to "other factors" the optimal sensor size was
> not 35 mm format (24 m x 36mm) but rather something about 2/3 this size.

Someone was lying to you (Nikon, maybe?). The optimal size of a sensor on a
DSLR is as large as possible, which is the same size as a 35mm film frame
since anything larger would require new lenses.

Even better, from an image quality point of view, are digital backs for
medium-format cameras (or something like the Mimaya digital MF). But these
are a different type of camera altogether.

<snip>
Related resources
Anonymous
February 10, 2005 10:23:46 AM

Archived from groups: rec.photo.digital (More info?)

Since the sensor size choice is always a compromise there is no such
thing as an optimal size. There are sizes that will work better for
different uses but no one optimal size that would work for everything.
Large sensors are both more sensitive and have better dynamic range
then smaller ones but try doing telephoto work using a 4 x 5 camera,
the lens would just be too large.

The optimal size is also dependent on how many pixels you are talking
about. If you are talking about 3 MP then it is much smaller then if
you are talking about 16 MP.

Scott
Anonymous
February 10, 2005 11:02:43 AM

Archived from groups: rec.photo.digital (More info?)

Don Stauffer in Minneapolis wrote:
> Mel Comisarow wrote:
>
> > For a fixed number of pixels, larger sensors give greater
sensitivity. I
> > read somewhere that due to "other factors" the optimal sensor size
was
> > not 35 mm format (24 m x 36mm) but rather something about 2/3 this
size.
> > Does anyone know of a web site that discusses this "optimal sensor
size"
> > issue? Thanks.
>
>
> Well, it was always my philosphy that one could not beat format size.

> Of course, that was for space photography and sensing, where the
> customer had deep pockets. Whether film or electronic photography
there
> is almost never a drawback to using bigger sensors except in certain
> cases where the job is detection of an object instead of imaging it.
>
> Economics are something else. If one has to dedicate a 12 inch wafer
of
> silicon to ONE focal plane chip, the cost would be astronomical.
You might also want to give some thought to the lens you would need for
this.
For me I need a camera that I can carry with me and I would like to be
able to hold the camera with the lens on it.
Anonymous
February 10, 2005 11:08:33 AM

Archived from groups: rec.photo.digital (More info?)

Eric Gill wrote:
> "Scott W" <biphoto@hotmail.com> wrote in news:1108049026.444600.76450
> @o13g2000cwo.googlegroups.com:
>
> > Since the sensor size choice is always a compromise there is no
such
> > thing as an optimal size.
>
> Sorry, bud - that conclusion doesn't follow, since "optimal" is
subjective
> anyways.
>
> <snip>

I think that was in fact my point, if optimal is subjective then there
will not be one number that would meet the needs of every person.

So to say 2/3 of 35mm is optimal does not make any more sense then to
say a full 35mm is better then 2/3. It depends on your needs.

Scott
Anonymous
February 10, 2005 12:37:13 PM

Archived from groups: rec.photo.digital (More info?)

Mel Comisarow wrote:

> For a fixed number of pixels, larger sensors give greater sensitivity. I
> read somewhere that due to "other factors" the optimal sensor size was
> not 35 mm format (24 m x 36mm) but rather something about 2/3 this size.
> Does anyone know of a web site that discusses this "optimal sensor size"
> issue? Thanks.


Well, it was always my philosphy that one could not beat format size.
Of course, that was for space photography and sensing, where the
customer had deep pockets. Whether film or electronic photography there
is almost never a drawback to using bigger sensors except in certain
cases where the job is detection of an object instead of imaging it.

Economics are something else. If one has to dedicate a 12 inch wafer of
silicon to ONE focal plane chip, the cost would be astronomical.
Anonymous
February 10, 2005 12:45:30 PM

Archived from groups: rec.photo.digital (More info?)

"If one has to dedicate a 12 inch wafer of
silicon to ONE focal plane chip, the cost would be astronomical."

Depends on how many defective pixels one can tollerate.
Anonymous
February 10, 2005 5:48:37 PM

Archived from groups: rec.photo.digital (More info?)

On 10 Feb 2005 08:02:43 -0800, "Scott W" <biphoto@hotmail.com> wrote:

snipped
>For me I need a camera that I can carry with me and I would like to be
>able to hold the camera with the lens on it.

But that may not be every photographer's need. Look at the people who
use an Linhof 8X10 and aren't concerned with carrying it around but in
the most limited sense.


*****************************************************

"Vietnam is what we had instead of happy childhoods."

Tim Page in
"Dispatches"
by Michael Herr
Anonymous
February 10, 2005 6:30:55 PM

Archived from groups: rec.photo.digital (More info?)

"Scott W" <biphoto@hotmail.com> wrote in news:1108049026.444600.76450
@o13g2000cwo.googlegroups.com:

> Since the sensor size choice is always a compromise there is no such
> thing as an optimal size.

Sorry, bud - that conclusion doesn't follow, since "optimal" is subjective
anyways.

<snip>
Anonymous
February 10, 2005 11:56:24 PM

Archived from groups: rec.photo.digital (More info?)

On Wednesday 09 February 2005 19:02, Mel Comisarow wrote:

> For a fixed number of pixels, larger sensors give greater sensitivity.
> I read somewhere that due to "other factors" the optimal sensor size
> was not 35 mm format (24 m x 36mm) but rather something about 2/3 this
> size. Does anyone know of a web site that discusses this "optimal
> sensor size"
> issue? Thanks.

What is "optimal" would depend entirely on your photographic
requirements. Are you looking for the most sensitivity, the least
noise, the greatest "sharpness" or dynamic range, etc? Do you intend
to make poster size art prints or are the photos for only print media
-- newspapers, magazines, brochures, etc?

In my case, the 3.1 MP, APS-C sized, CMOS sensor in the Canon D30 was
"optimal." Your mileage may vary.

--
Stefan Patric
NoLife Polymath Group
tootek2@yahoo.com
Anonymous
February 11, 2005 12:23:14 AM

Archived from groups: rec.photo.digital (More info?)

"Mel Comisarow" <melcom@shaw.ca> wrote in message
news:melcom-6B311F.19021209022005@news.vc.shawcable.net...
> For a fixed number of pixels, larger sensors give greater sensitivity. I
> read somewhere that due to "other factors" the optimal sensor size was
> not 35 mm format (24 m x 36mm) but rather something about 2/3 this size.
> Does anyone know of a web site that discusses this "optimal sensor size"
> issue? Thanks.

I'm sure Hasselbalad H-1 users would disagree with that.

KB
Anonymous
February 11, 2005 1:42:37 AM

Archived from groups: rec.photo.digital (More info?)

Roger N. Clark (change username to rnclark) wrote:
> Signal-to-noise, the larger sensor wins:
>
> Digital Cameras: Does Pixel Size Matter?
> Factors in Choosing a Digital Camera:
> http://clarkvision.com/imagedetail/does.pixel.size.matt...
>
> The question about whether more pixels in the same area,
> is a different one and is not so clear. But if you
> mean like 8 megapixels in 35mm format versus 8 megapixels
> in 2/3 the 35mm size, then the larger sensor wins in general.
> But if your lens was the resolution and you want telephoto
> reach then the smaller sensor might make sense. But given
> lenses covering the same field of view, the larger sensor
> wins, just like with film: large format cameras produce better
> images than medium format, which produce better images than
> 35mm. (The are always exceptions for some specific restricted
> cases.)
What you put silently in this argument is that you assume the larger sensor,
film or ccd, is combined with a lens capable to illuminate the
entire sensor, and that the aperture and view angle are constants. So with
the same aperture and view angle bigger sensors mean bigger front lens
obviously capturing more light from the object. So indeed in total a bigger
sensor catches more photons. A second issue is over how many cells you spread
these photons. If you keep the amount independent of the sensor size there
will be more photons per cell: less noise. That would be a waste of optical
information, but killing optical information in anti aliasing filters is
standard practice and fodder for a new thread.
There could be also factors in the manufacturing process of the ccd which
have an impact on photon efficiency, but IIRC a cell size of 6-7 micron is
standard in ultra high efficiency cooled ccd cameras. This suggests bigger
cells by themselves do not lead to higher efficiency.

But there is complicating factor in this reasoning: the cost of the lens
explodes as the sensor size grows.
Vice versa, lenses for small sensors are relatively cheap. Or, a fairly high
aperture lens can be made at 'reasonable' cost. For example, take the
28-200mm (35mm equivalent) f 2.0 zoom lens on the Sony F828. What would a
true 28-200mm f2.0 zoom lens of equivalent quality for 35mm cost?
To put it differently: the possibility to build larger aperture lenses for a
low price compensates for the loss of sensitivity for small sensors.
The useful smallest size of the ccd elements is ultimately limited by
diffraction, so there is a hard limit here.
*Somewhere* there must be an optimum, but it is greatly dependent on factors
like intended use, budget, copy number of lenses sold, etc, etc.

Cheers, Hans
Anonymous
February 11, 2005 1:42:38 AM

Archived from groups: rec.photo.digital (More info?)

HvdV wrote:

> Roger N. Clark (change username to rnclark) wrote:
>
>> Signal-to-noise, the larger sensor wins:
>>
>> Digital Cameras: Does Pixel Size Matter?
>> Factors in Choosing a Digital Camera:
>> http://clarkvision.com/imagedetail/does.pixel.size.matt...
>>
>> The question about whether more pixels in the same area,
>> is a different one and is not so clear. But if you
>> mean like 8 megapixels in 35mm format versus 8 megapixels
>> in 2/3 the 35mm size, then the larger sensor wins in general.
>> But if your lens was the resolution and you want telephoto
>> reach then the smaller sensor might make sense. But given
>> lenses covering the same field of view, the larger sensor
>> wins, just like with film: large format cameras produce better
>> images than medium format, which produce better images than
>> 35mm. (The are always exceptions for some specific restricted
>> cases.)
>
> What you put silently in this argument is that you assume the larger
> sensor, film or ccd, is combined with a lens capable to illuminate the
> entire sensor, and that the aperture and view angle are constants. So
> with the same aperture and view angle bigger sensors mean bigger front
> lens obviously capturing more light from the object. So indeed in total
> a bigger sensor catches more photons. A second issue is over how many
> cells you spread these photons. If you keep the amount independent of
> the sensor size there will be more photons per cell: less noise. That
> would be a waste of optical information, but killing optical information
> in anti aliasing filters is standard practice and fodder for a new thread.
> There could be also factors in the manufacturing process of the ccd
> which have an impact on photon efficiency, but IIRC a cell size of 6-7
> micron is standard in ultra high efficiency cooled ccd cameras. This
> suggests bigger cells by themselves do not lead to higher efficiency.
>
> But there is complicating factor in this reasoning: the cost of the lens
> explodes as the sensor size grows.
> Vice versa, lenses for small sensors are relatively cheap. Or, a fairly
> high aperture lens can be made at 'reasonable' cost. For example, take
> the 28-200mm (35mm equivalent) f 2.0 zoom lens on the Sony F828. What
> would a true 28-200mm f2.0 zoom lens of equivalent quality for 35mm cost?
> To put it differently: the possibility to build larger aperture lenses
> for a low price compensates for the loss of sensitivity for small sensors.
> The useful smallest size of the ccd elements is ultimately limited by
> diffraction, so there is a hard limit here.
> *Somewhere* there must be an optimum, but it is greatly dependent on
> factors like intended use, budget, copy number of lenses sold, etc, etc.
>
> Cheers, Hans
Hans,
You have thrown cost into optimum, which may or may
not have been in the OPs question (it was not specified).
If you define the problem of megapixels*signal-to-noise*cost,
then maybe there is such an optimum. Without cost,
bigger produces better images except in very restrictive
circumstances.

Someone made an argument you wouldn't do wildlife
photography with a 4x5. Sure I would, if I could
afford the system, and could focus easily,
and in fact I have done some wildlife with my 4x5.

You talk of "35mm equivalent." That implies scaling
the lens with the sensor. The Sony F828 has extremely
poor signal-to-noise with its 2.7-micron pixels.
There is no comparison to a DSLR. See:
http://www.dpreview.com/reviews/sonydscf828/page14.asp

It shows the worst noise of all those shown.
There is no optimum with the F828 in my opinion,
merely a reduction in cost but make it look
good on paper if you cover up the noise problems
that small sensors have.

Roger
Anonymous
February 11, 2005 11:31:09 AM

Archived from groups: rec.photo.digital (More info?)

Marvin wrote:
> I said much the same thing in other words. and nobody even bothered
to comment on it. Most folks seem to focus on a single
> factor, but a fact of engineering is that all choices have
trade-offs.

I must be missing something because I don't see a pervious post from
you on this thread.
Anonymous
February 11, 2005 11:50:22 AM

Archived from groups: rec.photo.digital (More info?)

Scott W wrote:

> Don Stauffer in Minneapolis wrote:
>
>>Mel Comisarow wrote:
>>
>>
>>>For a fixed number of pixels, larger sensors give greater
>
> sensitivity. I
>
>>>read somewhere that due to "other factors" the optimal sensor size
>
> was
>
>>>not 35 mm format (24 m x 36mm) but rather something about 2/3 this
>
> size.
>
>>>Does anyone know of a web site that discusses this "optimal sensor
>
> size"
>
>>>issue? Thanks.
>>
>>
>>Well, it was always my philosphy that one could not beat format size.
>
>
>>Of course, that was for space photography and sensing, where the
>>customer had deep pockets. Whether film or electronic photography
>
> there
>
>>is almost never a drawback to using bigger sensors except in certain
>>cases where the job is detection of an object instead of imaging it.
>>
>>Economics are something else. If one has to dedicate a 12 inch wafer
>
> of
>
>>silicon to ONE focal plane chip, the cost would be astronomical.
>
> You might also want to give some thought to the lens you would need for
> this.
> For me I need a camera that I can carry with me and I would like to be
> able to hold the camera with the lens on it.
>
Good point. The studies I referred to were for cameras or sensors
carried by spacecraft and aircraft. Yeah, though I have a couple of
large format film cameras I do not use them for ordinary photography,
only when I have a specific subject that requires them. I do intend to
start using my wife's old 2-1/4 square, though- she no longer uses it.
My new, lighter tripod will offset higher weight of heavier camera :-)
Anonymous
February 11, 2005 11:54:33 AM

Archived from groups: rec.photo.digital (More info?)

Ken Burns wrote:

> "Mel Comisarow" <melcom@shaw.ca> wrote in message
> news:melcom-6B311F.19021209022005@news.vc.shawcable.net...
>
>>For a fixed number of pixels, larger sensors give greater sensitivity. I
>>read somewhere that due to "other factors" the optimal sensor size was
>>not 35 mm format (24 m x 36mm) but rather something about 2/3 this size.
>>Does anyone know of a web site that discusses this "optimal sensor size"
>>issue? Thanks.
>
>
> I'm sure Hasselbalad H-1 users would disagree with that.
>
> KB
>
>
Nor the Hasselblad 500C users even :-)
February 11, 2005 2:25:09 PM

Archived from groups: rec.photo.digital (More info?)

HvdV wrote:
> Roger N. Clark (change username to rnclark) wrote:
>
<snip>
>
> What you put silently in this argument is that you assume the larger
> sensor, film or ccd, is combined with a lens capable to illuminate the
> entire sensor, and that the aperture and view angle are constants. So
> with the same aperture and view angle bigger sensors mean bigger front
> lens obviously capturing more light from the object. So indeed in total
> a bigger sensor catches more photons. A second issue is over how many
> cells you spread these photons. If you keep the amount independent of
> the sensor size there will be more photons per cell: less noise. That
> would be a waste of optical information, but killing optical information
> in anti aliasing filters is standard practice and fodder for a new thread.
> There could be also factors in the manufacturing process of the ccd
> which have an impact on photon efficiency, but IIRC a cell size of 6-7
> micron is standard in ultra high efficiency cooled ccd cameras. This
> suggests bigger cells by themselves do not lead to higher efficiency.
>

I said much the same thing in other words. and nobody even bothered to comment on it. Most folks seem to focus on a single
factor, but a fact of engineering is that all choices have trade-offs.
February 11, 2005 2:28:44 PM

Archived from groups: rec.photo.digital (More info?)

Eric Gill wrote:
> "Scott W" <biphoto@hotmail.com> wrote in news:1108049026.444600.76450
> @o13g2000cwo.googlegroups.com:
>
>
>>Since the sensor size choice is always a compromise there is no such
>>thing as an optimal size.
>
>
> Sorry, bud - that conclusion doesn't follow, since "optimal" is subjective
> anyways.
>
"Optimal" can be defined so that it isn't subjective. It takes some effort, and most folks don't seem willing to make the
effort. In properly-done product development, there are design goals that define "optimal", and they almost always include cost.
Anonymous
February 12, 2005 1:58:11 AM

Archived from groups: rec.photo.digital (More info?)

Hi Roger,
<snip>
>> But there is complicating factor in this reasoning: the cost of the
>> lens explodes as the sensor size grows.
>> Vice versa, lenses for small sensors are relatively cheap. Or, a
>> fairly high aperture lens can be made at 'reasonable' cost. For
>> example, take the 28-200mm (35mm equivalent) f 2.0 zoom lens on the
>> Sony F828. What would a true 28-200mm f2.0 zoom lens of equivalent
>> quality for 35mm cost?
>> To put it differently: the possibility to build larger aperture lenses
>> for a low price compensates for the loss of sensitivity for small
>> sensors.
>> The useful smallest size of the ccd elements is ultimately limited by
>> diffraction, so there is a hard limit here.
>> *Somewhere* there must be an optimum, but it is greatly dependent on
>> factors like intended use, budget, copy number of lenses sold, etc, etc.
>>
>> Cheers, Hans
>
> Hans,
> You have thrown cost into optimum, which may or may
> not have been in the OPs question (it was not specified).
> If you define the problem of megapixels*signal-to-noise*cost,
> then maybe there is such an optimum. Without cost,
> bigger produces better images except in very restrictive
> circumstances.
Agreed. But lens 'cost' or the possibility to build *is* a factor. OTOH as
lenses for waversteppers prove with enough $$ one can come an astonishing way.
>
> Someone made an argument you wouldn't do wildlife
> photography with a 4x5. Sure I would, if I could
> afford the system, and could focus easily,
> and in fact I have done some wildlife with my 4x5.
I guess a fast 4x5 tele, if you can afford it, is very very heavy...
>
> You talk of "35mm equivalent." That implies scaling
> the lens with the sensor. The Sony F828 has extremely
> poor signal-to-noise with its 2.7-micron pixels.
> There is no comparison to a DSLR. See:
> http://www.dpreview.com/reviews/sonydscf828/page14.asp
Hm, have a look at the low frequency noise components, very suspicious,
doesn't look like pure Poisson noise. The Canon has 7.4micron pixels, 7.5
times the surface of the Sony. It's a bit of work to measure signal to noise
from those pictures, but IMO the 7.5 factor largely accounts for the noise
difference. If you then further compare the Sigma 55-200 mm F4-5,6 lens for
the canon to the Sony's 28-200 f 2 lens you see the Sony lens is 2-3 stops
(4x-8x) faster, roughly compensating the lower sensitivity. I don't know
whether these lenses are of equal cost though.
>
> It shows the worst noise of all those shown.
> There is no optimum with the F828 in my opinion,
> merely a reduction in cost but make it look
> good on paper if you cover up the noise problems
> that small sensors have.
What I wonder about is whether the low sensitivity is fully due to the
geometrical optical scaling effect, or a property of the sensor itself, i.e.
lower quantum efficiency, higher readout noise, ineffective microlenses, etc.
Well, if cost is not a factor one could pile some Peltier coolers on the chip
:-).

Cheers, Hans
Anonymous
February 12, 2005 9:01:49 PM

Archived from groups: rec.photo.digital (More info?)

>>>
>>>> For a fixed number of pixels, larger sensors give greater
>>
>>
>> sensitivity. I
>

Is there a limit on sensitivity benefits?

When the sensor can tell 1 from 0 photons (or 25001 from 25000), that's
pretty much the limit of sensitivity.

The practical limit may be way lower tho. Excessive sensitivity seems
to invite noisiness.
Anonymous
February 12, 2005 10:54:09 PM

Archived from groups: rec.photo.digital (More info?)

Jack Zeal wrote:

>>>>
>>>>> For a fixed number of pixels, larger sensors give greater
>>>
>>>
>>>
>>> sensitivity. I
>>
>>
>
> Is there a limit on sensitivity benefits?
>
> When the sensor can tell 1 from 0 photons (or 25001 from 25000), that's
> pretty much the limit of sensitivity.
>
> The practical limit may be way lower tho. Excessive sensitivity seems
> to invite noisiness.

You will never know 25001 from 25000. Photon counting statistics
is the fundamental limit, so with 25000 photons, the noise is
square root 25000 or 158 photons.

Modern digital cameras are photon noise limited at high
signal levels, and read noise limited at low signal levels.
See:
http://clarkvision.com/imagedetail/does.pixel.size.matt...

Roger
Anonymous
February 22, 2005 7:27:07 AM

Archived from groups: rec.photo.digital (More info?)

On Thu, 10 Feb 2005 05:00:53 GMT, "Roger N. Clark (change username to
rnclark)" <username@qwest.net>, wrote in news:420AEA85.5010503@qwest.net:

> Signal-to-noise, the larger sensor wins:
>
> Digital Cameras: Does Pixel Size Matter?
> Factors in Choosing a Digital Camera:
> http://clarkvision.com/imagedetail/does.pixel.size.matt...

Rogers,

I would like you to check the correct sensitivity of the S60 at ISO 50. I
think it's true sensitivity at ISO 50, as well as that of many other
Canon P&S digicams, is closer to ISO 100, because it apparently uses the
same 5mgpxl 1/1.8" sensor on some Sony digicams that appear to have the
same noise performance and exposure index at ISO 100.

It won't change your general conclusion, but I think it would this
sentence "that the DSLRs have lower noise at ISO 400 than the point and
shoot S60 at ISO 100".

Regards,

--
T.N.T.

Lbh xabj jung gb qb vs lbh rire jnag gb rznvy zr.
Anonymous
February 25, 2005 9:55:13 AM

Archived from groups: rec.photo.digital (More info?)

T.N.T. wrote:
> On Thu, 10 Feb 2005 05:00:53 GMT, "Roger N. Clark (change username to
> rnclark)" <username@qwest.net>, wrote in news:420AEA85.5010503@qwest.net:
>
>
>>Signal-to-noise, the larger sensor wins:
>>
>>Digital Cameras: Does Pixel Size Matter?
>>Factors in Choosing a Digital Camera:
>>http://clarkvision.com/imagedetail/does.pixel.size.matt...
>
>
> Rogers,
>
> I would like you to check the correct sensitivity of the S60 at ISO 50. I
> think it's true sensitivity at ISO 50, as well as that of many other
> Canon P&S digicams, is closer to ISO 100, because it apparently uses the
> same 5mgpxl 1/1.8" sensor on some Sony digicams that appear to have the
> same noise performance and exposure index at ISO 100.
>
> It won't change your general conclusion, but I think it would this
> sentence "that the DSLRs have lower noise at ISO 400 than the point and
> shoot S60 at ISO 100".
>
> Regards,
>
It depends on how one defines ISO. If by full well, then
it is close to 50. At the time I did the S60 tests, I
imaged the same target with my Canon 1D Mark II, and
relative to the 1DII ISO 100, the S60 was 50. By this,
I mean the percentage of full well was similar for
these two cameras, with the S60 shutter speed 1/2 the 1DII
when the S60 was at iso 50, and the 1DII at 100.

Roger
!