[FFmpeg-trac] #4876(avcodec:open): DNx100 (DNxHD CID 1260) decode produces partially corrupted picture

[FFmpeg]

#4876: DNx100 (DNxHD CID 1260) decode produces partially corrupted picture
------------------------------------+-----------------------------------
             Reporter:  jeremy      |                    Owner:
                 Type:  defect      |                   Status:  open
             Priority:  normal      |                Component:  avcodec
              Version:  git-master  |               Resolution:
             Keywords:  dnxhd       |               Blocked By:
             Blocking:              |  Reproduced by developer:  1
Analyzed by developer:  0           |
------------------------------------+-----------------------------------

Comment (by kurosu):

 Replying to [comment:3 cehoyos]:
 > But it would not be overkill for 10bit iiuc?

 Well, I haven't checked the weighting and shift here and there, but 8bits
 looked really unlikely to need that many bits. 10bits neither, at the
 bitrates intended, but it looks like it would be 'physically' possible.

 That format looks like it was intended to be generic, so it's not
 unrealistic they made things use common denominators. Or the opposite:
 flags (like what now looks like interlaced mb) only existing in some
 profiles etc.

 > +    if (ctx->bit_depth == 8) {
 > +        skip_bits1(&ctx->gb);

 That bit really seems like 'interlaced mb', so it is either always here,
 or not dependent on bitdepth alone, but rather resolution/profile.

 > After some testing, we will only find out if you push;-)

 Well, it was useful as jeremy noticed the non-perfect decoding on
 following frames. See updated patch.

--
Ticket URL: <https://trac.ffmpeg.org/ticket/4876#comment:4>
FFmpeg <https://ffmpeg.org>
FFmpeg issue tracker