[homestead65]

My K10D underexposes in bright light only,with all metering modes. cloudy,dull light, metering is perfect,but in sunlight it can underexpose by up to one and a half stops. Is this normal?

[Marc Sabatella]

It's normal for people to expect a sunny day picture to look brighter than a cloudy day picture. But it's not correct. The camera has no idea one day is brighter than the other. it tries to make everything *average*. So bright scenes come out the same as dark scenes. Another issue is that sunny days will have cast shadows, and what ends up creating the right "average" exposure for the scene will obviously result in the shadow portion being darker than average. And a third issue is that bright days often include bright reflections, and the camera desperately tries to avoid overexposing those - resulting in the resr of the picture being underexposed by comparison.

None of this is K10D specific - it's common to pretty much all cameras that aren't specifically designed for beginners. Cameras specifically designed for beginners do often try to make
"guesses" as to what you want and will deliberately overexpose portions of the picture or deliberately reduce the contrast in the scene to avoid these problems. But most SLR's expect you to take control and make those adjustments yourself as you see fit.

There are lots of books out there that explain how exposure and metering work and the things you can do in order to get the results you want; I'd suggest a visit to the local library or bookstore. Lots of people specifically recommend "Understanding Exposure".

[Ron Kruger]

Marc, you have a way of shortening or ending threads with comprehensive and expertly stated replies. It's hard to add anything useful or constructive behind you.
The only thing I might add, is that some are never satisfied with bright day shots simply because the dynamic range of a camera (any camera) is not as wide as the human eye. While exposing to avoid blowing the highlights, the shadows are filled in more than what we see. Personally, I like this. Adds more contrast and richer colors, but you can regain some of this shadlow detail by adjusting the shadow/highlights or levels in PP.
Otherwise, time your shots for when the light angles are low and the sun is behind you.

[Kerrowdown]

Also shooting in RAW kinda gives flattish images which can always benefit from a wee tweak in PP.

[Lowell Goudge]

The only thing I would add is that in the past many cameras , in center weighted metering metered less off the upper half of the frame to compensate for the bright sky. I don't know about dslr's but you could test this by taking a shot right way up and the same shot upside-down and compare metering or histograms

[lol101]

It's normal for people to expect a sunny day picture to look brighter than a cloudy day picture. But it's not correct. The camera has no idea one day is brighter than the other. it tries to make everything *average*. So bright scenes come out the same as dark scenes. Another issue is that sunny days will have cast shadows, and what ends up creating the right "average" exposure for the scene will obviously result in the shadow portion being darker than average. And a third issue is that bright days often include bright reflections, and the camera desperately tries to avoid overexposing those - resulting in the resr of the picture being underexposed by comparison.

None of this is K10D specific - it's common to pretty much all cameras that aren't specifically designed for beginners. Cameras specifically designed for beginners do often try to make
"guesses" as to what you want and will deliberately overexpose portions of the picture or deliberately reduce the contrast in the scene to avoid these problems. But most SLR's expect you to take control and make those adjustments yourself as you see fit.

There are lots of books out there that explain how exposure and metering work and the things you can do in order to get the results you want; I'd suggest a visit to the local library or bookstore. Lots of people specifically recommend "Understanding Exposure".

I have used 2 K10 and metering on both units was far from perfect.

On the one I presently use, I have to stick to a permanent +0.7EV adjustment and even with that it underexposes under bright light too.

By underexposing, I mean that the histogram is almost completely located in the lower 2/3 of the range with nothing in the upper 1/3.

Things get even worse when a reflection or strong highlight is present in the frame.

The K10 is kind of a fine camera (its IQ is really excellent up to ISO 400) but AF and exposure are no less than flaky. So it is K10 specific (all IMO obviously) since K20 and subsequent models don't seem to suffer from this problem.

[jeffkrol]

I have used 2 K10 and metering on both units was far from perfect.

On the one I presently use, I have to stick to a permanent +0.7EV adjustment and even with that it underexposes under bright light too.

By underexposing, I mean that the histogram is almost completely located in the lower 2/3 of the range with nothing in the upper 1/3.

Things get even worse when a reflection or strong highlight is present in the frame.

The K10 is kind of a fine camera (its IQ is really excellent up to ISO 400) but AF and exposure are no less than flaky. So it is K10 specific (all IMO obviously) since K20 and subsequent models don't seem to suffer from this problem.

That's the tone curve mostly, not the exposure.....
Pentax seems to use notoriously tight tone curves... especially in the "upper"
models. It's NOT a problem unless it is inconsistent from other K10's.... it is a philosophy or feature, take your pick....
Image from a review of the k7... and was a small jpg so bear with the "not so perfect" correction. Things work much better on full size jpg or better still RAW..... and converted to TIFF (8 or 16bit).
The first image is pretty spot on what a true "old fashioned" meter would do.
You didn't mention type of metering, which changes things as well.... spot and center weighted (dumb metering modes)would behave similar but Matrix (semi-dumb metering mode)is always a shot in the dark as to WHAT the camera was thinking.

[KungPOW]

IMHO,

The K10D under exposes when used in Matrix Metering. The camera will do everything it can to not blow highlights. The problem comes into play when there is a single very bright light source in one of teh metering segments, but not in any of the others. The single bright spot will pull the exposure down resulting in an under exposed shot for the rest of the frame.

Spot metering and weighted average work just like any other camera. Meter off something you want to be 18% grey, lock the exposure, and shoot away. No exposure issues.

I consider the Matrix metering in the K10D to be one of the week points on the camera. There are other camera models, from the same year, that meter bang on every time. Even in tough light conditions. I consider matrix metering to be the "automatic transmision" of the exposure world. It is there to makes the photographer's job easier. The K10D is more of a semi-automatic, the driver still needs to get it in the correct gear, you just don't need to use a clutch.

[Marc Sabatella]

By underexposing, I mean that the histogram is almost completely located in the lower 2/3 of the range with nothing in the upper 1/3.

That is not the definition of underexposure. Correct exposure does not require anything to be in the top third. That's a technique / religion called "expose to the right" that some adhere to, but it's not really what books on exposure or ISO standards will actually describe as "correct". A grey card or other similarly flat scene, for instance, should meter about a third to half a stop *left* of center on a histogram.

[stevebrot]

A grey card or other similarly flat scene, for instance, should meter about a third to half a stop *left* of center on a histogram.

Marc, this discussion opens a huge can of worms. I did a quick test for several of my lenses metered against a blank wall. Sure enough, as you say, the histogram is usually a bit to the left of center. The amount of skew also varied by lens This was true regardless of exposure mode. I then repeated the series using the built-in flash. Guess what? No skew or slightly to the right when using PTTL flash.

Pulling out the hand-held meter (Gossen Luna Lux) and shooting in M mode, I get the same left skew as with the built-in meter.

Is there any chance that the shift is related to white balance adjustment?


Steve

[jeffkrol]

Marc, this discussion opens a huge can of worms. I did a quick test for several of my lenses metered against a blank wall. Sure enough, as you say, the histogram is usually a bit to the left of center. The amount of skew also varied by lens This was true regardless of exposure mode. I then repeated the series using the built-in flash. Guess what? No skew or slightly to the right when using PTTL flash.

Pulling out the hand-held meter (Gossen Luna Lux) and shooting in M mode, I get the same left skew as with the built-in meter.

Is there any chance that the shift is related to white balance adjustment?


Steve

I'm not Mark but I doubt it... one thing though, you can really only make base comparisons in RAW and software that doesn't pull any tricks..... About the last thing to look at is WB (except in the processing software part).
Since pTTL does a low level pre flash and guesstimates from there there are quite a few reasons it could turn out the way it did for you... most likely it's just programmed that way......
Looks like the K7 is the same: Apparently Pentax treats the flash as a "dumb mode" and intercedes.....
Compare this to the first photo I posted above.
http://www.ephotozine.com/article/Pentax-K7-11613\

[homestead65]

Hey ,thanks very much,i shouldn't have said exposure problem,as i know its a tone curve problem, that can be brought back in photoshop quite nicely,i was just wondering if this was normal for pentax to "protect" the highlights so much.
thanks again

[lol101]

That is not the definition of underexposure. Correct exposure does not require anything to be in the top third. That's a technique / religion called "expose to the right" that some adhere to, but it's not really what books on exposure or ISO standards will actually describe as "correct". A grey card or other similarly flat scene, for instance, should meter about a third to half a stop *left* of center on a histogram.

What I am talking about is K10 rendition of a sunny day where you get dark blue skies with grey clouds (histogram bunched to the left) instead of light blue skies with white clouds (full histogram representing exactly the luminance gradations from dark to light tones, as a handheld meter would give you).

A correct exposure in my book is when whites are rendered white (not grey) and blacks black.

[lol101]

IMHO,

The K10D under exposes when used in Matrix Metering. The camera will do everything it can to not blow highlights. The problem comes into play when there is a single very bright light source in one of teh metering segments, but not in any of the others. The single bright spot will pull the exposure down resulting in an under exposed shot for the rest of the frame.

Same here. Must be that you don't understand exposure.

It also happens when the overall luminosity of the scene is high, resulting in an underexposed shot where "white skies" turn to grey and the rest of the scene gets quite dark. Center weighted would give you mostly the same result than matrix on the K10 (but matrix is supposed to be "smarter" no?).

[jeffkrol]

What I am talking about is K10 rendition of a sunny day where you get dark blue skies with grey clouds (histogram bunched to the left) instead of light blue skies with white clouds (full histogram representing exactly the luminance gradations from dark to light tones, as a handheld meter would give you).

A correct exposure in my book is when whites are rendered white (not grey) and blacks black.

This is more a matter of semantics...you are right that would be the correct exposure but generally that would not be what your meter recommends....... as noted what you are referring to is the same as shooting a white wall or snow... they will not be white based on a "normal" dumb meter. Clouds and sky in certain proportions average out to 110 RGB grey with the "correct" uncompensated meter reading (forgive a tad of a stretch here) therefore without user intervention will be bunched to the left a tad when using the "recommended" settings. Then there is the "preventing blowout" theories.....

There are literally years of discussion on this very "issue"........ and the easiest way to define a large part of it is the meter may be accurate to it's programming but not be correct for the shot or your intent........
Someone else's words a long time ago.........

It was possible to compensate for this easily when converting RAW to JPEG, but that shouldn’t be necessary. I found I was often using the spot meter and AE-L (Autoexposure Lock) in situations where I’ve come to expect a multzone meter to take care of it for me, so it’s something to watch out for.

Pentax K10D Review - Digital SLR Photography
you can spend years discussing "why" when a simple "it is what it is" suffices.......