[FFmpeg-devel] [PATCH v1 1/1] libavformat/amr.c: remove mode range check

[Paul B Mahol]

On Sat, Sep 4, 2021 at 1:52 PM Sun Zhenliang <hisunzhenliang at outlook.com>
wrote:

>
> >
> > On Sep 4, 2021, at 16:01, Hendrik Leppkes <h.leppkes at gmail.com> wrote:
> >
> > On Sat, Sep 4, 2021 at 9:10 AM sunzhenliang <hisunzhenliang at outlook.com>
> wrote:
> >>
> >> Those comfort noise frames and empty frames should be
> >> considered the correct frame. And amr.c/amr_read_packet()
> >> also takes them as correct frames too.
> >>
> >> Signed-off-by: sunzhenliang <hisunzhenliang at outlook.com>
> >> ---
> >> libavformat/amr.c | 4 ++--
> >> 1 file changed, 2 insertions(+), 2 deletions(-)
> >>
> >> diff --git a/libavformat/amr.c b/libavformat/amr.c
> >> index 836b276fd5..540d8033d8 100644
> >> --- a/libavformat/amr.c
> >> +++ b/libavformat/amr.c
> >> @@ -178,7 +178,7 @@ static int amrnb_probe(const AVProbeData *p)
> >>
> >>     while (i < p->buf_size) {
> >>         mode = b[i] >> 3 & 0x0F;
> >> -        if (mode < 9 && (b[i] & 0x4) == 0x4) {
> >> +        if ((b[i] & 0x4) == 0x4) {
> >>             int last = b[i];
> >>             int size = amrnb_packed_size[mode];
> >>             while (size--) {
> >> @@ -234,7 +234,7 @@ static int amrwb_probe(const AVProbeData *p)
> >>
> >>     while (i < p->buf_size) {
> >>         mode = b[i] >> 3 & 0x0F;
> >> -        if (mode < 10 && (b[i] & 0x4) == 0x4) {
> >> +        if ((b[i] & 0x4) == 0x4) {
> >>             int last = b[i];
> >>             int size = amrwb_packed_size[mode];
> >>             while (size--) {
> >
> > This is just probing, and expecting to find a significant amount of
> > non-empty/non-noise frames seems like something you would want in
> > probing.
> > Does this actually fix detection of any samples?
> >
> > It seems like this has the potential of quite substantially reducing
> > the number of checked bits and thus invalid detections.
> >
> > - Hendrik
> > ____________________________________________
>
> Thanks for reviewing.
>
> Probing is expecting to find “correct” frames, which includes those kinds
> of comfort noise and NO_DATA frames mentioned in 3GPP Specifications.
>
> Besides, those frames are considered as normal frames while reading
> packets. It seems better to take the same standard of frames in both
> reading and probing.
>
>
Yes, but in past IIRC it caused misdetection when probing.


> - sunzhenliang
>
>
>
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel at ffmpeg.org
> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>
> To unsubscribe, visit link above, or email
> ffmpeg-devel-request at ffmpeg.org with subject "unsubscribe".
>