Consistent underexposure on K10D

switters · 08-01-2007, 06:34 PM

Okay, I received my K10D yesterday and the first lens (of several) I'm trying today - the 35/2. I'm freshly over from the Canon team (20D), so this is my first time taking the Pentax kit for a spin.

I'm amazed at how differently the two cameras meter and expose. I did some side by side tests and the K10D was underexposing consistently compared to the 20D. Often times the right side of the histogram was somewhere in the middle of the tonal range. These weren't particularly difficult shooting conditions, just some portraits indoors against a medium-toned wall. The Canon was getting it right every time.

I started out in evaluative metering, but then tried center-weighted and spot after the first few underexposures. I figured the K10D might be picking up on the lighter tones in the wall and biasing the exposure downward from that. When I spot and center metered directly on the model, however, there wasn't that much of a change in exposure.

I didn't read about this in any reviews and haven't seen it mentioned on forums. Has anyone here experience this? What metering mode do you find works best in general, and are there any "tricks" for getting proper exposure with the K10D that you've figured out.

Please note that although there is always room for user error

, I was shooting the exact same picture with the Canon 20D and it was getting the exposure I expected.

Thanks for your help.

SpecialK · 08-01-2007, 06:48 PM

Samples with exif required for best replies.

MJB DIGITAL · 08-01-2007, 07:01 PM

switters
the same thing happens with both my ist*D and my K10D. I think that Pentax metering does actually consistantly underexpose. I also think you are right about the brightest highlights being favored and the rest being metered downward from that.
This is something that I wish would change, but I work with it and i like my pentax and trust it with my professional work

....i actually dont call myself a professional yet...i call my self a 'working photographer'

jeffkrol · 08-01-2007, 07:09 PM

Originally posted by switters

Okay, I received my K10D yesterday and the first lens (of several) I'm trying today - the 35/2. I'm freshly over from the Canon team (20D), so this is my first time taking the Pentax kit for a spin.

I'm amazed at how differently the two cameras meter and expose. I did some side by side tests and the K10D was underexposing consistently compared to the 20D. Often times the right side of the histogram was somewhere in the middle of the tonal range. These weren't particularly difficult shooting conditions, just some portraits indoors against a medium-toned wall. The Canon was getting it right every time.

I started out in evaluative metering, but then tried center-weighted and spot after the first few underexposures. I figured the K10D might be picking up on the lighter tones in the wall and biasing the exposure downward from that. When I spot and center metered directly on the model, however, there wasn't that much of a change in exposure.

I didn't read about this in any reviews and haven't seen it mentioned on forums. Has anyone here experience this? What metering mode do you find works best in general, and are there any "tricks" for getting proper exposure with the K10D that you've figured out.

Please note that although there is always room for user error

, I was shooting the exact same picture with the Canon 20D and it was getting the exposure I expected.

Thanks for your help.

Yep your Canon is overexposing......Welcome to the world of correctly calibrated to ANSI standards metering and iso ratings.......

ALL kidding aside, the best place to start is in the fact that the meters are diffent, not necessarily one right or wrong....
Start by reading this paper regarding Canon and by thinking that the Pentax meter is just plain dumb in the old school sense. It will not "second guess" what your shooting except possibly in Matrix metering mode. which unfortunately is a bit primitive compaired to say Nikon w/ it's 10000 entry look up table....
Canon exposure meterhttp://doug.kerr.home.att.net/pumpkin/Exposure_Calibration.pdf
To quote: (Mr. Kerr has more papers that are quite interesting and I recommend reading them)
So that leaves Canon with only the possibility of achieving their target value of Hu/Hsat by using a non-standard rating of ISO sensitivity—a
rating that is about 0.74 that which would be determined under ISO
12232. In other words, the sensitivity that is designated “ISO 100” by
Canon would probably be rated at about ISO 135 under ISO 12232.
Sooo this means, in general the Canon overexposes in regard to a standard. Now this is not a bad thing, most seem to find it helpful BUT it is not the only way to play the game.

MJB DIGITAL · 08-01-2007, 07:17 PM

thanks jeff!!

pmperry · 08-01-2007, 08:23 PM

I just made a similar jump and I wasn't so surprised as Nikon's system was a similar experience.

One thing I did notice after selling my 20D was that the manuals for the XT, XTi, 20D, 30D and 5D all used the term weighted for every metering mode with the exception of Evaluative.

I found that term interesting because it implies that the metering isn't exclusive (such as spot metering) when it should be. Instead it implies that the camera is using the whole evaluative system and biasing towards your 3.5% spot area (for the 30D and 5D) which is kind of a cheap implementation.

Also, not that your older MF Lenses are limited to Center Weighted Metering (I believe it is the MF lenses but I'll have to double check) as noted in the cameras manual. One thing you can do though is use the Digital Preview mode to get a good look at the histogram first as the stop down metering seems less accurate than those that actually speak to the camera.

Anyway, I hope this helps.

pmperry · 08-01-2007, 08:35 PM

Originally posted by jeffkrol

Yep your Canon is overexposing......Welcome to the world of correctly calibrated to ANSI standards metering and iso ratings.......

ALL kidding aside, the best place to start is in the fact that the meters are diffent, not necessarily one right or wrong....
Start by reading this paper regarding Canon and by thinking that the Pentax meter is just plain dumb in the old school sense. It will not "second guess" what your shooting except possibly in Matrix metering mode. which unfortunately is a bit primitive compaired to say Nikon w/ it's 10000 entry look up table....
Canon exposure meterhttp://doug.kerr.home.att.net/pumpkin/Exposure_Calibration.pdf
To quote: (Mr. Kerr has more papers that are quite interesting and I recommend reading them)
So that leaves Canon with only the possibility of achieving their target value of Hu/Hsat by using a non-standard rating of ISO sensitivity—a
rating that is about 0.74 that which would be determined under ISO
12232. In other words, the sensitivity that is designated “ISO 100” by
Canon would probably be rated at about ISO 135 under ISO 12232.
Sooo this means, in general the Canon overexposes in regard to a standard. Now this is not a bad thing, most seem to find it helpful BUT it is not the only way to play the game.

While this certainly helps to explain some of it, it doesn't explain why a 20D using Partial Metering wouldn't require any EV boost when pointed at caucasian skin.

Ole · 08-01-2007, 08:45 PM

It could explain that. If it is correct that ISO 100 set on the Canon really is ISO 135 then that discrepancy corresponds to a "built in" positive EV compensation that you were not aware of.

If you dial in +0.5 or +1 EV on the Pentax you should get similar results as the Canon

pmperry · 08-02-2007, 05:42 AM

Originally posted by Ole

It could explain that. If it is correct that ISO 100 set on the Canon really is ISO 135 then that discrepancy corresponds to a "built in" positive EV compensation that you were not aware of.

If you dial in +0.5 or +1 EV on the Pentax you should get similar results as the Canon

True, it could, and if you read DPReview they note that the Canon's are actually ISO 125 for 100 and they're pretty much like this through the whole range.

Anyway, even a 3rd of a stop wouldn't make up for a whole stop so the images should look 2/3 underexposed. I think it's a mixture of both issues where Canon has their ISO tuned wrong and their dedicated metering modes aren't dedicated at all but rather just a stronger bias.

jeffkrol · 08-02-2007, 06:04 AM

Originally posted by pmperry

True, it could, and if you read DPReview they note that the Canon's are actually ISO 125 for 100 and they're pretty much like this through the whole range.

Anyway, even a 3rd of a stop wouldn't make up for a whole stop so the images should look 2/3 underexposed. I think it's a mixture of both issues where Canon has their ISO tuned wrong and their dedicated metering modes aren't dedicated at all but rather just a stronger bias.

This user test gives some validity to the 'correct" metering w/ the G10 vs a hand held meter. There is a difference but the ANSI standard has about a 1/2 stop "fudge factor" built in so.....
My 2 Pence...: Pentax SLR Talk Forum: Digital Photography Review
And if I remember correctly the camera "rounds" f stops to the nearest whole(?) stop but it is infinitly stepped in practice.
As to the Caucasion skin thing, I actually assume the Canon meter is more sofisticated than the Pentax (old fashioned so to speak) as to things like "fuzzy logic" ect. and could have a built in adjustment circuit hooked to all metering modes but that's just a guess.
Again Mr. Kerr has an interesting paper:
http://doug.kerr.home.att.net/pumpkin/300D_metering_DAK.pdf

switters · 08-02-2007, 07:07 AM

Hmmm. I can't say I'm totally thrilled about this. I know every camera/system is different and has its quirks. But accurate exposure metering is just something I didn't have to think much about with the 20D, and I liked that.

Making sure the +1/2 or +1 exposure comp is always set, and remembering to overexpose in manual by that amount (which is the mode I shoot most often in) is kind of a hassle. But I guess if I love everything else about the K10D it's one I can learn to live with.

I could configure the User settings in Aperture Priority, for example, with +1/2 or +1EV compensation. I'll check the manual, but I'm assuming exposure compensation resets when the camera is turned off in other exposure modes?

Any other tips would be appreciated. I'll post a few pics later as well.

switters · 08-02-2007, 09:02 AM

Okay, I've noticed something really interesting. I took several test shots with the K10D and imported them into Aperture. I had the camera set at normal exposure (no compensation), and as I reported earlier in the thread the histogram would often indicate underexposure.

However, when I imported these shots and opened them up in Aperture, they appear to be properly exposed! The tonal values go right up to the right side of the histogram, with a nice even distribution across the range. (I'm actually very impressed with the natural rendering of tones in comparison with the 20D).

Why is this discrepancy occurring (where the histogram in camera is significantly different than the histogram in my RAW processor - Aperture)? Does it have something to do with the K10D's in-camera display? Has anyone else noticed such a discrepancy?

I guess I'll have to learn how to "translate" my camera's histogram results into what will actually appear on my computer.

arbutusq · 08-02-2007, 09:02 AM

Be patient and give it time. When I switched from Nikon I was bugged by the meter, but now that I know it it works well, and far better than Nikon at preserving highlights. It is easily fooled by bright objects but each meter has its quirks, once you know them you're fine.

Another thing to be aware of is the tone curve applied to the RAW data in camera. Applying a steeper tone curve can change the apparant exposure by quite a bit.

switters · 08-02-2007, 09:13 AM

Ahhh. So the tone curve you're referring to could affect the apparent exposure, and thus histogram, in the camera? And that would explain the difference between the in-camera histogram and what appears in my RAW converter?

You're absolutely right, I'm sure - I just need to give it some time. And don't get me wrong - I love the Pentax K10D! Today I receive a gaggle of lenses to try - 14/2.8, DA 21, DA 40, 43/1.9, 50/1.4, DA 70. I'm gonna be a busy guy!

pmperry · 08-02-2007, 11:58 AM

Originally posted by jeffkrol

This user test gives some validity to the 'correct" metering w/ the G10 vs a hand held meter. There is a difference but the ANSI standard has about a 1/2 stop "fudge factor" built in so.....
My 2 Pence...: Pentax SLR Talk Forum: Digital Photography Review
And if I remember correctly the camera "rounds" f stops to the nearest whole(?) stop but it is infinitly stepped in practice.
As to the Caucasion skin thing, I actually assume the Canon meter is more sofisticated than the Pentax (old fashioned so to speak) as to things like "fuzzy logic" ect. and could have a built in adjustment circuit hooked to all metering modes but that's just a guess.
Again Mr. Kerr has an interesting paper:
http://doug.kerr.home.att.net/pumpkin/300D_metering_DAK.pdf

Hey that was my DPReview Post!

08-01-2007, 06:34 PM	#1
switters Veteran Member Join Date: Jul 2007 Posts: 346	Consistent underexposure on K10D Okay, I received my K10D yesterday and the first lens (of several) I'm trying today - the 35/2. I'm freshly over from the Canon team (20D), so this is my first time taking the Pentax kit for a spin. I'm amazed at how differently the two cameras meter and expose. I did some side by side tests and the K10D was underexposing consistently compared to the 20D. Often times the right side of the histogram was somewhere in the middle of the tonal range. These weren't particularly difficult shooting conditions, just some portraits indoors against a medium-toned wall. The Canon was getting it right every time. I started out in evaluative metering, but then tried center-weighted and spot after the first few underexposures. I figured the K10D might be picking up on the lighter tones in the wall and biasing the exposure downward from that. When I spot and center metered directly on the model, however, there wasn't that much of a change in exposure. I didn't read about this in any reviews and haven't seen it mentioned on forums. Has anyone here experience this? What metering mode do you find works best in general, and are there any "tricks" for getting proper exposure with the K10D that you've figured out. Please note that although there is always room for user error , I was shooting the exact same picture with the Canon 20D and it was getting the exposure I expected. Thanks for your help.

08-02-2007, 07:07 AM	#11
switters Veteran Member Join Date: Jul 2007 Posts: 346 Original Poster	Hmmm. I can't say I'm totally thrilled about this. I know every camera/system is different and has its quirks. But accurate exposure metering is just something I didn't have to think much about with the 20D, and I liked that. Making sure the +1/2 or +1 exposure comp is always set, and remembering to overexpose in manual by that amount (which is the mode I shoot most often in) is kind of a hassle. But I guess if I love everything else about the K10D it's one I can learn to live with. I could configure the User settings in Aperture Priority, for example, with +1/2 or +1EV compensation. I'll check the manual, but I'm assuming exposure compensation resets when the camera is turned off in other exposure modes? Any other tips would be appreciated. I'll post a few pics later as well. Last edited by switters; 08-02-2007 at 08:02 AM.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
MZ-7 and underexposure	photolady95	Film SLRs and Compact Film Cameras	89	07-09-2009 09:44 AM
Weird K10D underexposure problem	Kirivon	Pentax DSLR Discussion	9	12-19-2008 04:54 AM
underexposure	karq	Photographic Technique	11	12-10-2008 07:51 AM
K10D underexposure + Pentax Colorado	jshurak	Pentax DSLR Discussion	17	04-12-2008 10:54 AM
K10D - drastic underexposure!!	SupremeMoFo	Pentax DSLR Discussion	26	08-14-2007 01:27 AM

08-01-2007, 06:48 PM	#2
SpecialK Pentaxian Join Date: Dec 2006 Location: So California Photos: Gallery Posts: 16,480	Samples with exif required for best replies.

08-01-2007, 07:01 PM	#3
MJB DIGITAL Veteran Member Join Date: Jul 2007 Location: st. louis Posts: 1,170	switters the same thing happens with both my ist*D and my K10D. I think that Pentax metering does actually consistantly underexpose. I also think you are right about the brightest highlights being favored and the rest being metered downward from that. This is something that I wish would change, but I work with it and i like my pentax and trust it with my professional work ....i actually dont call myself a professional yet...i call my self a 'working photographer'

08-01-2007, 07:09 PM	#4
jeffkrol Veteran Member Join Date: Sep 2006 Location: Wisconsin USA Photos: Gallery Posts: 8,434	Originally posted by switters Okay, I received my K10D yesterday and the first lens (of several) I'm trying today - the 35/2. I'm freshly over from the Canon team (20D), so this is my first time taking the Pentax kit for a spin. I'm amazed at how differently the two cameras meter and expose. I did some side by side tests and the K10D was underexposing consistently compared to the 20D. Often times the right side of the histogram was somewhere in the middle of the tonal range. These weren't particularly difficult shooting conditions, just some portraits indoors against a medium-toned wall. The Canon was getting it right every time. I started out in evaluative metering, but then tried center-weighted and spot after the first few underexposures. I figured the K10D might be picking up on the lighter tones in the wall and biasing the exposure downward from that. When I spot and center metered directly on the model, however, there wasn't that much of a change in exposure. I didn't read about this in any reviews and haven't seen it mentioned on forums. Has anyone here experience this? What metering mode do you find works best in general, and are there any "tricks" for getting proper exposure with the K10D that you've figured out. Please note that although there is always room for user error , I was shooting the exact same picture with the Canon 20D and it was getting the exposure I expected. Thanks for your help. Yep your Canon is overexposing......Welcome to the world of correctly calibrated to ANSI standards metering and iso ratings....... ALL kidding aside, the best place to start is in the fact that the meters are diffent, not necessarily one right or wrong.... Start by reading this paper regarding Canon and by thinking that the Pentax meter is just plain dumb in the old school sense. It will not "second guess" what your shooting except possibly in Matrix metering mode. which unfortunately is a bit primitive compaired to say Nikon w/ it's 10000 entry look up table.... Canon exposure meterhttp://doug.kerr.home.att.net/pumpkin/Exposure_Calibration.pdf To quote: (Mr. Kerr has more papers that are quite interesting and I recommend reading them) So that leaves Canon with only the possibility of achieving their target value of Hu/Hsat by using a non-standard rating of ISO sensitivity—a rating that is about 0.74 that which would be determined under ISO 12232. In other words, the sensitivity that is designated “ISO 100” by Canon would probably be rated at about ISO 135 under ISO 12232. Sooo this means, in general the Canon overexposes in regard to a standard. Now this is not a bad thing, most seem to find it helpful BUT it is not the only way to play the game.

08-01-2007, 07:17 PM	#5
MJB DIGITAL Veteran Member Join Date: Jul 2007 Location: st. louis Posts: 1,170	thanks jeff!!

08-01-2007, 08:23 PM	#6
pmperry Forum Member Join Date: Jul 2007 Location: Tampa, Florida Posts: 50	I just made a similar jump and I wasn't so surprised as Nikon's system was a similar experience. One thing I did notice after selling my 20D was that the manuals for the XT, XTi, 20D, 30D and 5D all used the term weighted for every metering mode with the exception of Evaluative. I found that term interesting because it implies that the metering isn't exclusive (such as spot metering) when it should be. Instead it implies that the camera is using the whole evaluative system and biasing towards your 3.5% spot area (for the 30D and 5D) which is kind of a cheap implementation. Also, not that your older MF Lenses are limited to Center Weighted Metering (I believe it is the MF lenses but I'll have to double check) as noted in the cameras manual. One thing you can do though is use the Digital Preview mode to get a good look at the histogram first as the stop down metering seems less accurate than those that actually speak to the camera. Anyway, I hope this helps.

08-01-2007, 08:35 PM	#7
pmperry Forum Member Join Date: Jul 2007 Location: Tampa, Florida Posts: 50	Originally posted by jeffkrol Yep your Canon is overexposing......Welcome to the world of correctly calibrated to ANSI standards metering and iso ratings....... ALL kidding aside, the best place to start is in the fact that the meters are diffent, not necessarily one right or wrong.... Start by reading this paper regarding Canon and by thinking that the Pentax meter is just plain dumb in the old school sense. It will not "second guess" what your shooting except possibly in Matrix metering mode. which unfortunately is a bit primitive compaired to say Nikon w/ it's 10000 entry look up table.... Canon exposure meterhttp://doug.kerr.home.att.net/pumpkin/Exposure_Calibration.pdf To quote: (Mr. Kerr has more papers that are quite interesting and I recommend reading them) So that leaves Canon with only the possibility of achieving their target value of Hu/Hsat by using a non-standard rating of ISO sensitivity—a rating that is about 0.74 that which would be determined under ISO 12232. In other words, the sensitivity that is designated “ISO 100” by Canon would probably be rated at about ISO 135 under ISO 12232. Sooo this means, in general the Canon overexposes in regard to a standard. Now this is not a bad thing, most seem to find it helpful BUT it is not the only way to play the game. While this certainly helps to explain some of it, it doesn't explain why a 20D using Partial Metering wouldn't require any EV boost when pointed at caucasian skin.

08-01-2007, 08:45 PM	#8
Ole Administrator Join Date: Sep 2006 Location: Copenhagen, Denmark Photos: Gallery \| Albums Posts: 4,408	It could explain that. If it is correct that ISO 100 set on the Canon really is ISO 135 then that discrepancy corresponds to a "built in" positive EV compensation that you were not aware of. If you dial in +0.5 or +1 EV on the Pentax you should get similar results as the Canon

08-02-2007, 05:42 AM	#9
pmperry Forum Member Join Date: Jul 2007 Location: Tampa, Florida Posts: 50	Originally posted by Ole It could explain that. If it is correct that ISO 100 set on the Canon really is ISO 135 then that discrepancy corresponds to a "built in" positive EV compensation that you were not aware of. If you dial in +0.5 or +1 EV on the Pentax you should get similar results as the Canon True, it could, and if you read DPReview they note that the Canon's are actually ISO 125 for 100 and they're pretty much like this through the whole range. Anyway, even a 3rd of a stop wouldn't make up for a whole stop so the images should look 2/3 underexposed. I think it's a mixture of both issues where Canon has their ISO tuned wrong and their dedicated metering modes aren't dedicated at all but rather just a stronger bias.

08-02-2007, 06:04 AM	#10
jeffkrol Veteran Member Join Date: Sep 2006 Location: Wisconsin USA Photos: Gallery Posts: 8,434	Originally posted by pmperry True, it could, and if you read DPReview they note that the Canon's are actually ISO 125 for 100 and they're pretty much like this through the whole range. Anyway, even a 3rd of a stop wouldn't make up for a whole stop so the images should look 2/3 underexposed. I think it's a mixture of both issues where Canon has their ISO tuned wrong and their dedicated metering modes aren't dedicated at all but rather just a stronger bias. This user test gives some validity to the 'correct" metering w/ the G10 vs a hand held meter. There is a difference but the ANSI standard has about a 1/2 stop "fudge factor" built in so..... My 2 Pence...: Pentax SLR Talk Forum: Digital Photography Review And if I remember correctly the camera "rounds" f stops to the nearest whole(?) stop but it is infinitly stepped in practice. As to the Caucasion skin thing, I actually assume the Canon meter is more sofisticated than the Pentax (old fashioned so to speak) as to things like "fuzzy logic" ect. and could have a built in adjustment circuit hooked to all metering modes but that's just a guess. Again Mr. Kerr has an interesting paper: http://doug.kerr.home.att.net/pumpkin/300D_metering_DAK.pdf

08-02-2007, 09:02 AM	#12
switters Veteran Member Join Date: Jul 2007 Posts: 346 Original Poster	Okay, I've noticed something really interesting. I took several test shots with the K10D and imported them into Aperture. I had the camera set at normal exposure (no compensation), and as I reported earlier in the thread the histogram would often indicate underexposure. However, when I imported these shots and opened them up in Aperture, they appear to be properly exposed! The tonal values go right up to the right side of the histogram, with a nice even distribution across the range. (I'm actually very impressed with the natural rendering of tones in comparison with the 20D). Why is this discrepancy occurring (where the histogram in camera is significantly different than the histogram in my RAW processor - Aperture)? Does it have something to do with the K10D's in-camera display? Has anyone else noticed such a discrepancy? I guess I'll have to learn how to "translate" my camera's histogram results into what will actually appear on my computer.

08-02-2007, 09:02 AM	#13
arbutusq Veteran Member Join Date: Sep 2006 Location: Vancouver BC canada Photos: Gallery Posts: 380	Be patient and give it time. When I switched from Nikon I was bugged by the meter, but now that I know it it works well, and far better than Nikon at preserving highlights. It is easily fooled by bright objects but each meter has its quirks, once you know them you're fine. Another thing to be aware of is the tone curve applied to the RAW data in camera. Applying a steeper tone curve can change the apparant exposure by quite a bit.

08-02-2007, 09:13 AM	#14
switters Veteran Member Join Date: Jul 2007 Posts: 346 Original Poster	Ahhh. So the tone curve you're referring to could affect the apparent exposure, and thus histogram, in the camera? And that would explain the difference between the in-camera histogram and what appears in my RAW converter? You're absolutely right, I'm sure - I just need to give it some time. And don't get me wrong - I love the Pentax K10D! Today I receive a gaggle of lenses to try - 14/2.8, DA 21, DA 40, 43/1.9, 50/1.4, DA 70. I'm gonna be a busy guy!

08-02-2007, 11:58 AM	#15
pmperry Forum Member Join Date: Jul 2007 Location: Tampa, Florida Posts: 50	Originally posted by jeffkrol This user test gives some validity to the 'correct" metering w/ the G10 vs a hand held meter. There is a difference but the ANSI standard has about a 1/2 stop "fudge factor" built in so..... My 2 Pence...: Pentax SLR Talk Forum: Digital Photography Review And if I remember correctly the camera "rounds" f stops to the nearest whole(?) stop but it is infinitly stepped in practice. As to the Caucasion skin thing, I actually assume the Canon meter is more sofisticated than the Pentax (old fashioned so to speak) as to things like "fuzzy logic" ect. and could have a built in adjustment circuit hooked to all metering modes but that's just a guess. Again Mr. Kerr has an interesting paper: http://doug.kerr.home.att.net/pumpkin/300D_metering_DAK.pdf Hey that was my DPReview Post!