[FFmpeg-devel] [PATCH] swscale/utils: Remove bpc==8 gating init_range_convert

[Neil Birkbeck]

On Thu, Nov 30, 2017 at 9:52 AM, Michael Niedermayer <michael at niedermayer.cc
> wrote:

> > > Perfect, thanks Michael. Let me check those samples out.
>
> there are 2 more in 2939 which change:
> https://trac.ffmpeg.org/ticket/2939
>

It seems the swscale_unscaled code paths do not get reconfigured when
sws_setColorSpaceDetails is called. The unscaled code path also seems to
get called for yuv420p16le->yuv420p, so the patch only affected those
samples when doing some scaling. But the newer results looks consistent
with direct conversion to rgb (and is consistent when I decode the samples
in matlab).


> > >
> > For that sample, I feel like it may be incorrectly tagged as pc/full.
>
> is it stored in the file or taken from:
> ff_generate_avci_extradata()
> maybe theres a bug in the AVCIntra handling
>
> It seems avci100_1080i_extradata may be the one that is signalling full
range for the AVCI100.mov sample. I tested changing the range flag:
-        0x3c, 0x60, 0x20, 0x20, 0x28, 0x00, 0x00, 0x03,
+        0x3c, 0x20, 0x20, 0x20, 0x28, 0x00, 0x00, 0x03,
There is an unused ACLR atom in the mov that also appears to signal full
range (parsing of that atom is skipped for h264)

>
> > > Report generated with:
> > > https://raw.githubusercontent.com/nbirkbeck/ffmpeg-test-
> > > samples/master/color-range/run.sh
>
> Can you turn this into a fate test ?
>

Will do.