[issue2516] Assertion ost-audio_resample failed at ffmpeg.c:901

[Jean Delvare]

Jean Delvare kh...@linux-fr.org added the comment:

Well done mshell. I can confirm that the problem was introduced by r25939.
Stefano, can you please take a look?

--
nosy: +stefano_sa
substatus: needs_more_info - reproduced


FFmpeg issue tracker iss...@roundup.ffmpeg.org
https://roundup.ffmpeg.org/issue2516

[issue2496] ffmpeg mm potential buffer over-read bug

[Peter Ross]

Peter Ross pr...@xvid.org added the comment:

The input buffer is always padded with FF_INPUT_BUFFER_PADDING_SIZE (8) 
bytes, so a segfault will never happen here.


FFmpeg issue tracker iss...@roundup.ffmpeg.org
https://roundup.ffmpeg.org/issue2496

[issue2565] AC3 DTS wav muxer

[Benjamin Larsson]

New submission from Benjamin Larsson ba...@ludd.ltu.se:

DTS and AC3 in wav have a header saying these files have 2 channel 16 bit 
pcm samples. This to make cd recording software able to record DTS and AC3 
cd directly. So a muxer that can write this header instead of a proper 
header is then needed.

--
messages: 13531
priority: normal
status: new
substatus: new
title: AC3 DTS wav muxer
type: feature_request


FFmpeg issue tracker iss...@roundup.ffmpeg.org
https://roundup.ffmpeg.org/issue2565

[issue2566] Synced stream demux

[Benjamin Larsson]

New submission from Benjamin Larsson ba...@ludd.ltu.se:

When using ffmpeg to split streams the streams desync when remuxed as the 
pts/dts is not preserved. I use ffmpeg -i file.tbf -an -vcodec copy 
file_v.raw and then ffmpeg -i file.tbf -vn -acodec copy file_a.raw. An 
option that would take a ms 'x' argument to split the streams when they 
first are found to be that close would be useful to have.

--
messages: 13532
priority: normal
status: new
substatus: new
title: Synced stream demux
type: feature_request


FFmpeg issue tracker iss...@roundup.ffmpeg.org
https://roundup.ffmpeg.org/issue2566

[issue2567] dvdauthor does not accept ffmpeg remuxed vobs

[Benjamin Larsson]

New submission from Benjamin Larsson ba...@ludd.ltu.se:

Files remuxed fail at the end of the dvd image creation with ffmpeg. 
Splitting the streams and using mplex to mux the streams work without any 
problem. But you need to manually sync the streams then and rebuild the 
streams a few times to get it right.

--
messages: 13533
priority: normal
status: new
substatus: new
title: dvdauthor does not accept ffmpeg remuxed vobs
type: bug


FFmpeg issue tracker iss...@roundup.ffmpeg.org
https://roundup.ffmpeg.org/issue2567

[issue2393] avcodec_decode_video craches with H.264 steram

[Ronald S. Bultje]

Ronald S. Bultje rsbul...@gmail.com added the comment:

The attached patch fixes the crash for me. The problem is that on 
resolution change, the decoded NALs are free'ed in free_tables() and 
then reused directly after that (after reinit) during decoding of the 
actual data, this of course crashes. The attached patch fixes that.

There's still some corrupt data at the bottom of the images, is that 
also a bug or was that always there?


FFmpeg issue tracker iss...@roundup.ffmpeg.org
https://roundup.ffmpeg.org/issue2393
diff --git a/libavcodec/h264.c b/libavcodec/h264.c
index 40dc276..2dc8045 100644
--- a/libavcodec/h264.c
+++ b/libavcodec/h264.c
@@ -614,7 +614,7 @@ static void hl_motion(H264Context *h, uint8_t *dest_y, 
uint8_t *dest_cb, uint8_t
 }
 
 
-static void free_tables(H264Context *h){
+static void free_tables(H264Context *h, int free_rbsp){
 int i;
 H264Context *hx;
 av_freep(h-intra4x4_pred_mode);
@@ -637,10 +637,12 @@ static void free_tables(H264Context *h){
 av_freep(hx-top_borders[1]);
 av_freep(hx-top_borders[0]);
 av_freep(hx-s.obmc_scratchpad);
+if (free_rbsp){
 av_freep(hx-rbsp_buffer[1]);
 av_freep(hx-rbsp_buffer[0]);
 hx-rbsp_buffer_size[0] = 0;
 hx-rbsp_buffer_size[1] = 0;
+}
 if (i) av_freep(h-thread_context[i]);
 }
 }
@@ -748,7 +750,7 @@ int ff_h264_alloc_tables(H264Context *h){
 
 return 0;
 fail:
-free_tables(h);
+free_tables(h, 1);
 return -1;
 }
 
@@ -1776,7 +1778,7 @@ static int decode_slice_header(H264Context *h, 
H264Context *h0){
 || av_cmp_q(h-sps.sar, s-avctx-sample_aspect_ratio))) {
 if(h != h0)
 return -1;   // width / height changed during parallelized decoding
-free_tables(h);
+free_tables(h, 0);
 flush_dpb(s-avctx);
 MPV_common_end(s);
 }
@@ -1784,6 +1786,7 @@ static int decode_slice_header(H264Context *h, 
H264Context *h0){
 if(h != h0)
 return -1;  // we cant (re-)initialize context during parallel 
decoding
 
+printf(Reinit\n);
 avcodec_set_dimensions(s-avctx, s-width, s-height);
 s-avctx-sample_aspect_ratio= h-sps.sar;
 av_assert0(s-avctx-sample_aspect_ratio.den);
@@ -3331,7 +3334,7 @@ av_cold void ff_h264_free_context(H264Context *h)
 {
 int i;
 
-free_tables(h); //FIXME cleanup init stuff perhaps
+free_tables(h, 1); //FIXME cleanup init stuff perhaps
 
 for(i = 0; i  MAX_SPS_COUNT; i++)
 av_freep(h-sps_buffers + i);

[issue2393] avcodec_decode_video craches with H.264 steram

[Ronald S. Bultje]

Ronald S. Bultje rsbul...@gmail.com added the comment:

Applied.

Please open a new issue for the bottom screen corruption if appropriate.

--
status: open - closed
substatus: reproduced - fixed


FFmpeg issue tracker iss...@roundup.ffmpeg.org
https://roundup.ffmpeg.org/issue2393

[issue2393] avcodec_decode_video craches with H.264 steram

[Sergey]

Sergey svbyst...@gmail.com added the comment:

Thank you! We will test patch shortly.


FFmpeg issue tracker iss...@roundup.ffmpeg.org
https://roundup.ffmpeg.org/issue2393