Canon 20D - ISO 100 performance?

In article ,
says...
The difference in noise is smaller, and Canon users should be thankful
for having a ISO 100 in the first place. Most of the cameras using
Sony's 6 megapixel CCD start at ISO 200.

I wonder why the Sony-CCD cameras don't at least give you the option of
going to 100 ISO if you really want to. Would the results be too
horrible to bear, or are they worried about comparisons against other
cameras that do provide this capability, or what?

Well, what I don't get is why Nikon will start at 200, yet the Minolta
7D (using the same Sony sensor, IIRC) can start at 100, with excellent
noise characteristics.
--
http://www.pbase.com/bcbaird

In message ZasNe.46991$084.11335@attbi_s22,
"Colic" wrote:

The 20D is behind the D60 (in how clean the image is) at ISO 100, but
by 200 they are essentially even, and at 400 the 20D has a slight lead.
At 800 the 20D simply shines, often showing images that are very close
to the D60 at 200. I use ISO 1600 on the 20D far more often than I
dared use 400 on the D60.

If this is true, it might be because of the 20D's faster readout, and
the root cause may be the same issue that causes banding; poor
digitization to make faster burst speeds possible. This would affect
the lowest ISO the most, as the limit for shadows tends to more related
to digitization than noise.

Of course, it is also possible that the shadow blackpoint is treated
differently, and noise is clipped away, but also detail (blackened), in
the D60 shadows.

I know the 10D and 20D are totally different in the way the converters
tend to treat the deepest shadows.

Anyway, comments like yours and the person you responded to leave a lot
of questions, as you don't specify whether you are talking about JPEGs,
RAW renders in a specific program, or the RAW data itself.
--


John P Sheehy

-hh wrote:


Now to verify that I understand this correctly from a "big picture"
perspective, what this suggests is that the traditional paradigm of
film that "lower ISO always results in finer grain" is functionally not
equivalently true for digital...correct?

One key advantage of digital is that the grain is constant size
regardless of ISO. The dynamic is noisier with higher ISO, esp. above
800 or so.

The "lowest" iso of a camera is slightly noisier in some cameras than at
one stop faster ISO. In the end (make a print) it's hardly perceptible.

Cheers,
Alan
--
-- r.p.e.35mm user resource: http://www.aliasimages.com/rpe35mmur.htm
-- r.p.d.slr-systems: http://www.aliasimages.com/rpdslrsysur.htm
-- [SI] gallery & rulz: http://www.pbase.com/shootin
-- e-meil: Remove FreeLunch.

SMS wrote:

Brian Baird wrote:

The difference in noise is smaller, and Canon users should be thankful
for having a ISO 100 in the first place. Most of the cameras using
Sony's 6 megapixel CCD start at ISO 200.


I wonder why the Sony-CCD cameras don't at least give you the option of
going to 100 ISO if you really want to. Would the results be too
horrible to bear, or are they worried about comparisons against other
cameras that do provide this capability, or what?

Maxxum 7D uses the same sensor line and goes 100 to 3200. 3200 is, not
surprisingly, quite noisy.

I'm not convinced that there is anything at all wrong with the sensor at
100 that Nikon and Pentax could not have included it. OTOH, many
kit/consumer lenses go with the likes of D70's and *ist's, so that extra
stop of speed at the slow end is a benefit of sorts.

per
http://www.dpreview.com/reviews/koni...lumi-graph.gif

you can see that the 7D is slightly noisier at 100 than 200.

For faster glass, it is germane to note, that having the ISO 100 does
allow shallower depth of field for a brighly lit situation.

Cheers,
Alan.

--
-- r.p.e.35mm user resource: http://www.aliasimages.com/rpe35mmur.htm
-- r.p.d.slr-systems: http://www.aliasimages.com/rpdslrsysur.htm
-- [SI] gallery & rulz: http://www.pbase.com/shootin
-- e-meil: Remove FreeLunch.

In message ,
Alan Browne wrote:

Maxxum 7D uses the same sensor line and goes 100 to 3200.

These numbers are fairly arbitrary. I would not be concerned about a
camera only going down to "ISO 200", if you can set the camera to "ISO
200" and shoot just like it was ISO 100, without blowing reasonable
highlights, or if you could do it at ISO 125. Different camera metering
system have different ways of utilizing the 12 bits of RAW data.

I have 2 DSLRs, the Canon 10D and 20D, and they are from the same
series, and do things quite differently. If I use the same f-stop and
shutter speed, on a tripod-mounted lens, and switch cameras set to the
same ISO, the 10D registers darker in the RAW data. If I let each
camera meter on its own, the 10D gives a higher absolute exposure by 2/3
stop, nearly closing the gap in the RAW data, but the default JPEG
render of the in-camera converter, combined with the metering, will blow
out highlights much quicker on the 10D. The native ISO of the sensor
(as windowed by the digitizer) is clearly lower for the 10D; the 10D,
however, meters so as to push it about 2/3 stop. IOW, if we can say
that native ISO of the 20D is 100, then the native ISO of the 10D is 64,
relatively speaking; if we consider the native ISO of the 10D to be 100,
then the native ISo of the 20D is 160. IOW, the 20D has moved to a
higher native ISO than the 10D.

3200 is, not
surprisingly, quite noisy.

Of course, this could be due to the defaults leaving a lot of RAW
headroom. Unless you take a shot with manual exposure with a known
light intensity, you don't know what is really going on.

I'm not convinced that there is anything at all wrong with the sensor at
100 that Nikon and Pentax could not have included it.

Maybe not, but you could lose up to a stop of the highlight headroom
present at ISO 200.



--


John P Sheehy

wrote:

In message ,
Alan Browne wrote:


Maxxum 7D uses the same sensor line and goes 100 to 3200.


These numbers are fairly arbitrary. I would not be concerned about a
camera only going down to "ISO 200", if you can set the camera to "ISO
200" and shoot just like it was ISO 100, without blowing reasonable
highlights, or if you could do it at ISO 125. Different camera metering
system have different ways of utilizing the 12 bits of RAW data.

In D/SLR's the metering system does not use the sensor data.


I have 2 DSLRs, the Canon 10D and 20D, and they are from the same
series, and do things quite differently. If I use the same f-stop and
shutter speed, on a tripod-mounted lens, and switch cameras set to the
same ISO, the 10D registers darker in the RAW data. If I let each
camera meter on its own, the 10D gives a higher absolute exposure by 2/3
stop, nearly closing the gap in the RAW data, but the default JPEG
render of the in-camera converter, combined with the metering, will blow
out highlights much quicker on the 10D. The native ISO of the sensor
(as windowed by the digitizer) is clearly lower for the 10D; the 10D,
however, meters so as to push it about 2/3 stop. IOW, if we can say
that native ISO of the 20D is 100, then the native ISO of the 10D is 64,
relatively speaking; if we consider the native ISO of the 10D to be 100,
then the native ISo of the 20D is 160. IOW, the 20D has moved to a
higher native ISO than the 10D.


3200 is, not
surprisingly, quite noisy.


Of course, this could be due to the defaults leaving a lot of RAW
headroom. Unless you take a shot with manual exposure with a known
light intensity, you don't know what is really going on.


I'm not convinced that there is anything at all wrong with the sensor at
100 that Nikon and Pentax could not have included it.


Maybe not, but you could lose up to a stop of the highlight headroom
present at ISO 200.





--
-- r.p.e.35mm user resource: http://www.aliasimages.com/rpe35mmur.htm
-- r.p.d.slr-systems: http://www.aliasimages.com/rpdslrsysur.htm
-- [SI] gallery & rulz: http://www.pbase.com/shootin
-- e-meil: Remove FreeLunch.

In message ,
Alan Browne wrote:

wrote:

In message ,
Alan Browne wrote:

Maxxum 7D uses the same sensor line and goes 100 to 3200.

These numbers are fairly arbitrary. I would not be concerned about a
camera only going down to "ISO 200", if you can set the camera to "ISO
200" and shoot just like it was ISO 100, without blowing reasonable
highlights, or if you could do it at ISO 125. Different camera metering
system have different ways of utilizing the 12 bits of RAW data.

In D/SLR's the metering system does not use the sensor data.

Your definition of "utilizing" is a bit narrow, don't you think?
"Utilizing" does not mean "deriving directly from".

The design of the camera acknowledges the absolute sensitivity of the
camera, and how sensor exposure maps to the RAW data at a given ISO.
--


John P Sheehy

wrote:

In message ,
Alan Browne wrote:


Maxxum 7D uses the same sensor line and goes 100 to 3200.


These numbers are fairly arbitrary. I would not be concerned about a
camera only going down to "ISO 200", if you can set the camera to "ISO
200" and shoot just like it was ISO 100, without blowing reasonable
highlights, or if you could do it at ISO 125. Different camera metering
system have different ways of utilizing the 12 bits of RAW data.

What do you mean? The metering systems in D/SLR's is not based on the
imaging sensor ... it's based on the sensor in the optical/prism path
before exposure.


I have 2 DSLRs, the Canon 10D and 20D, and they are from the same
series, and do things quite differently. If I use the same f-stop and
shutter speed, on a tripod-mounted lens, and switch cameras set to the
same ISO, the 10D registers darker in the RAW data. If I let each
camera meter on its own, the 10D gives a higher absolute exposure by 2/3
stop, nearly closing the gap in the RAW data, but the default JPEG
render of the in-camera converter, combined with the metering, will blow
out highlights much quicker on the 10D. The native ISO of the sensor
(as windowed by the digitizer) is clearly lower for the 10D; the 10D,
however, meters so as to push it about 2/3 stop. IOW, if we can say
that native ISO of the 20D is 100, then the native ISO of the 10D is 64,
relatively speaking; if we consider the native ISO of the 10D to be 100,
then the native ISo of the 20D is 160. IOW, the 20D has moved to a
higher native ISO than the 10D.

No standard? I think what really hurts 10D/20D shooters is the lack of
a spot meter in the camera. That would allow critical setting for
highlight exposure.



3200 is, not
surprisingly, quite noisy.


Of course, this could be due to the defaults leaving a lot of RAW
headroom. Unless you take a shot with manual exposure with a known
light intensity, you don't know what is really going on.

Even when I do, there is so much pre and post A/D amplification going on
that the image can't help but be noisy.

I'm not convinced that there is anything at all wrong with the sensor at
100 that Nikon and Pentax could not have included it.


Maybe not, but you could lose up to a stop of the highlight headroom
present at ISO 200.

Not if I meter to place the highlight at the top for that sensitivity.
And if I need a faster speed or smaller aperture, then it's no big deal
to go to 200. On the other hand, if I need less speed or larger
aperture I at least have the 100. Slower than 100 would be a benefit in
some instances as well.

Cheers,
Alan


--
-- r.p.e.35mm user resource: http://www.aliasimages.com/rpe35mmur.htm
-- r.p.d.slr-systems: http://www.aliasimages.com/rpdslrsysur.htm
-- [SI] gallery & rulz: http://www.pbase.com/shootin
-- e-meil: Remove FreeLunch.

In message ,
Alan Browne wrote:

wrote:

In message ,
Alan Browne wrote:


Maxxum 7D uses the same sensor line and goes 100 to 3200.


These numbers are fairly arbitrary. I would not be concerned about a
camera only going down to "ISO 200", if you can set the camera to "ISO
200" and shoot just like it was ISO 100, without blowing reasonable
highlights, or if you could do it at ISO 125. Different camera metering
system have different ways of utilizing the 12 bits of RAW data.

What do you mean? The metering systems in D/SLR's is not based on the
imaging sensor ... it's based on the sensor in the optical/prism path
before exposure.

Is the response of the sensor and the digitization of the voltages
unknown, and not utilized?

I have 2 DSLRs, the Canon 10D and 20D, and they are from the same
series, and do things quite differently. If I use the same f-stop and
shutter speed, on a tripod-mounted lens, and switch cameras set to the
same ISO, the 10D registers darker in the RAW data. If I let each
camera meter on its own, the 10D gives a higher absolute exposure by 2/3
stop, nearly closing the gap in the RAW data, but the default JPEG
render of the in-camera converter, combined with the metering, will blow
out highlights much quicker on the 10D. The native ISO of the sensor
(as windowed by the digitizer) is clearly lower for the 10D; the 10D,
however, meters so as to push it about 2/3 stop. IOW, if we can say
that native ISO of the 20D is 100, then the native ISO of the 10D is 64,
relatively speaking; if we consider the native ISO of the 10D to be 100,
then the native ISo of the 20D is 160. IOW, the 20D has moved to a
higher native ISO than the 10D.

No standard? I think what really hurts 10D/20D shooters is the lack of
a spot meter in the camera. That would allow critical setting for
highlight exposure.

I wouldn't mind having one, but I'd probably use it very rarely. I
don't really have a lot of experiences where I wish I had a spotmeter,
and unless the camera had a +3 EC setting, it is pretty useless for
automation. Point the spotmeter at the brightest highlight and +2 EC,
and you've wasted about half of the RAW values in the green channel;
more in the others. +3 EC on the dial is essential for maximal use of a
spotmeter.

Spot for the subject, and you may blow out the background, so what good
is it? Film paradigms do not work optimally with digital.

3200 is, not
surprisingly, quite noisy.

Of course, this could be due to the defaults leaving a lot of RAW
headroom. Unless you take a shot with manual exposure with a known
light intensity, you don't know what is really going on.

Even when I do, there is so much pre and post A/D amplification going on
that the image can't help but be noisy.

What about the other end; ISO 200. If you expose a grey-card based on
an external meter, where do the RAW values fall?

I'm not convinced that there is anything at all wrong with the sensor at
100 that Nikon and Pentax could not have included it.

Maybe not, but you could lose up to a stop of the highlight headroom
present at ISO 200.

Not if I meter to place the highlight at the top for that sensitivity.

I really have a hard time understanding you. Which sensitivity?

If you expose so that the highlights are just short of clipping the RAW
data when the camera is set to ISO 200, you may actually have an
exposure index of ISO 100, or even less, of the subject is low-contrast.
The limit of pull-ability is the contrast of the scene. If you can't
pull a typical-contrast scene one stop with the camera set to ISO 200,
then you can't use an exposure index of ISO 100 without clipping. I
would say that you need about 2.5 stops of headroom above middle grey
before you can adequately claim an exposure index.

And if I need a faster speed or smaller aperture, then it's no big deal
to go to 200. On the other hand, if I need less speed or larger
aperture I at least have the 100. Slower than 100 would be a benefit in
some instances as well.

If the sensor doesn't collect enough measurable photons, though, you can
only achieve the lower exposure index through neutral density filters
(actually, a magenta filter would be better on many DSLRs; I do most of
my strong-light, low-ISO shooting with a magenta filter over the lens,
for better blue and red shadows).
--


John P Sheehy

In article , Alan Browne
wrote:

No standard? I think what really hurts 10D/20D shooters is the lack of
a spot meter in the camera. That would allow critical setting for
highlight exposure.

I don't need/want a spot meter in my camera. It's probably the primary
reason I never bought Canon stuff way back when. An averaging meter
combined with many years of experience works just fine.