Recently I have been playing a bit with @patdavid’s Mairi_Troisieme.ORF RAW file, and I have noticed that there is a sort of red cast on it when opened with default settings.
Here is a screenshot of a crop of the greyish background, with exposure and saturation bumped up to better see the histogram:
I would expect this background to be much more neutral, although it might depend on the actual WB settings (I’ve kept the as-shot setting). But then I looked at the RAW black point settings, and noticed that the values are 1 digit lower for RAW channel 0 (red) and RAW channel 3 (green2). Is that really correct? Lowering the red channel and just one of the green channels?
If I set all values to 255, the result is visibly more neutral:
Something went wrong with the Math (typically a division by zero or some unsupported operation on infinity) that resulted in Not a Number (NaN).
Subsequent numerical operations on the NaN just yield another NaN so it keeps on iterating. Clearly a bug.
A crop on the pullover also shows a red cast, which I doubt was there in reality… again, this might come either from the black point settings or the camera WB settings, but still I find the default black point settings at least quite odd.
Are you working from @patdavid’s XMP or with camera WB? In the latter case I would be surprised if the WB wasn’t just whatever the camera set automatically. That’s (one of) the reason(s) to shoot raw: Never touch WB in camera again and stop wasting your time.
Actually, Patrick’s XMP does not change the WB setting, which is left to “camera”. This in turn seems to correspond to “auto WB” because the camera WB does not match any of the other presets…
Anyhow, I simply found strange that the default configuration lowers the black level of the red channel and only one of the green channels, and since the effect is visible by eye I was wondering if there could be a mistake there (for example, the first green channel should be lowered instead of the red one?).
There’s been a recent discussion in the Adobe Lightroom forum about a very similar observation with an Olympus E-M5 II. Some raws show a noticeable color shift in Lightroom and Rawdigger (using Dcraw) relative to Olympus Viewer 3. These photos too have recorded black levels in some of the channels differing by 1.
In that other thread, someone hypothesized that Olympus Viewer 3 may be reading black levels or computing them from some other place in the raw that isn’t known to Lightroom and Rawdigger. For example, Canon’s record per-row black lives, and perhaps the Olympus is doing something similar.
