[WorksAsIntended]

I am sure it is not about white balance. The wb system is pretty simple in how it works and this cannot be produced even when trying to abuse it.

[Digitalis]

he problem is it's very intermittent. The majority of shots are just fine but it randomly over-exposes a shot here and there, but I'll do what I can to keep an eye on things like that.

IF you are using the custom function "link AE to AF point" even slight shifts in composition can cause the AE system to make adjustments in the middle of a sequence of shots.



Pentax K5IIs - SMCP-K 50mm f/1.2 @ f/5.6 1/15th ISO 80

[WorksAsIntended]

Right, by adjuting either shutter time, aperture or iso, which are all the same in this case.

[Lowell Goudge]

The one key point I missed is that the OP is taking the raw images and doing post processing afterwards.

Just a thought but since the camera is not doing the Post processing, by making jpegs, is it perhaps an error in post processing. To determine that, we would need the raw images.

Also RAW also excludes all discussions of things like WB etc.

High frame rate combined with aperture could be the other possible contributor. Nothing else makes sense, as I and others have commented shutter failure is completely different and since all camera settings are the same, shutter, iso, and aperture setting. The only logical remaining possibilities are aperture performance if hardware, or post processing error.

That is, unless the exposure did change due to a real change in lighting, but given who little the birds have moved, I doubt there is that much change in the cloud cover and overall lighting.

[DeadJohn]

My vote is for aperture issue. The more you stop down, the more the aperture needs to move, and if it hasn't closed all the way before the shutter fires you'll get overexposure.

It might be the lens or the camera. Do you get the problem with different lenses? Test with a different lens and stop all the way down to f/22 or f/32 to maximize the chance of the problem occurring.


Your EXIF says DFA 100 WR macro. I had an aperture issue with that same lens. It took 3 tries for Precision Camera to fix it because they weren't paying attention to the note I included with the lens.

[PushTheRedButton]

IF you are using the custom function "link AE to AF point" even slight shifts in composition can cause the AE system to make adjustments in the middle of a sequence of shots.

I checked, and I did have it set to link AE to AF point, but could the flock of birds that far away really make that huge of a difference? After giving it a lot of thought, if the sensor really is that sensitive, I can see how a flock of black birds flying right through the middle of the exposure zone could make sense, as I only use the center focus sensor while shooting sunsets as the bright light messes with the AF system and can cause the lens to zoom all the way in using the full array of focus sensors.

edit: and after thinking about it more, I've taken lots of shots with birds taking up more of the center focus point without this effect happening.

I changed the setting so they're not linked. Next chance I get I'm going to try changing the settings back and forth while focusing on getting a flock right in the middle. I know the K-5 autofocus sensors are pretty big, but I can't recall how many exposure sensors/zones there are.

The one key point I missed is that the OP is taking the raw images and doing post processing afterwards.

Just a thought but since the camera is not doing the Post processing, by making jpegs, is it perhaps an error in post processing. To determine that, we would need the raw images.

Also RAW also excludes all discussions of things like WB etc.

High frame rate combined with aperture could be the other possible contributor. Nothing else makes sense, as I and others have commented shutter failure is completely different and since all camera settings are the same, shutter, iso, and aperture setting. The only logical remaining possibilities are aperture performance if hardware, or post processing error.

That is, unless the exposure did change due to a real change in lighting, but given who little the birds have moved, I doubt there is that much change in the cloud cover and overall lighting.

I mentioned in another reply I noticed the problem while reviewing the shots in-camera. These shots have been exported via software, but not edited, but I attached the raw files themselves anyway.

edit: The raw photos didn't upload. Gonna try again.
edit 2: my raw files are "invalid" (DNG)

My vote is for aperture issue. The more you stop down, the more the aperture needs to move, and if it hasn't closed all the way before the shutter fires you'll get overexposure.

It might be the lens or the camera. Do you get the problem with different lenses? Test with a different lens and stop all the way down to f/22 or f/32 to maximize the chance of the problem occurring.


Your EXIF says DFA 100 WR macro. I had an aperture issue with that same lens. It took 3 tries for Precision Camera to fix it because they weren't paying attention to the note I included with the lens.

So here's why I'm confused about that. Shooting at f/2.8 will leave the lens wide open and the blades are as far apart as they can be. The more the lens is stopped down the close the blades come together, so to speak. Maybe you can explain a little more because I would think the opposite would occur.

[ehrwien]

I checked, and I did have it set to link AE to AF point, but could the flock of birds that far away really make that huge of a difference?

Should not be the cause of the exposure difference here, since your example pictures show the same aperture, exposure time and ISO values. All any exposure automatic does is adjust those three values; you can have one (T or Av with ISO range), two (T or AV with fixed ISO, TAv with ISO range) or all three (M) values locked down, and the camera will adjust the ones that are not locked (none in M). Even exposure compensation doesn't change this, it just adjusts the measurement taken by the exposure meter for your selected metering method in one or the other direction.

edit: The raw photos didn't upload. Gonna try again.
edit 2: my raw files are "invalid" (DNG)

You can't upload RAW files here, you'll have to upload them elsewhere (e.g. Dropbox) and link where they are.

So here's why I'm confused about that. Shooting at f/2.8 will leave the lens wide open and the blades are as far apart as they can be. The more the lens is stopped down the close the blades come together, so to speak. Maybe you can explain a little more because I would think the opposite would occur.

Go into M mode, select an exposure time of a second and set the aperture at around f/16. Look into the front of your lens, press the shutter button and watch. The physical aperture (the hole created by the aperture blades) should go from wide open before you pressed the shutter button, to closed down almost completely for the one second after you pressed the shutter button, and wide open again when the shot is done. (there's also a button for optical preview on some bodies or you could assign a button to trigger the optical preview; it will stop down the aperture to the chosen value (make the hole smaller) for as long as you hold down that button)
This is done so that the camera can use all the light that comes in through a larger aperture (smaller value, like f/2.8) to more accurately focus and meter the scene, and it also allows you to see the scene brighter in your viewfinder than it would be with the lens stopped down.
The camera only stops down the lens right before the shot. When you take multiple images in a short burst, the aperture in the lens may not be fast enough to stop down all the way it should in time for the image, and that's what some people already suggested here and that's what I also think might be causing the difference in exposure here despite showing the exact same exposure values.

[Digitalis]

Go into M mode, select an exposure time of a second and set the aperture at around f/16. Look into the front of your lens, press the shutter button and watch. The physical aperture (the hole created by the aperture blades) should go from wide open before you pressed the shutter button, to closed down almost completely for the one second after you pressed the shutter button, and wide open again when the shot is done. (there's also a button for optical preview on some bodies or you could assign a button to trigger the optical preview; it will stop down the aperture to the chosen value (make the hole smaller) for as long as you hold down that button)

You could do the same thing while looking through the lens and holding the aperture lever down. Upon releasing the lever, the aperture blades should open instantaneously, if there is any delay in the opening of the lens aperture: there is your smoking gun.

checked, and I did have it set to link AE to AF point, but could the flock of birds that far away really make that huge of a difference?

I agree it is unlikely, but each variable in the imaging chain has to be quantified and taken into account. So far the most incriminating variable is the aperture mechanism in the lens you are using. Personally I have seldom encountered any issues with the aperture mechanism with my lenses, but in situations where they have cropped up on my Pentax cameras I have experienced similar issues to yours.

[mikesbike]

Should not be the cause of the exposure difference here, since your example pictures show the same aperture, exposure time and ISO values. All any exposure automatic does is adjust those three values; you can have one (T or Av with ISO range), two (T or AV with fixed ISO, TAv with ISO range) or all three (M) values locked down, and the camera will adjust the ones that are not locked (none in M). Even exposure compensation doesn't change this, it just adjusts the measurement taken by the exposure meter for your selected metering method in one or the other direction.

You can't upload RAW files here, you'll have to upload them elsewhere (e.g. Dropbox) and link where they are.

Go into M mode, select an exposure time of a second and set the aperture at around f/16. Look into the front of your lens, press the shutter button and watch. The physical aperture (the hole created by the aperture blades) should go from wide open before you pressed the shutter button, to closed down almost completely for the one second after you pressed the shutter button, and wide open again when the shot is done. (there's also a button for optical preview on some bodies or you could assign a button to trigger the optical preview; it will stop down the aperture to the chosen value (make the hole smaller) for as long as you hold down that button)
This is done so that the camera can use all the light that comes in through a larger aperture (smaller value, like f/2.8) to more accurately focus and meter the scene, and it also allows you to see the scene brighter in your viewfinder than it would be with the lens stopped down.
The camera only stops down the lens right before the shot. When you take multiple images in a short burst, the aperture in the lens may not be fast enough to stop down all the way it should in time for the image, and that's what some people already suggested here and that's what I also think might be causing the difference in exposure here despite showing the exact same exposure values.

I agree with this^^. But I also agree that having your AF point and exposure metering linked as a regular practice is not a good idea. Only good for certain circumstances you may be dealing with. That is like having a traveling spot meter. The result is not always optimum for the exposure of the overall scene. In this case, however, it is obviously not the cause of the exposure difference between the two shots.

[rogerstg]

It should be easy enough to determine the cause by testing at home. Using a lamp and tripod, try to reproduce the fault using settings when the problem occurred. Then change one thing at a time until you identify the issue; ether by eliminating or worsening the occurrence. It seems there are two possibilities: the aperture is not stopping down entirely, or the trailing curtain of the shutter is slightly delayed.

[DeadJohn]

So here's why I'm confused about that. Shooting at f/2.8 will leave the lens wide open and the blades are as far apart as they can be. The more the lens is stopped down the close the blades come together, so to speak. Maybe you can explain a little more because I would think the opposite would occur.

The camera keeps the lens wide open to gather as much light as possible during autofocus and to give you a bright viewfinder. The camera light meter measures while wide open then estimates how much light will come through at whatever aperture you are set at. The aperture closes as the shutter fires. A photo at f2.8 needs no aperture movement, a photo at f5.6 will have some aperture movement, and a photo at f22 will have a lot of aperture movement from the wide open state.

*If* the lens aperture is moving too slowly, the shutter will fire when the aperture is still moving somewhere between wide open and the actual aperture setting you selected. You'll probably notice it more as you stop down more. Taken to an extreme, a completely broken non-moving aperture won't have any effect f2.8, will be a little overexposed at 3.5, and will be all-white at f/22.

Faster shutter speeds help make a slow aperture more noticeable, since the shutter might open, expose, and close before the aperture had a chance to get into the correct stopped down position.

[stevebrot]

Summary:

• As noted in the original post, the exposure settings for the two images are the same
• Light may have changed between the two exposures, but the two appear to have been taken in quick succession, on the order of 0.5-1.0s between.
• XMP metadata indicates identical processing with the exception of color temp and tint

Color temp/tint were:

Top: 5000K, +6
Bottom: 5150K, +9

Assuming that the color temp shifted due to the change in actual light to the sensor, we are back to the question of improper aperture stop-down or a shutter problem. The two may be differentiated by attempting to reproduce using the same shutter and burst setting in M mode with an evenly lit blank wall as subject, but with the lens diaphragm set wide open. Brightness variation for that case would most likely be due to shutter timing. Lack of such may be then due to lazy aperture in the lens, though intermittent aperture control problems in the body (lever not moving when it is supposed to) cannot be ruled out.


Steve

[mikesbike]

As DeadJohn says. An easy test is to make it happen with continuous on high, then try it on slow or just take a series of shots in AF-S mode. If those turn out fine and it only happens in spray-n-pray mode, the aperture blades are just not able to keep up.

[WorksAsIntended]

Even simplier, "spray-n-pray" on an object very close so you can see a wieder open aperture by a change in bokeh.

[PushTheRedButton]

Should not be the cause of the exposure difference here, since your example pictures show the same aperture, exposure time and ISO values. All any exposure automatic does is adjust those three values; you can have one (T or Av with ISO range), two (T or AV with fixed ISO, TAv with ISO range) or all three (M) values locked down, and the camera will adjust the ones that are not locked (none in M). Even exposure compensation doesn't change this, it just adjusts the measurement taken by the exposure meter for your selected metering method in one or the other direction.

You can't upload RAW files here, you'll have to upload them elsewhere (e.g. Dropbox) and link where they are.

Go into M mode, select an exposure time of a second and set the aperture at around f/16. Look into the front of your lens, press the shutter button and watch. The physical aperture (the hole created by the aperture blades) should go from wide open before you pressed the shutter button, to closed down almost completely for the one second after you pressed the shutter button, and wide open again when the shot is done. (there's also a button for optical preview on some bodies or you could assign a button to trigger the optical preview; it will stop down the aperture to the chosen value (make the hole smaller) for as long as you hold down that button)
This is done so that the camera can use all the light that comes in through a larger aperture (smaller value, like f/2.8) to more accurately focus and meter the scene, and it also allows you to see the scene brighter in your viewfinder than it would be with the lens stopped down.
The camera only stops down the lens right before the shot. When you take multiple images in a short burst, the aperture in the lens may not be fast enough to stop down all the way it should in time for the image, and that's what some people already suggested here and that's what I also think might be causing the difference in exposure here despite showing the exact same exposure values.

You could do the same thing while looking through the lens and holding the aperture lever down. Upon releasing the lever, the aperture blades should open instantaneously, if there is any delay in the opening of the lens aperture: there is your smoking gun.





I agree it is unlikely, but each variable in the imaging chain has to be quantified and taken into account. So far the most incriminating variable is the aperture mechanism in the lens you are using. Personally I have seldom encountered any issues with the aperture mechanism with my lenses, but in situations where they have cropped up on my Pentax cameras I have experienced similar issues to yours.

The camera keeps the lens wide open to gather as much light as possible during autofocus and to give you a bright viewfinder. The camera light meter measures while wide open then estimates how much light will come through at whatever aperture you are set at. The aperture closes as the shutter fires. A photo at f2.8 needs no aperture movement, a photo at f5.6 will have some aperture movement, and a photo at f22 will have a lot of aperture movement from the wide open state.

*If* the lens aperture is moving too slowly, the shutter will fire when the aperture is still moving somewhere between wide open and the actual aperture setting you selected. You'll probably notice it more as you stop down more. Taken to an extreme, a completely broken non-moving aperture won't have any effect f2.8, will be a little overexposed at 3.5, and will be all-white at f/22.

Faster shutter speeds help make a slow aperture more noticeable, since the shutter might open, expose, and close before the aperture had a chance to get into the correct stopped down position.

Summary:

• As noted in the original post, the exposure settings for the two images are the same
• Light may have changed between the two exposures, but the two appear to have been taken in quick succession, on the order of 0.5-1.0s between.
• XMP metadata indicates identical processing with the exception of color temp and tint

Color temp/tint were:

Top: 5000K, +6
Bottom: 5150K, +9

Assuming that the color temp shifted due to the change in actual light to the sensor, we are back to the question of improper aperture stop-down or a shutter problem. The two may be differentiated by attempting to reproduce using the same shutter and burst setting in M mode with an evenly lit blank wall as subject, but with the lens diaphragm set wide open. Brightness variation for that case would most likely be due to shutter timing. Lack of such may be then due to lazy aperture in the lens though intermittent aperture control problems in the body (lever not moving when it is supposed to) cannot be ruled out.


Steve

I completely understand now, and I think everyone talking about stopping down so much affecting it is right. The blades open and close smoothly using the little lever on the inside, but the lens has had a tremendous amount of use, so maybe this is the first sign it needs servicing.

I appreciate everyone's input, thank you all