p.2 #1 · Sony silently applying gain at certain apertures
ImDaJim wrote:
ruthenium -> I used the same ISO and shutter speed on both aperture settings obviously. And as you can see it happens in real time on video at fixed settings so it's certainly not gonna show up in RAW files since it's applying separate gain somewhere else in the processing pipeline. But you're right I should test it with two separate RAW files. I'll try to do that test ASAP.
I am not sure if this misunderstanding or deliberate but your polite responses are repeatedly evasive. I asked "Please, tell what was your shutter speed and ISO with f/1 and then with f/1.1" and you responded "the same" but you didn't respond to the question.
Who cares what a camera display shows when "it's certainly not gonna show up in RAW files"? But, you say this when you in fact haven't looked at the raw files.
Your video shows extreme noise. Why is that?
When reporting a problem attributed to a camera, this report must be specific. Some of the key camera settings need to be stated (shutter speed, ISO). The raw files can be most useful. Make the two raw files, taken with f/1 and f/1.1, available, please.
p.2 #2 · Sony silently applying gain at certain apertures
I said I don't have the lens with me currently and can't do tests with RAW pics. I'm not dismissing your request or methodology, and I'm willing to make that RAW test as soon as I have the lens with me. I'm just saying the video footage itself demonstrates the behavior pretty clearly.
Only changing the aperture, in full manual mode, without touching any other parameter, shows the exposure immediately adjusting when I close the diaphragm and resetting when I go back to f/1.0. This exposure adjustment seems totally independent from any ISO or shutter speed setting.
"Your video shows extreme noise. Why is that?" -> If you just read my reply entirely and looked at the info I provided in the video itself, you'd see I literally boosted the exposure +10 EV in post to point out the change in noise pattern, indicating there's some really involved / complex image processing going on, not just a shift in shutter speed value or ISO.
For more context, users have already discussed Sony cameras secretly applying lens corrections even when the settings were turned off. I also think I found the exact post Jonas B was referring to:
p.2 #3 · Sony silently applying gain at certain apertures
Canon DSLRs did this. I read about it around 5D - 5D2 era.
The sensor could not take much advantage of wider aperture than f1.4.
So gain was added at wider apertures so, if using Av, Shutter speed would match the expected value.
I don't think this was only done by Canon.
Microlens design was the reason and microlenses did improve from 5D to 5D2 meaning less gain was required.
p.2 #4 · Sony silently applying gain at certain apertures
Choderboy wrote:
Canon DSLRs did this. I read about it around 5D - 5D2 era.
The sensor could not take much advantage of wider aperture than f1.4.
So gain was added at wider apertures so, if using Av, Shutter speed would match the expected value.
I don't think this was only done by Canon.
Microlens design was the reason and microlenses did improve from 5D to 5D2 meaning less gain was required.
Thank you Choderboy, I recall that from when the 5D ("classic") was my main camera. (At that time anything was possible eith a mirror and 12MP.)
p.2 #5 · Sony silently applying gain at certain apertures
ImDaJim wrote:
[...]
For more context, users have already discussed Sony cameras secretly applying lens corrections even when the settings were turned off. I also think I found the exact post Jonas B was referring to:
Hi,
Every post has a link. Can you find that post again and post the link (rather than a screenshot). At the top left of your post you can see the text "p.2 #2". That's the link, please right click and post it. I'm by now pretty curious about the function of my shady memory.
p.2 #6 · Sony silently applying gain at certain apertures
Choderboy -> Yep, I actually found more info on this while doing a deep search yesterday evening. This article for instance, was heavily criticized for its overly dramatic conclusions (any lens faster than f/2.0 is almsot useless... lol) but it was technically right about cameras actively trying to compensate for assumed light loss. It was discussed here and someone could even replicate the behavior by taking two pictures at the same settings, with and without contacts.
p.2 #7 · Sony silently applying gain at certain apertures
Most digital cameras silently increase the sensor gain (through a variety of methods) at wide f-numbers to compensate for their sensors having worse sensitivity to light that strikes at an angle.
With electronic-communicating ultrafast lenses, simply unmounting them so the camera doesn't know the current aperture can disable this.
I remember having someone with a Fuji 50/1 doing that and finding around a stop of compensation, which is lost dynamic range, since the highlights got clipped in the raws after boosting exposure.
Canon at least has headroom in its raws between the white point and 2^14 so you can recover some of the highlights.
p.2 #8 · Sony silently applying gain at certain apertures
ImDaJim wrote:
Choderboy -> Yep, I actually found more info on this while doing a deep search yesterday evening. This article for instance, was heavily criticized for its overly dramatic conclusions (any lens faster than f/2.0 is almsot useless... lol) but it was technically right about cameras actively trying to compensate for assumed light loss. It was discussed here and someone could even replicate the behavior by taking two pictures at the same settings, with and without contacts.
p.2 #9 · Sony silently applying gain at certain apertures
Thanks a lot for the valuable info CarVac. 1 stop of compensation is pretty crazy... in my case the boost seems to only be in the 0.3-0.4 EV zone, so not as dramatic but still an issue for my use.
By now the only question remaining is, how do more recent Sony bodies handle this? I wish someone could do the test with an a7V or similar. My gut feeling is that this kind of behavior is still hard-coded in Sony's current firmware, and sadly there's no way to disable it.
p.2 #10 · Sony silently applying gain at certain apertures
ImDaJim wrote:
I said I don't have the lens with me currently and can't do tests with RAW pics. I'm not dismissing your request or methodology, and I'm willing to make that RAW test as soon as I have the lens with me. I'm just saying the video footage itself demonstrates the behavior pretty clearly.
Only changing the aperture, in full manual mode, without touching any other parameter, shows the exposure immediately adjusting when I close the diaphragm and resetting when I go back to f/1.0. This exposure adjustment seems totally independent from any ISO or shutter speed setting.
"Your video shows extreme noise. Why is that?" -> If you just read my reply entirely and looked at the info I provided in the video itself, you'd see I literally boosted the exposure +10 EV in post to point out the change in noise pattern, indicating there's some really involved / complex image processing going on, not just a shift in shutter speed value or ISO.
For more context, users have already discussed Sony cameras secretly applying lens corrections even when the settings were turned off. I also think I found the exact post Jonas B was referring to:
So I wouldn't even be surprised at all if a similar hidden processing was going on behind the scenes....Show more →
This is a technical forum, and a serious technical discussion requires some data. You speculate instead: "Sony silently applying gain at certain apertures." Yet, you have not even looked at the raw files, and have been unable to tell the shutter speed and the ISO in your observations. This specualtive approach is a real problem.
Another problem is that what you report LITERALLY makes sense, however you seem to imply a problem. Just above, you wrote, quote: "Only changing the aperture, in full manual mode, without touching any other parameter, shows the exposure immediately adjusting when I close the diaphragm and resetting when I go back to f/1.0. This exposure adjustment seems totally independent from any ISO or shutter speed setting."
Exposure is under the exclusive control of the aperture and shutter speed (ISO does not directly affect exposure, by definition). Therefore this is perfectly normal that changing the aperture should result in a change in exposure. This is also perfectly normal that, in full manual, the change in exposure should be "totally independent from any ISO or shutter speed setting." I sense that you wanted to describe a problem, yet you have not succeeded, and literally reported the expected behaviour.
I am not being unfriendly. What I am saying is that you must report your observations succinctly and don't speculate. Share as much technical details as possible, especially the key settings: shutter speed, ISO. Repeat the comparison of f/1 vs f/1.1 in the aperture-priority mode with auto-ISO. Finally, share the raw files. The raw files are important for two reasons. First, one can replicate your observations (or fail to replicate); second is that one can compare the EXIF and see if indeed "Sony silently applying gain at certain apertures" - this must be written in the EXIF.
p.2 #11 · Sony silently applying gain at certain apertures
ruthenium wrote:
[...] second is that one can compare the EXIF and see if indeed "Sony silently applying gain at certain apertures" - this must be written in the EXIF.
I agree with you about the methodology and the need of being careful. Personally I'm not that upset, or surprised, though as I have seen this happen before.
Your last claim about any change of gain "must be written in EXIF" may be true or not. I wouldn't put it that way as I suspect the gain is _silently_ increased. A careful test setup will tell.
p.2 #12 · Sony silently applying gain at certain apertures
Ruthenium, I appreciate you trying to help but really, your two last replies suggest you're not properly reading my posts and didn't watch the video I've put together. I'd be great if you were less condescending and applied the same scrutiny you expect from others
Is a ~0.4 EV *increase* in exposure across the entire image when *stopping down* the lens normal behavior to you? All these explanations about how exposure works come across as if I were completely clueless about photography, when you could simply try to understand what I'm facing first and not quote me without any context. I might not be the best at explaining stuff but I think everyone here understood the nature of the issue.
p.2 #13 · Sony silently applying gain at certain apertures
ImDaJim wrote:
Ruthenium, I appreciate you trying to help but really, your two last replies suggest you're not properly reading my posts and didn't watch the video I've put together. I'd be great if you were less condescending and applied the same scrutiny you expect from others
Is a ~0.4 EV *increase* in exposure across the entire image when *stopping down* the lens normal behavior to you? All these explanations about how exposure works come across as if I were completely clueless about photography, when you could simply try to understand what I'm facing first and not quote me without any context. I might not be the best at explaining stuff but I think everyone here understood the nature of the issue....Show more →
I am not condescending, neither I am being unfriendly. I can repeat my questions, and would appreciate it if you answer these questions:
1) What was the shutter speed and ISO with f/1 and f/1.1 respectively? - tell the numbers, please.
2) Are all your observations all about visual changes in brightness on the camera back screen? - Yes, or no?
3) Have you compared the raw files obtained with f/1 and f/1.1 - Yes, or no?
4) You notion of "a ~0.4 EV *increase* in exposure across the entire image when *stopping down* the lens" - is this a visual impression that you formed by qualitatively evaluating the brightness of the images with f/1 and f/1.1 on the camera screen? I am asking this to understand whether you actually refer to the exposure (on the sensor), or might be referring to changes in how light or dark you saw the images on the screen of the camera. Please, don't take this as a condescending remark, I am sure you know that how light or dark an image appears on the screen of a camera may have nothing to do with the exposure.
Questions of this kind are normal in a technical discussion, and can help understanding the problem -I hope you should be able to answer these for me - thank you!
p.2 #14 · Sony silently applying gain at certain apertures
ruthenium -> Happy to reply to any questions.
1) If you're asking about the video test, then it was ISO 100 and 1/400. Checked with Catalyst Browse. Both of those were unchanged during the whole test since I was in manual mode and only closing / opening the aperture.
2) Yes, it's all "visual" if you want to put it this way, since the end result is an exposure change. It's not only on the camera screen but also baked in the final video file, as you can see.
3) Once again, I haven't done RAW file comparisons yet. Something I did though, as said in the initial post, is un-mounting the lens to remove communication. I did this while the correction was being applied and guess what, the exposure reset as soon as I disconnected the lens. It came back when reconnecting it.
4) It's just an estimate of what I think the increase in brightness roughly is, but this doesn't matter really. This being accurate or not doesn't change anything to the behavior itself. The increase is consistent and reproducible. A/B testing with and without the lens chip connected shows something happens regardless of what my settings are.
Nothing wrong with asking questions but again I have the impression you're trying to discuss the very existence of the issue where I already established the phenomenon in a purely empirical way. You're right to apply Occam's razor here as this seems totally illogical at first, but we already have proof of this happening (see previous messages).
My main concern was more about how other Sony cameras handle this, as I'll soon be switching to whatever Sony decides to release for the next gen of a7s / FX cameras, and this behavior truly bothers me for video work in low light. I'd hate having to put electrical tape on the contacts and lose the benefits of EXIF (focus distance for IBIS, aperture info...).
p.2 #15 · Sony silently applying gain at certain apertures
Sony A1, fw 4.00, Voigtländer 50/1.2 SE
for each exposure there is a f-number and under that the averaged level of the pixels (in aRGB before converting to web-fomat and sRGB).
I guess the difference between f1.0 and 1.0 and f1.1 and 1.1 are bigger.
p.2 #16 · Sony silently applying gain at certain apertures
Thanks a million for this test Jonas! I just measured the exposure difference in your pics and it's about 0.2 EV brighter with the compensation. I also measured my own difference at f/1.1 and it's right at +0.4 EV, so my estimate was accurate.
Your assumption that the compensation is bigger at wider apertures is correct, Sony likely overrides the ISO by different amounts depending on the lens max aperture. It assumes the faster the lens, the more loss due to light hitting at more oblique angles. As to why it "turns off" wide open in my case, it's maybe simply because the firmware doesn't know what to do with f/1.0 as this value wasn't part of the expected apertures at the time of coding. Would be interesting to test my lens on a recent body.
EXIF showing the exact same ISO... what a surprise It's indeed an entirely silent process in the imaging pipeline that can't be avoided unfortunately. Really disappointed it still does that on your a1. This behavior should be part of the lens compensation settings in the menu and not a hidden override you can't even control. Come to think of it, maybe it could be the case in the newer firmware? What are your lens comp settings?
p.2 #17 · Sony silently applying gain at certain apertures
The term for what's being discussed is "pixel vignetting", which is the loss of light at the pixel well from the oblique light angles from large aperture lenses. The amount of vignetting is a function of pixel pitch, the smaller the pitch (ie, the higher the resolution), the more loss of light.
With the A7s II being a native 12MP FF sensor (not quad bayer 48MP -> 12 MP like the A7s III), I would expect its sensor to have relatively modest amounts of pixel vignetting as compared to higher-resolution sensors.
p.2 #18 · Sony silently applying gain at certain apertures
ImDaJim wrote:
[...]
EXIF showing the exact same ISO... what a surprise It's indeed an entirely silent process in the imaging pipeline that can't be avoided unfortunately. Really disappointed it still does that on your a1. This behavior should be part of the lens compensation settings in the menu and not a hidden override you can't even control. Come to think of it, maybe it could be the case in the newer firmware? What are your lens comp settings?
I forgot to mention that all lens compensating items in the menu are disabled (always). The crops I posted above are about 5% av the total image area and taken from the dead center of the images.
NB: My A1 is v1 so far from the newest of all Sony cameras. Was it 4 or 5 years ago now it was announced?
p.2 #19 · Sony silently applying gain at certain apertures
Jonas B -> Yep, it was released in 2021. Just out of curiosity, what are the lens comp parameters available when you go to that page? Is it still vignette, distortion and aberration only?
p.2 #20 · Sony silently applying gain at certain apertures
ImDaJim wrote:
Jonas B -> Yep, it was released in 2021. Just out of curiosity, what are the lens comp parameters available when you go to that page? Is it still vignette, distortion and aberration only?
Vignette, distortion and CA, that's correct if you shoot still images. The anti distortion setting is sometimes locked to Auto, something that depends on the lens.
In video mode there is also a choice for focus breathing showing (also lens dependable).
It's indeed an entirely silent process in the imaging pipeline that can't be avoided unfortunately. Really disappointed it still does that on your a1. This behavior should be part of the lens compensation settings in the menu and not a hidden override you can't even control. Come to think of it, maybe it could be the case in the newer firmware? What are your lens comp settings?