Klaus Post
Klaus Post
ok, I will take a look. Thanks for reporting!
I just want to make sure that when you say "quite substantial", is there a bias, or is it just "different"? Isn't the differences just bigger because of response (base)...
Oh yeah. WB scaling is also applied, which is why red differences are so dominant.
Did some stats: ``` C++ r->scaleBlackWhite(); r_ref->scaleBlackWhite(); double delta = 0; int delta_max = 0; int delta_min = 0; for (int y = 0; y < r->dim.y; y++) { for...
ok, without dithering and when correcting the bias, there is a 0.01 of 65535 difference between SSE and non-SSE, with all values being between -2 and +1 when scaling with...
Well, you cannot base it on an absolute value. A single digit difference could mean that the de-mosaic takes a different approact and therefor interpolates missing values in a different...
Only the definitions themselves and some formal way of supporting the quattro image format. I do however think I have to do Sigma decoding into a planar format, or something...
You are welcome to add the hint to new cameras, even if you cannot confirm it, but I don't like changing the default.
Ok, you have convinced me. Unless @LibRaw has any objections, that change can be made.
@pedrocr ok - go ahead! @LibRaw - you should just have said so! Adding a "applyDither" to the RawDecoder will just be a few lines. edit: btw, the dither should...