[UncleVanya]

I’m not so certain it is strange behavior. I think you are experiencing Exposure Value algorithm decisions which affect multiple parameters beyond just Av, Sv and ISO, and are controlled by multiple Menu settings beyond just eDials. It is a highly complex subject. Full EXIF exports will help track down the variables.

My impression is that this is with the camera set to M mode but I could be wrong - perhaps Av? I agree we need to be clear that ISO and highlight protection and so on are the same in both shots.

[monochrome]

My impression is that this is with the camera set to M mode but I could be wrong - perhaps Av? I agree we need to be clear that ISO and highlight protection and so on are the same in both shots.

The OP originally asked about different Iris size when Aperture value was selected in A on the lens in M camera mode versus setting the aperture on the lens manually while in M Mode. The Exposure Value and White Balance questions came later.

My observation related to such things as selecting MTF Priority in P Mode (if there is a lens Profile for an A20/2.8) versus one of the other Priority Profiles (Landscape, Portrait, Action and the like). Those Mode-dependent choices are embedded throughout the Menu for all camera Modes.

In essence I’m suggesting it’s possible the camera selects a slightly different Aperture in A on the lens to make what it thinks will be a good Exposure based on automatic settings other than Sv and ISO.

[UncleVanya]

The OP originally asked about different Iris size when Aperture value was selected in A on the lens in M camera mode versus setting the aperture on the lens manually while in M Mode. The Exposure Value and White Balance questions came later.

My observation related to such things as selecting MTF Priority in P Mode (if there is a lens Profile for an A20/2.8) versus one of the other Priority Profiles (Landscape, Portrait, Action and the like). Those Mode-dependent choices are embedded throughout the Menu for all camera Modes.

In essence I’m suggesting it’s possible the camera selects a slightly different Aperture in A on the lens to make what it thinks will be a good Exposure based on automatic settings other than Sv and ISO.

That makes sense.

[stevebrot]

I’m not so certain it is strange behavior. I think you are experiencing Exposure Value algorithm decisions which affect multiple parameters beyond just Av, Sv and ISO, and are controlled by multiple Menu settings beyond just eDials. It is a highly complex subject. Full EXIF exports will help track down the variables.

Agreed. Unless the camera is confirmed to be in fully manual mode with no dynamic range automation or exposure compensation (auto or manual), flash-on, or auto-ISO, it is hard to say what is happening. The example photos of the diaphragm show a difference of about 1 stop between body-controlled and aperture ring controlled. A full stop might indicate intent.

The problem can't be the sticky blades because is related to the size of the iris, the second test checks this.

I beg to differ based on a fair amount of experience with vintage lenses and troubleshooting experience on this site. Your images show different opening sizes. They do not indicate why. It is quite possible that there is an issue with the aperture actuation linkage that results in an improper opening size at exposure time. After all, the lens is probably at least 30 years old.

If it were me doing the testing, I would do the set up in M mode with fixed ISO, no EV comp, no auto-DR adjust, and no auto EV comp. I would then do a blank wall/constant light exposure series with equivalent EV through the range of available apertures with the expectation of similar capture histogram for all exposures. I would also test for physical resistance in the mechanism first since it is easy to do and becomes the stop point for fault identification if the movement is anything other than smooth and snappy within the range. There is no sense looking further.

To be clear, unless similar problems have been seen with other lenses, it is highly unlikely this is a problem with the camera body. If the lens is a recent purchase and fault may be reasonably assigned, a return to the seller might be in order. Otherwise, repair might be appropriate.


Steve

---

Last edited by stevebrot; 08-28-2018 at 03:37 PM.

[Not a Number]

I'm still not sure on what the flow is here.

Assuming - M mode.

1. Ring set to "A"
2. F-stop set with eDial (e.g. f/8.0)
3. Exposure zeroed using the shutter speed eDial or Green button (e.g. 1/125th sec)
4. Exposure Taken

1. Ring set to above F-stop (e.g f/8.0)
2. Shutter should still be 1/125th or adjusted to same with eDial
3. Exposure taken

Problem: Exposures differ by perhaps 1 or so stops. Aperture closes down less with ring set to "A".

Does it only do this with the one lens? I tried this with various Pentax-A lenses on my K10D and there is no noticeable difference. But a K10D is quite a few generations separated from a K-1ii.

What happens if you set the aperture to f/5.6 or f/11?

If the problem behavior is limited to the one one lens odds are the lens is at fault.

Comments: Sticky blades is most likely the issue. With the ring in "A" the aperture control lever is only going to move as far as it needs to to close down the iris to the set f-stop. With the ring off "A" mode the control lever is going to move to the fully closed position relying on the stop in the lens ring mechanism to keep the iris open at f/8.0. Force equals Mass times Acceleration. The lever travels a longer distance = more acceleration = more Force.

[Andrea K]

Dear all.

It is impossible for a software setting such the D-Range to change time or aperture values, thereby having an effect on movement or DOF creativity.

Here the PEF of my tests:
f/8 with ring
f/8 with dial
f/8 with ring at hyperfocal focus
f/8 with dial at hyperfocal focus

Here a video of the blades in action:

Youtu.be

Youtu.be

Here a picture of the lever:


Blades are clean and snappy.
The problem there is also with all my other bodies from the SuperA to the z1p, then the lens is the cause. Can I solve this behavior?
I think that also other lenses have not precise iris actuation but in my collection this 20mm seems the worst.

At last my thoughts are that is not normal that the AWB changes using the dial or the ring to set the same exposure with the same metering mode (center weigthed) of the same scene, this is a bug.

Thank you to all.

[D1N0]

I downloaded the first two PEFs they show no difference in white balance. It is exactly the same. When I do an auto levels correction the dial file gets a -0.03 correction and the ring file gets a +0.41 correction. Rawtherapee thinks the dial file was exposed more correctly than the ring file. I tend to think my raws are always a little under exposed (and so thinks light room). There seems to be a discrepancy between mechanical stop down and camera operated stop down by about half a stop. Could be a tolerance issue because of sample variation.

[Andrea K]

I downloaded the first two PEFs they show no difference in white balance. It is exactly the same. When I do an auto levels correction the dial file gets a -0.03 correction and the ring file gets a +0.41 correction. Rawtherapee thinks the dial file was exposed more correctly than the ring file. I tend to think my raws are always a little under exposed (and so thinks light room). There seems to be a discrepancy between mechanical stop down and camera operated stop down by about half a stop. Could be a tolerance issue because of sample variation.

These shots are after the AWB affair, I used for each the Sun WB because in this way the only difference is the exposure. Mind that the real big problem isn't the different "luminosity" but the wrong DOF due to an incorrect iris hole when the ring is in A position and I use the dial to set the aperture value.

[Not a Number]

Some of comments on the video.

You are "snapping" the lever. On my K10D with the lens removed and the DOF preview lever actuated you can see the lever move from wide open to the fully closed position (see video). By counting the number of frames it takes about 1/6th or 0.1667 seconds*. WIthout a lens mounted the K10D operates the same as with a non-"A" or ring off the "A" position. WIth an "A" enabled lens the lever would only move as far as it needs to set the desired f-stop. Again, less travel less force to overcome any stickiness/slowness of the iris.

Try letting the lever "snap" only a short distance. Try moving the lever more slowly/gently - following it with your finger instead of "snap" releasing.

Youtu.be

Youtu.be

* Granted the K10D uses the older electromagnet aperture control block rather then the motor driven control on the k-1 series I don't see there being much difference in speed.

[stevebrot]

It is impossible for a software setting such the D-Range to change time or aperture values, thereby having an effect on movement or DOF creativity.

"Impossible" is such a strong word. It happens all the time and depends on camera settings. That is why access to the straight-out-of-camera EXIF is so useful. Where it never happens is M mode with fixed ISO*.


Steve

* On previous flagship and all current models, auto-ISO is not available in M mode. On other models, the camera defaults (silently) to TAv mode.

[stevebrot]

OK...I downloaded your first two examples and made one small change for the sake of comparison using DCU 5. I changed the custom image setting from "Bright" to "Flat. The histograms are virtually identical. I will repeat using dcraw to produce a linear tiff just to check, but I suspect to see similar.

I would have used ACR, but I don't have support for K-1ii PEF with my version.


Steve

[stevebrot]

I will repeat using dcraw to produce a linear tiff just to check, but I suspect to see similar.

Nope, a linear TIFF shows a definitely increase in exposure with the "dial" version...a bit more than about 0.5 stops. Another reason to NOT use DCU, eh?

Which aperture opening (ring vs dial) is more accurate is difficult to determine, though conventional wisdom would suggest the aperture ring since it is based on a mechanical stop.


Steve

[Andrea K]

Nope, a linear TIFF shows a definitely increase in exposure with the "dial" version...a bit more than about 0.5 stops. Another reason to NOT use DCU, eh?

Which aperture opening (ring vs dial) is more accurate is difficult to determine, though conventional wisdom would suggest the aperture ring since it is based on a mechanical stop.


Steve

No, isn't difficult to determine. You can check the hyperfocal focus samples, here you can see that the by ring setting is the right.

---------- Post added 30-08-18 at 10:32 ----------

OK...I downloaded your first two examples and made one small change for the sake of comparison using DCU 5. I changed the custom image setting from "Bright" to "Flat. The histograms are virtually identical. I will repeat using dcraw to produce a linear tiff just to check, but I suspect to see similar.

I would have used ACR, but I don't have support for K-1ii PEF with my version.

Steve

Please, the problem isn't the luminosity or the colors, even if it is annoying to have a wrong exposure, but is the DOF.

---------- Post added 30-08-18 at 10:34 ----------

"Impossible" is such a strong word. It happens all the time and depends on camera settings. That is why access to the straight-out-of-camera EXIF is so useful. Where it never happens is M mode with fixed ISO*.


Steve

* On previous flagship and all current models, auto-ISO is not available in M mode. On other models, the camera defaults (silently) to TAv mode.

I'm speaking about time and aperture values that are the base of the photography. These settings aren't affect by software options.

[wtlwdwgn]

Andrea, I sent two images to you by PM. I just think your copy of the A 20 needs to be serviced. Eric is the go to guy we like to send our gear to.

[stevebrot]

No, isn't difficult to determine. You can check the hyperfocal focus samples, here you can see that the by ring setting is the right.

OK, which opening is the "real" f/8? (neither?)

I'm speaking about time and aperture values that are the base of the photography. These settings aren't affect by software options.

I would generally agree, though I suggest that you spend a few years answering help requests on this site before saying something never happens that is counter to expectation.

I agree that the lens as the most likely source of the variance and that it is more likely that the aperture lever calibration is at fault. Whether this may resolved by a competent repair person is hard to say. It is unfortunate that this has resulted in uncertainty in regards to DOF with what is considered an ultra-wide angle lens.* If the lens is a recent purchase, you may want to be in conversation with the seller to arrange a return.


Steve

* Fiendishly difficult to control due to short focus throw and resulting abrupt transition.