On 07/03/2012 12:34 PM, Måns Rullgård wrote: > Justin Ruggles <[email protected]> writes: > >> On 07/02/2012 08:21 PM, Mans Rullgard wrote: >>> --- >>> libavcodec/flacdec.c | 10 +++++++-- >>> libavcodec/flacdsp.c | 27 +++++++++++++++++++++++ >>> libavcodec/flacdsp_template.c | 47 >>> ++++++++++++++++++++++++++++------------- >>> 3 files changed, 67 insertions(+), 17 deletions(-) >> >> Currently we're also setting AVCodec.sample_fmts for decoders which can >> output different sample formats. This does not guarantee that a request >> of one of those formats will always work, but it at least gives some >> information to the user as to whether there is a chance to get back a >> certain sample format if requested. > > Thinking about it, we should probably return 32-bit samples if asked > for, even if 16-bit would be wide enough. If the caller asks for 16-bit > when 32-bit required, we should still honour the interleaved/planar > aspect of the request.
Yes, that sounds like a good idea. -Justin _______________________________________________ libav-devel mailing list [email protected] https://lists.libav.org/mailman/listinfo/libav-devel
