10D ISO ratings a lie?

MrFoopy said:

Since you didn't report it, and more so because I've made the
mistake myself more times than I'd like to admit, you don't have
any exposure compensation set, do you? The first time I ever
checked my body against sunny f/16, I made that mistake and didn't
figure it out for a while. Sigh...

Click to expand...

I thought I was clear in my first post, but let me try it again; if I set the camera in manual mode (where the camera doesn't have any exposure compensation; it merely reports how much it thinks you are under- or over-exposing the scene) to ISO 100, 1/100, and f16, and shoot things directly illuminated by a high sun on a day with dinstict shadows, grey subjects should become midtones, but they become bright shadows instead. Clean white automobiles should be near the top, but they are over a stop from the top. White autos, and white houses and white trim are the only things that register in the upper half of the histogram at all.
In direct, blinding sunlight.

--
John

Ron Parr said:

Now, it is true that film and electronic sensors have different
methods for establishing sensitivity. Electronic sensors are
supposed to be measured in accordance with ISO 12232:1998. This is
standard is supposed to produce results close to film, but the
measurement methods are fundamentally different.

Click to expand...

The bottom line, though, is that my camera exposes well in auto mode. It knows what the real sensitivity of the sensor is, and is obviously calibrated for it. It says it is ISO 100, but it is really ISO 40, and that's exactly what the auto-exposure system is calibrated for.

I don't have time to read the sekonic stuff now, but it looks interesting. Cameras and external meters should be looking at R, G, and B separately, in my opinion. They are pretty useless for digital in luminance mode. A metering system for digital should be able to know the sensitivity of the color channels, and have the ability to give maximum exposure without clipping in RAW and JPEG modes. I am constantly amazed at how stone-age the whole world of camera tech, especially digital, is.
--
John

Al Pacheco said:

That rule of thumb is mainly for negative film. Digital requires
some adjustment to the rule.

Click to expand...

You're not getting it - the camera's auto-exposure and manual metering agree pretty closely to the sunny f16 rule. There is a discrepancy only when actually shooting in manual by not looking at the camera's metering, but by sunny f16, or an external meter.

IOW, the ISO that the camera is set to is not the real ISO, that is what I'm talking about. The camera itself knows exactly what the sensitivity of the sensor and amplifier are, and auto-expose properly.

My "ISO 100" is actually ISO 40. My ISO 800 is actually ISO 320, etc.

--
John

Holden said:

Sunny 16 is EV15... whether digital, analog, motion or still.

Click to expand...

Holden said:

It is the value to which most test benches are calibrated. The
sensor of a digital camera should get pretty close... at least to
medium gray. Trying to hold all the dynamic range of such a bright
scene may be beyond the specification of the given chip.. and yes
would require some adjustment..

Click to expand...

Any explanation about the difference with digital is totally irrelevant here. The camera, used as intended, with auto-exposure or manual mode's metering, exposes a well-balanced scene correctly. It is only when manual mode is used and the setting chosen from an external meter or sunny f16, that the discrepancy exists. It exist because the camera says "ISO 100", but it's metering for ISO 40. Even if, for the sake of argument, the camera were under-exposing to avoid clipping highlights, 1.3 stops of underexposure leaves a full 5.1 stops of headroom above middle grey, and 4.5 stops of green and blue (in white daylight) in RAW mode, and 3.5 stops in JPEG mode. But, it clearly does not do that. With its own metering, it frequently blows out highlights in JPEG mode.

The bottom line is that taking the ISO values my 10D claims to be using, literally, results in gross under-exposure. Using sunny f16, in raw mode, a bright white object in direct sun registers about 1000 out of 4095 RAW levels in the blue and green channels, and about 600 in the red channel. That's 2 to 2.5 bits per color channel, down the drain, unused.

Holden said:

this is what Fred Mirandas DRI module is all about...

Click to expand...

--
John

Al Pacheco said:

And it does not apply as a hard and fast rule for digital. I never
calibrated a bench but I make my living calibrating photographic
equipment and digital is different enough to warrant a new rule of
thumb.

Click to expand...

You're completely missing the point.

The camera does not have any problems in its metering, or auto-exposure; do you understand that part? If I shoot a grey card or a black wall, or a white wall in the direct sun, with the camera's metering, they come out to 128, and the middle of the histogram on the back of the camera. If I shoot the grey card with sunny f16, or measured by my external meter, it comes out a little bit more than 1 stop left of center.

The camera is designed to put even illumination, or the average illumination, at the middle of the histogram (128). If it under-exposes taking the ISO literally, then it can not possibly be an adaptation for "digital", because the camera doesn't use that difference for anything, except to state the ISO as something far from what it actually is. It meters and exposes just as it would for film, as far as the histogram is concerned.

--
John

John,

My personal experience has been that the rule of Sunny F16 is applicable to digital cameras - including my 10d. I get near perfect exposures when I shoot in bright sun with the 10d - and I rarely shoot in an idiot mode...manual only.

-JM

John Sheehy said:

Ron Parr said:

Now, it is true that film and electronic sensors have different
methods for establishing sensitivity. Electronic sensors are
supposed to be measured in accordance with ISO 12232:1998. This is
standard is supposed to produce results close to film, but the
measurement methods are fundamentally different.

Click to expand...

The bottom line, though, is that my camera exposes well in auto
mode. It knows what the real sensitivity of the sensor is, and is
obviously calibrated for it. It says it is ISO 100, but it is
really ISO 40, and that's exactly what the auto-exposure system is
calibrated for.

I don't have time to read the sekonic stuff now, but it looks
interesting. Cameras and external meters should be looking at R,
G, and B separately, in my opinion. They are pretty useless for
digital in luminance mode. A metering system for digital should be
able to know the sensitivity of the color channels, and have the
ability to give maximum exposure without clipping in RAW and JPEG
modes. I am constantly amazed at how stone-age the whole world of
camera tech, especially digital, is.
--
John

Click to expand...

This discussion motivated me to pull my old A-1 (Canon FD mount film SLR) from my attic and do some tests with my D60 (predecessor to 10D with same sensitivity in Phil's test shots).

Bot cameras were set to ISO 100, no compensation, and both had 50mm 1.8 lenses on. Obviously the A-1 would have a wider field of view, so I needed to pick uniformly illuminated areas for the test. (This is also necessary to avoid issues with the cameras using different averaging methods to determine exposure.)

I walked around the house pointing cameras at things like: Blank beige walls, darkly colored walls, our deck outstide, the sky, etc. I used aperture priority, keepting the aperture at f/5.6 (except for the sky) and checking the chosen shutter speed.

The D60 was always the same or at most 0.5 stops faster than the A-1. By faster, I mean that if the A-1 chose 1/125, the D60 would choose 1/90.

I suppose the next step would be to take shots with both and compare, but I'm sufficiently convinced that they meter essentially the same. Besides, I'm not sure if they sell film anymore in my neighborhood. ;-)

--
Ron Parr
FAQ: http://www.cs.duke.edu/~parr/photography/faq.html
Gallery: http://www.pbase.com/parr/

You say the camera doesn't have any problems metering nor with auto-exposure. But it doesn't follow the "sunny 16" rule. What does that tell you? The rule is not a rule for digital.

That's the point.

John Sheehy said:

Al Pacheco said:

And it does not apply as a hard and fast rule for digital. I never
calibrated a bench but I make my living calibrating photographic
equipment and digital is different enough to warrant a new rule of
thumb.

Click to expand...

You're completely missing the point.

The camera does not have any problems in its metering, or
auto-exposure; do you understand that part? If I shoot a grey card
or a black wall, or a white wall in the direct sun, with the
camera's metering, they come out to 128, and the middle of the
histogram on the back of the camera. If I shoot the grey card with
sunny f16, or measured by my external meter, it comes out a little
bit more than 1 stop left of center.

The camera is designed to put even illumination, or the average
illumination, at the middle of the histogram (128). If it
under-exposes taking the ISO literally, then it can not possibly be
an adaptation for "digital", because the camera doesn't use that
difference for anything, except to state the ISO as something far
from what it actually is. It meters and exposes just as it would
for film, as far as the histogram is concerned.

--
John

Click to expand...

John Sheehy said:

You're completely missing the point.

Click to expand...

Careful analysis of the ISO of a few 10Ds has revealed an ISO 100 that is right about in the 110-120 range. It does vary by camera and the calibration that occurs on those cameras however.

John Sheehy said:

The camera does not have any problems in its metering, or
auto-exposure; do you understand that part? If I shoot a grey card
or a black wall, or a white wall in the direct sun, with the
camera's metering, they come out to 128, and the middle of the
histogram on the back of the camera.

Click to expand...

The is NOT where the histogram should lie. The histogram should lie at 13% gray, which is NOT in the middle. This means your camera is over-exposing. This would partially explain your errors.

Jason

John Sheehy said:

The bottom line is that taking the ISO values my 10D claims to be
using, literally, results in gross under-exposure. Using sunny
f16, in raw mode, a bright white object in direct sun registers
about 1000 out of 4095 RAW levels in the blue and green channels,
and about 600 in the red channel. That's 2 to 2.5 bits per color
channel, down the drain, unused.

Click to expand...

As I mentioned above, the basis if your confusion is you don't understand where middle gray is. Beyond that, my D60, 300D, and 1DII all have clipped white areas on me when shooting at sunny 16. Usually it is 1/400th @ ISO 100 @ f/8. Maybe you should check the Tstop of your lens you use, and check the filtration % of the atmosphere in your area.

I guess your goal is to be the first guy that finds this big problem that everyone overlooked, when in fact you just don't understand the basics of this rule of thumb.

Jason

Al Pacheco said:

You say the camera doesn't have any problems metering nor with
auto-exposure. But it doesn't follow the "sunny 16" rule. What does
that tell you? The rule is not a rule for digital.

That's the point.

Click to expand...

Your point is poor circular reasoning, designed to defend the exaggerated ISO ratings of the camera.

If there were some range of tones where the difference between film and digital warranted higher ISO ratings for digital, you might have a point, but EVERY zone is under-exposed when I shoot sunny F16. The shadows are almost gone, the midtones are bright shadows, and the highlights are light grey. The dynamic range above bright whites is totally wasted.

--
John

I tend to agree with Ron Parr's results and in addition, I seem to recall all reviews pointing out that Canon's ISO ratings are actually on the conservative side. This is definitely true when compared to the Nikon counterparts since the Canon's require 0.3-0.6eV less light for the same exposure! This really says a lot about Canon's "low noise at high ISO" since the 10D/300D is still beating these other digital cameras in noise by a reasonable margin, despite the shorter exposure. Now the skeptic might suggest that this is due to differences in metering, however when the two shots are compared the Canon is actually slightly more exposed...go figure!

-Sean

Ron Parr said:

This discussion motivated me to pull my old A-1 (Canon FD mount
film SLR) from my attic and do some tests with my D60 (predecessor
to 10D with same sensitivity in Phil's test shots).

Bot cameras were set to ISO 100, no compensation, and both had 50mm
1.8 lenses on. Obviously the A-1 would have a wider field of view,
so I needed to pick uniformly illuminated areas for the test.
(This is also necessary to avoid issues with the cameras using
different averaging methods to determine exposure.)

I walked around the house pointing cameras at things like: Blank
beige walls, darkly colored walls, our deck outstide, the sky, etc.
I used aperture priority, keepting the aperture at f/5.6 (except
for the sky) and checking the chosen shutter speed.

The D60 was always the same or at most 0.5 stops faster than the
A-1. By faster, I mean that if the A-1 chose 1/125, the D60 would
choose 1/90.

Click to expand...

--
http://www.pbase.com/compuminus/cambridge
(recently changed my screen name from zappa69)

John,

I understand what you are talking about...and understand your concerns.

The sky here today is partly cloudy....sun is in and out so when (if) I get some nice bright sunlight today I will take my 10D outside and give it a shot with my 10D to see if it is consistent with the results your seeing.

Could be that it's just your camera...or you could be on to something.

John Sheehy said:

Has anyone tried using "Sunny f16" in manual mode with a 10D on a
sunny day?

When I do, the pictures are under-exposed by over a stop. When I
check exposure of a flat, grey surface or blue sky in the camera,
and compare it to my Sekonic meter, there is about 1.3 stops
difference.

This casts a shadow on the integrity of the camera's "low noise at
high ISO", when ISO 800 is really ISO 320, no?

--
John

Click to expand...

He's not talking about metering. He's talking about if the sensor is producing a proper expsosure at f/16, 1/100, 100 ISO on a sunny day.

Ron Parr said:

This discussion motivated me to pull my old A-1 (Canon FD mount
film SLR) from my attic and do some tests with my D60 (predecessor
to 10D with same sensitivity in Phil's test shots).

Bot cameras were set to ISO 100, no compensation, and both had 50mm
1.8 lenses on. Obviously the A-1 would have a wider field of view,
so I needed to pick uniformly illuminated areas for the test.
(This is also necessary to avoid issues with the cameras using
different averaging methods to determine exposure.)

I walked around the house pointing cameras at things like: Blank
beige walls, darkly colored walls, our deck outstide, the sky, etc.
I used aperture priority, keepting the aperture at f/5.6 (except
for the sky) and checking the chosen shutter speed.

The D60 was always the same or at most 0.5 stops faster than the
A-1. By faster, I mean that if the A-1 chose 1/125, the D60 would
choose 1/90.

I suppose the next step would be to take shots with both and
compare, but I'm sufficiently convinced that they meter essentially
the same. Besides, I'm not sure if they sell film anymore in my
neighborhood. ;-)

--
Ron Parr
FAQ: http://www.cs.duke.edu/~parr/photography/faq.html
Gallery: http://www.pbase.com/parr/

Click to expand...

JohnnyMnemonic said:

He's not talking about metering. He's talking about if the sensor
is producing a proper expsosure at f/16, 1/100, 100 ISO on a sunny
day.

Click to expand...

I know exactly what he's talking about. The sunny 16 thing is meaningless and not what's he's really talking about. He's talking about the camera having lower than rated ISO and having a meter that is adjusted to compensate for this, i.e., ISO 100 isn't really ISO 100.

Under the assumption that my both my A-1 and D60 produce correctly exposed images when I use their built in meters, (seems to be the case, to me recollection), my test proved that they both have essentially the same notion of what ISO 100 means.

--
Ron Parr
FAQ: http://www.cs.duke.edu/~parr/photography/faq.html
Gallery: http://www.pbase.com/parr

At sporting events, especially baseball, the sunny f/16 rule works better for me than automatic metering. In general for sporting events I use manual metering, whether it's the sunny f/16 rule, or taking some shot prior to the start of action and dialing the exposure in via the histogram.

It's been my experience that the camera's ISO sensitivity is very, very close.

These were initially set up with the sunny f/16 rule, but I dialed back the exposure to avoid blowing out the highlights on the white uniform's. As you can see, I had to dodge Zambtano's face a bit to compensate for the intentional underexposure:

ISO 200, f/5.6, 1/1000 sec.





= Ed =

John Sheehy said:

Has anyone tried using "Sunny f16" in manual mode with a 10D on a
sunny day?

When I do, the pictures are under-exposed by over a stop. When I
check exposure of a flat, grey surface or blue sky in the camera,
and compare it to my Sekonic meter, there is about 1.3 stops
difference.

This casts a shadow on the integrity of the camera's "low noise at
high ISO", when ISO 800 is really ISO 320, no?

--
John

Click to expand...

--
= Ed Rotberg =

'A waist is a terrible thing to mind'
http://www.edrotberg.org/gallery

Could you explain how the Sunny 16 is meaningless?

If the sensor ISO is calibrated properly...1/100 at ISO 100 at f/16 should produce a properly exposed photograph. So I don't see how it is meaningless.

One could argue that you're comparisons are meaningless to what the poster is talking about because he's referring to a 10D...and you are comparing metering of a D60 and an A1. A 10D does not equal a D60.

I just did some actual tests with a 10D...which I'll be posting in a few minutes. My tests disagree with the original poster's observations...which I can back up with actual photographs.

Ron Parr said:

JohnnyMnemonic said:

He's not talking about metering. He's talking about if the sensor
is producing a proper expsosure at f/16, 1/100, 100 ISO on a sunny
day.

Click to expand...

I know exactly what he's talking about. The sunny 16 thing is
meaningless and not what's he's really talking about. He's talking
about the camera having lower than rated ISO and having a meter
that is adjusted to compensate for this, i.e., ISO 100 isn't really
ISO 100.

Under the assumption that my both my A-1 and D60 produce correctly
exposed images when I use their built in meters, (seems to be the
case, to me recollection), my test proved that they both have
essentially the same notion of what ISO 100 means.

--
Ron Parr
FAQ: http://www.cs.duke.edu/~parr/photography/faq.html
Gallery: http://www.pbase.com/parr

Click to expand...

JohnnyMnemonic said:

Could you explain how the Sunny 16 is meaningless?

Click to expand...

Sunny 16 is a rule of thumb - that's all. Using it to judge ISO calibration is absurd.

JohnnyMnemonic said:

One could argue that you're comparisons are meaningless to what the
poster is talking about because he's referring to a 10D...and you
are comparing metering of a D60 and an A1. A 10D does not equal a
D60.

Click to expand...

As I ALREADY said, Phil's review shows that they meter the same so the test is quite meaningful.

--
Ron Parr
FAQ: http://www.cs.duke.edu/~parr/photography/faq.html
Gallery: http://www.pbase.com/parr/

John Sheehy said:

Any explanation about the difference with digital is totally
irrelevant here.

Click to expand...

You are right about this we are actually discussing a metering system.. which is quite similar to film and hand held meters, digital cameras and video cameras.

John Sheehy said:

The camera, used as intended, with auto-exposure
or manual mode's metering, exposes a well-balanced scene correctly.

Click to expand...

Then are you trying to do?

John Sheehy said:

It is only when manual mode is used and the setting chosen from an
external meter or sunny f16, that the discrepancy exists.

Click to expand...

My 10D is spot on incident, slightly overexposes (1/3 stop) in reflective, when metering fom my Gossen Luna Pro F, using sunny 16... and full manual on camera

It exist

John Sheehy said:

because the camera says "ISO 100", but it's metering for ISO 40.
Even if, for the sake of argument, the camera were under-exposing
to avoid clipping highlights, 1.3 stops of underexposure leaves a
full 5.1 stops of headroom above middle grey, and 4.5 stops of
green and blue (in white daylight) in RAW mode, and 3.5 stops in
JPEG mode. But, it clearly does not do that. With its own
metering, it frequently blows out highlights in JPEG mode.

Click to expand...

You are now discussing the "range" of exposure that your chip has... remember all meters are calibrated to 18% medium gray... let me say that another way "all meters are calibrated to produce an accurate 18% medium gray". This is taught in the zone system, beyond the zone system and virtually every competant photo guide out there.

John Sheehy said:

The bottom line is that taking the ISO values my 10D claims to be
using, literally, results in gross under-exposure.

Click to expand...

If so, you need to have something calibrated, either your camera or your handheld meter. This is not an inherent defect in this camera. Again mine is spot on..... to med gray mind you, it is irrelevant to check anything else. At least for purposes of testing your meter as you are then discussing the dynamic range of your chip.

Using sunny

John Sheehy said:

f16, in raw mode, a bright white object in direct sun registers
about 1000 out of 4095 RAW levels in the blue and green channels,
and about 600 in the red channel. That's 2 to 2.5 bits per color
channel, down the drain, unused.

Click to expand...

You are once again discussing the dynamic range of your CMOS chip, not the metering system

John, I am not sure what your real agenda is, but a number of people here have said this same thing to you in a myriad of ways... you are confusing your issues I am afraid... best of luck...

G.

I just shot a gray card under partly cloudy skies at f/16, ISO 100, 1/100.

The results seem to show that my camera is pretty much on the mark with the Sunny-16 rule. Even though it shows about a half stop under...I'd consider this minimal and would blame it on the 'partly cloudy' skies....the sun is not fully out today but I did grab a chance where the sun was fairly bright. So I'd account for the half stop difference with the fact that it's partly cloudy today.

I wouldn't consider these results a problem and if you're seeing something much different I'd consider taking it up with Canon...it doesn't seem that it is a systematic defect with 10D's in general.

Here's my results:



Here's the EC slider adjusted to bring it right to the middle....+.55

John Sheehy said:

Has anyone tried using "Sunny f16" in manual mode with a 10D on a
sunny day?

When I do, the pictures are under-exposed by over a stop. When I
check exposure of a flat, grey surface or blue sky in the camera,
and compare it to my Sekonic meter, there is about 1.3 stops
difference.

This casts a shadow on the integrity of the camera's "low noise at
high ISO", when ISO 800 is really ISO 320, no?

--
John

Click to expand...