On Sun, May 15, 2022 at 12:49:51PM -0400, Ronald S. Bultje wrote: > Hi, > > On Fri, May 13, 2022 at 5:22 PM Michael Niedermayer > wrote: > > > On Fri, May 13, 2022 at 11:42:08AM -0400, Leo Izen wrote: > > > This commit moves some of the functionality from avfilter/colorspace > > > into avutil/csp and exposes it as a public API so it can be used by > > > libavcodec and/or libavformat. > > [...] > > > diff --git a/libavutil/csp.h b/libavutil/csp.h > > > new file mode 100644 > > > index 0000000000..1bcde7ddd3 > > > --- /dev/null > > > +++ b/libavutil/csp.h > > > @@ -0,0 +1,49 @@ > > > +/* > > > + * Copyright (c) 2016 Ronald S. Bultje > > > + * This file is part of FFmpeg. > > > + * > > > + * FFmpeg is free software; you can redistribute it and/or > > > + * modify it under the terms of the GNU Lesser General Public > > > + * License as published by the Free Software Foundation; either > > > + * version 2.1 of the License, or (at your option) any later version. > > > + * > > > + * FFmpeg is distributed in the hope that it will be useful, > > > + * but WITHOUT ANY WARRANTY; without even the implied warranty of > > > + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU > > > + * Lesser General Public License for more details. > > > + * > > > + * You should have received a copy of the GNU Lesser General Public > > > + * License along with FFmpeg; if not, write to the Free Software > > > + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA > > 02110-1301 USA > > > + */ > > > + > > > +#ifndef AVUTIL_CSP_H > > > +#define AVUTIL_CSP_H > > > + > > > +#include "libavutil/frame.h" > > > +#include "libavutil/pixfmt.h" > > > + > > > +struct LumaCoefficients { > > > + double cr, cg, cb; > > > +}; > > > + > > > +struct PrimaryCoefficients { > > > + double xr, yr, xg, yg, xb, yb; > > > +}; > > > + > > > +struct WhitepointCoefficients { > > > + double xw, yw; > > > +}; > > > > I think we should avoid floating point so as to ensure reproduceable > > results and simplify regerssion testing > > > > To explain: when I designed this stuff, I chose to keep them in float so > that we can use the literal values from the specs, which are themselves in > floating point. That would not be possible anymore, and would therefore > make it slightly harder to read for a casual observer. > > (Otherwise no opinion, I'm typically used to fixed-point rather than > floating-point myself also.) I looked at Rec. ITU-T H.264 (02/2016) (was the first thing laying around, i assume but did not check that this matches the specs these numbers came from) (randomly picking "whitepoints") the whitepoints for "10" contain some 1/3 values but if i just look at "1" the values are 0.3127, 0.3290 i dont think these can be represented as m*2^E either 3127/10000 = m*2^E 3127 * 2^-E = m * 10000 both sides would be all integers the right side contains 5 as factor, the left doesnt of course i may be missing something thx [...] -- Michael GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB During times of universal deceit, telling the truth becomes a revolutionary act. -- George Orwell