We aren't the folks from Adobe. This is a user-to-user forum, we are all just users like yourself. If you can correct that green cast in the shadows your self, you can then save new camera defaults and that adjustment will be made automatically whenever you import images. Just a suggestion that might give you some immediate relief.
Thanks! But how can this be communicated to Adobe since they have messed up this camera profile? There must be a way to contact them?
Thanks for the offer!
Here is the link to the dpreview thread, there is a comparison of the JPGs straight at the beginning (Lightroom versus Capture 1):
In the thread there is a link to the RAW file:
A workaround is to export the RAW into DNG in Capture 1, then open that into Lightroom which then works much better, possibly based on the camera profile embedded now from C1. Another indication that its only the camera profile that is the culprit, not LR in general. That is illustrated with the people shot later in the thread.
Here also the link to Reeves comparison, this page there is a link to the raw files of both A6000 and A7 so you can see the different shadow treatment of both in LR:
By the way there is another thread going into a similar direction on the green cast of A6000 pictures when you use Lightroom and push them by 3 EV.
There a comparison is done to the Nikon Coolpix A that also uses a Sony sensor, with quite perfect rendering - in Lightroom. The last conclusion there is
that shadow noise in green is higher and the black points are incorrectly set, Raw Therapee is supposed to do that better (have to admit I didn't fully understand it).
If anybody understands how to correct this similarily in LR, please tell me :-):
I also took the RAW files from Reeve and run the A/ through ACR, the A6000 through ACR, the ACR through DNG conversion in Capture One and then through ACR, and lastly the A6000 through Capture One into JPG.
What clearly can be seen is a very good color match between A7 rendering by Adobe and the A6000 by Adobe IF THE RAW file was previously converted into DNG by Capture One and this DNG then opened in Adobe.
Since the colors in the bright parts of the pictures match, but only the shadow rendering goes bonkers with the Adobe camera profile for the A6000, I think this is where the problem lies.
I put the different crops on my Flickr pages:
Looking at the two backlit-tree examples, one from an a7 and one from an a6000, it seems one is taken with some sort of HDR or other non-linear processing mode because the shutter speed is 1/80th on one and 1/40th and one looks brighter if you look at the camera-embedded preview JPG but LR renders them with a very similar brightness, as if they were taken with the same 1/80th shutter speed. I would guess LR is trying to simulate some non-linear processing of the camera and that messes with the shadows quite a bit. The 1/80th one is what has the extra green whereas the 1/40th one looks more normal colored in the shadows.
That doesn't mean there isn't some in-camera or in-LR processing difference in some other example images, like the one of the color-checker calibration target where there is a clear green cast to the black areas, it's just there are very few raws to test with, mostly just screenshots.
Those pictures (Reeves) were not taken with HDR. Two different cameras: A7 = Fullframe with aperture f11(?), A6000 = APSC with aperture f8, therefore different shutter speeds. The JPG pictures then were modified by using the "brighten dark areas" slider in Lightroom.
Since the cameras are different, logically different camera profiles are used by LR. The camera profile for the A7 doesn't have this bug, the one for the A6000 does.
I am talking about the two raw files linked in the very last post of page 6 of the Reeves comparison thread, not JPGs or screenshots of unknown manipulation. In LR they both say f/11 and ISO 100 and have a similar tone, but one is shot at twice the shutter speed as the other. This side-by-side is using Adobe Factory Defaults for each image where the profile is Adobe Standard, the toning sliders are all 0, tone-curve linear, and the Sharpening and Color NR are 25:
However, when I look at the camera-embedded JPGs, side-by-side, the 1/80th image on the left is darker than the 1/40th image on the right, as expected given the reported shutter-speed difference:
And it is indeed the darker image that has more shadow problems, also as expected.
When I do manipulate the darker (a6000) raw file in LR, I do see that the shadows are greenish when brightened unlike most other cameras which usually show magenta, which is the combination of the blue and red sensors showing more noise because there are half-as-many photosites as compared to the green, so I am seeing what people are complaining about and would expect shadow noise to be purple not green, I'm more concerned that these two example raws aren't actually comparable due to apparent differences in exposure. It is also odd that the lens listed in LR is different than the lens listed in the filename, so either someone has mistakenly or deliberately renamed the file wrong or someone or something has changed the EXIF parameters that LR is reporting to be the same when they're really not.
The photographer Reeve explained this, he used f8 on the A6000 and f11 on the A7. Since the Canon FD 20 is a fully manual lens, EXIF information was added through the use of Exiftool, where he mistakenly overwrote correct EXIF information for the used aperture, and lens.
Canon FD 20 mm lens used only on A7, Sony SEL1018 used on the A6000.
So the RAW pictures should be about equally bright, since they were equally exposed. The shown JPGs have been pushed in LR differently so are not that important. So its not different exposure that causes the shadow problems, but the Adobe camera profile used for the A6000. Of course the A6000 has somewhat higher noise than the fullframe A7, but as can be seen after running the ARW file through Capture One and THEN processing it in Adobe (see flickr link above), the difference in noise is small and doesn't explain the green cast.
Despite my last name I don't read German, so thanks for the translation/explanation.
That still leaves the difference in brightness in the previews embedded in the raws (my second screenshot) that presumably have nothing to do with LR, or was EXIFtool also used to replace the original camera-embedded previews with something of unknown origin?
From what I am reading here and elsewhere, the noisy shadows of the a6000 are green (and I would add instead of magenta like most cameras), and other raw converters, C1 in particular, processes shadow noise much better than Adobe, and from what I understand tend to desaturate shadows much more, which would reduce the green to more of a neutral gray if the sensor, itself, is the source of the greenishness.
For me to bother an Adobe employee about this, I think I'd want to see a set of comparison raws with both neutral dark and light areas in approximately daylight or studio lights color-balance so the white-balance can be set from the white area and the shift in deep shadow color can be seen in the dark areas. Deep shadows in a backlit scenario are lit slightly with whatever light or reflective material are behind the camera. In the two outdoor scenes, the fence and the tree, this could include reflections from green foliage as well as whatever color the sky is opposite the sun at sunrise and sunset.
At this point it's not clear that it's not the camera sensor that produces the green cast rather than a mishandling of the sensor data by the Adobe raw converter.
following information can be found here:
A simple exiftool command that corrects the blacklevel values removes the massive greenshift.
It would be great if Adobe would correct the camera profile for the A6000.