for13to1

> Could you share which formula and rounding you used for this? Has been a while since I calculated these. Since we are rounding to RGB 8 bit I don't...

> Also do not forget that [AviSynth/AviSynthPlus#259](https://github.com/AviSynth/AviSynthPlus/issues/259) > > > That means that instead of +/-127 (or +/-32767), it must be handled as 128+/-127.5 (or 32768 +/-32767.5) before it gets...

> > The coefficient 224/255 you mentioned is for RGB2YUV, more specifically for full RGB to limited YUV > > Yes, you are right. > > Cb = Clip1_C (...

> So I also took a look and I think the whole conversion here is not really correct. I looked at the standard (https://www.itu.int/rec/R-REC-BT.2020/en) and inverted the formulas in there...

> Maybe this code needs a good rework overall. @ChristianFeldmann The calculation in the original code is OK. As you can see, there is only little difference in the patch...

> ITU-T Series H Supplement 15 (01/2017) says (https://www.itu.int/itu-t/recommendations/rec.aspx?rec=13243) > > what the inverted matrix should be. Remember, the main matrix is the matrix from R'G'B' to Y'Cb'Cr'. That matrix...

> > This PR is about BT.601/BT.709/BT.2020 conversion, and more precisely, is about linear matrix conversion, with no consideration about gamma/EOTF/OETF. > > And that is what that document is...

> > What is wrong with "k_R, k_G, k_B = 0.2627, 0.6780, 0.0593 are derived from BT.2020 primaries and white point"? Could you explain it? > > Nothing is wrong....

> > When talking about BT.601/BT.709/BT.2020, shouldn't we just refer to BT.601/BT.709/BT.2020 > > Where is your math provided in those docs? No, we need to read all related to...

Thanks for replying. I suppose that showing Bayer raw file in gray is OK, because they are actually of one channel 's size, and in the other hand, this is...