Strange Color Issues

[inspiredtimothy]

Went out to shoot some on a local trail in mid to late afternoon yesterday with the S5 II and the 14-28 and 28-200. The sun was extremely bright and I encountered a lot of situations where I needed to bracket. On one particular set I have noticed some strange color patterns appearing in highlights, as seen here in the clouds:



Exposure metering is highlight metering +2. Shadows have been adjusted and highlights lowered. It's also present in the neutral +0 shot:



Color differences are due to the red shadow noise that appears when lifting exposure an extra 2 stops (as far as I know, that is my guess). I can remove the color for the most part in the highlights by creating a luma mask in Capture One and reducing saturation + changing tint in the bright tonal ranges, however this introduces an extra layer of complexity and annoyance to my processing and often will affect highlights in the rest of the image. If I were to use all frames from the bracket, the resulting image would have the bad color issues in it, unless I fix it in every single one. Here is the +0 image with adjusted saturation:



Importantly, this doesn't appear when highlights are blown out completely in the +4 to +5 range. I was shooting in bracketing mode, which I believe I have seen in a thread recently drops down to 12 bit color data due to continuous shooting. I am wondering if this was the culprit, but then it is very strange that it shows in the neutrally metered image very prominently. I also have heard people speculate that Lumix raw files are all lossy compressed by default with no option to turn it off.

My only other worry is that I was shooting several other brackets at small apertures f11-f22 throughout the day that featured the sun in the shots, and I am worried that my sensor was damaged somehow. There is no way that my sensor was exposed to sunlight for more than a minute at a time and at wide angles and very small apertures, however I am concerned after seeing this issue.

I'm concerned because part of my main use for this camera is shooting real estate, and bracketing is a requirement. Regardless of the reason, if these issues show up when bracketing that is a major impediment to my workflow and a huge annoyance that I never experienced on Sony. Has anybody else ever seen this in their images?

 

[PJD]

FWIW, I haven't seen this in my S5iiX RAW files. However, I've seen examples of over-exposed images that display a pink discoloration, including scenes shot using Blackmagic & Lumix cameras. The common denominator might be Sony sensors, but that may not be a factor. (I'd link to a couple of example images on Blackmagic's forum, but linking to another forum is discouraged here.)

I doubt your sensor is damaged.

Some things to check:

• I don't know if it can make a difference, but were you shooting using iA mode? If so, maybe some "auto-magic" effect is causing the issue. If so, test-shoot in manual mode instead.

• Does this happen with different lenses (although I doubt it's a lens issue)? If you were using a filter in front of the lens, does it happen if you shoot similar scenes without a filter?

• You might double-check if your RAW workflow is somehow a factor. Test against different software (perhaps a free trial of another app) to see if it produces a different result.

• Since S5_ bracketed & burst RAW images are 12-bit instead of 14-bit (according to Panasonic tech support), instead of using these features for this type of scene, shoot some tests using the camera locked on a tripod and shoot multiple "bracketed" frames _manually_. IOW, not using the camera's bracketing feature, but adjust exposure manually for each image.

Let us know what you find out. I'm interested to hear others comments on this too.

 

[Pete_W]

I haven't noticed this in any of my images either Tim. It's a bit strange.

Do you have another raw processor you can try to see if there is a difference?

 

[CH_CH]

Did you post something like this in a different forum? I think I remember similar pictures with similar pink-violet aberrations.
If you can, I would return the unit and get a different one, just in case.
I never had something like this. The closest I got was using Lumix lenses on Oly bodies or the opposite, in MfT, as some UV is not filtered- do not ask me which combination. But it looked somewhat different.

 

[CharlesH]

12 bits could be part of the issue. The S9 has 12 bits, and it seems they have taken the 2 bits off the top end, or at least one bit off the top. So you don't have the over exposure recovery you have with a 14 bit camera. If I have a scene with bright areas I usually under expose with the S9 to compensate for this, so I don't saturate the bright areas. As best I can tell all the colors seem to saturate at the same time so that doesn't necessarily explain strange color patterns.

 

[inspiredtimothy]

FWIW, I haven't seen this in my S5iiX RAW files. However, I've seen examples of over-exposed images that display a pink discoloration, including scenes shot using Blackmagic & Lumix cameras. The common denominator might be Sony sensors, but that may not be a factor. (I'd link to a couple of example images on Blackmagic's forum, but linking to another forum is discouraged here.)

I doubt your sensor is damaged.

Some things to check:

• I don't know if it can make a difference, but were you shooting using iA mode? If so, maybe some "auto-magic" effect is causing the issue. If so, test-shoot in manual mode instead.

• Does this happen with different lenses (although I doubt it's a lens issue)? If you were using a filter in front of the lens, does it happen if you shoot similar scenes without a filter?

• You might double-check if your RAW workflow is somehow a factor. Test against different software (perhaps a free trial of another app) to see if it produces a different result.

• Since S5_ bracketed & burst RAW images are 12-bit instead of 14-bit (according to Panasonic tech support), instead of using these features for this type of scene, shoot some tests using the camera locked on a tripod and shoot multiple "bracketed" frames _manually_. IOW, not using the camera's bracketing feature, but adjust exposure manually for each image.

Let us know what you find out. I'm interested to hear others comments on this too.

No auto modes on that I am aware of. idynamic and iauto were switched off to my knowledge.

I am not sure if this appeared with the 14-28 as well. I have definitely seen it on the 28-200, will have to check with other lenses. I'm not sure what I should use for methodology.


This issue does appear when viewing the raw previews in Windows and also in DXO Photolab and Luminar Neo in addition to Capture One. I don't use Lightroom since I don't have an Adobe subscription. I did however, convert to DNG using the free Adobe converter and the issue still presents itself when opening in raw software. It doesn't appear to be an incompatibility with specific software.

I can bracket manually and see if the issue presents still. I am hoping that the resolution to the color errors is something more convenient than inputting everything manually. That would still be a big impediment to my workflow and at that point I may have to consider other options.

One other consideration is that I was in auto shutter mode which I believe usually defaults to EFC except at fast shutter speeds. I can try switching to full manual and see if that resolves it.

One problem here is that I'm not sure how to replicate this except to find a cloud in the sky at roughly the same position and try a variety of settings and then check back at my computer to see if it showed up at all. If that's what I have to do it may be some time before I figure this out

 

[inspiredtimothy]

I haven't noticed this in any of my images either Tim. It's a bit strange.

Do you have another raw processor you can try to see if there is a difference?

No change from what I can tell except in Luminar Neo and DXO Photolab it appears to look worse (probably due to differences in color reproduction).

 

[inspiredtimothy]

Did you post something like this in a different forum? I think I remember similar pictures with similar pink-violet aberrations.
If you can, I would return the unit and get a different one, just in case.
I never had something like this. The closest I got was using Lumix lenses on Oly bodies or the opposite, in MfT, as some UV is not filtered- do not ask me which combination. But it looked somewhat different.

I haven't posted this anywhere else, but if you're able to find other posts where people experience this and share them here, that could be very helpful!

 

[inspiredtimothy]

12 bits could be part of the issue. The S9 has 12 bits, and it seems they have taken the 2 bits off the top end, or at least one bit off the top. So you don't have the over exposure recovery you have with a 14 bit camera. If I have a scene with bright areas I usually under expose with the S9 to compensate for this, so I don't saturate the bright areas. As best I can tell all the colors seem to saturate at the same time so that doesn't necessarily explain strange color patterns.

The strangest part of this is that it seems to be worse when all the highlight information is retained, so it doesn't appear to be related to isolated clipping in the red, blue, or green channels. I was using highlight metering and this was present even at the +0 setting. I believe it was in the underexposed shots as well, but I can double check later when I have time. In the overexposed shots the color issue is gone, other than all information being missing