| Message ID | a6d7c5e4-13c8-61d5-c945-3828acbd1a75@googlemail.com |
|---|---|
| State | Accepted |
| Headers | show |
On Wed, Oct 19, 2016 at 09:18:51PM +0200, Andreas Cadhalpun wrote: > This is similar to commit c143a9c. > > Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com> > --- > libavformat/aiffdec.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) can aiff work without block_align ? either way, block_duration is from the header reading if its still accurate then using it together with 1 instead of the matching block align is quite likely not correct OTOH if block_duration does not represent the actual content then the duration would only be correct by pure chance Its a bit unfortunate that theres no usecase with an undamaged sample which would have clear correct values one has to work on the assumptation of a use case where the user needs to override the codec and then ask "what is correct to do" that makes this a bit tricky ... [...]
diff --git a/libavformat/aiffdec.c b/libavformat/aiffdec.c index de82787..d96fc1d 100644 --- a/libavformat/aiffdec.c +++ b/libavformat/aiffdec.c @@ -391,7 +391,7 @@ static int aiff_read_packet(AVFormatContext *s, pkt->flags &= ~AV_PKT_FLAG_CORRUPT; /* Only one stream in an AIFF file */ pkt->stream_index = 0; - pkt->duration = (res / st->codecpar->block_align) * aiff->block_duration; + pkt->duration = (res / (st->codecpar->block_align ? st->codecpar->block_align : 1)) * aiff->block_duration ; return 0; }
This is similar to commit c143a9c. Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com> --- libavformat/aiffdec.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)