[issue2570] configure incorrectly enables #define HAVE_AMD3DNOW 1 in config.h
Justin Ruggles justin.rugg...@gmail.com added the comment: this should be fixed now that libswscale uses the libavutil runtime cpu flags. if not, please reopen. -- status: open - closed substatus: open - fixed topic: -build system, swscaler __ Libav issue tracker iss...@roundup.libav.org https://roundup.libav.org/issue2570 __
[issue2639] frame copy of flac audio produces broken file
Justin Ruggles justin.rugg...@gmail.com added the comment: i don't have mpc-hc to duplicate the issue, and the resulting mkv files now seem to be valid. closing the issue. if there are still problems with latest git master please reopen. -- status: open - closed substatus: open - fixed __ Libav issue tracker iss...@roundup.libav.org https://roundup.libav.org/issue2639 __
[issue2678] ffmpeg wav to aac using libfaac wrapper has wrong channel mapping
Justin Ruggles justin.rugg...@gmail.com added the comment: . -- status: new - open substatus: new - open title: ffmpeg wav to aac/ac3 has wrong channels mapping - ffmpeg wav to aac using libfaac wrapper has wrong channel mapping topic: +avcodec type: bug - feature_request __ Libav issue tracker iss...@roundup.libav.org https://roundup.libav.org/issue2678 __
[issue2658] Generating a Tiff from a video throws a seg fault
Justin Ruggles justin.rugg...@gmail.com added the comment:
i have no idea if the attached patch is correct, but it fixes the crash
__
Libav issue tracker iss...@roundup.libav.org
https://roundup.libav.org/issue2658
__diff --git a/libavcodec/tiffenc.c b/libavcodec/tiffenc.c
index 049b28f..a10b833 100644
--- a/libavcodec/tiffenc.c
+++ b/libavcodec/tiffenc.c
@@ -223,7 +223,6 @@ static int encode_frame(AVCodecContext * avctx, unsigned char *buf,
*p = *pict;
p-pict_type = FF_I_TYPE;
p-key_frame = 1;
-avctx-coded_frame= s-picture;
s-compr = TIFF_PACKBITS;
if (avctx-compression_level == 0) {
@@ -442,12 +441,22 @@ fail:
return ret;
}
+static av_cold int encode_init(AVCodecContext *avctx)
+{
+TiffEncoderContext *s = avctx-priv_data;
+
+avcodec_get_frame_defaults((AVFrame*)s-picture);
+avctx-coded_frame = (AVFrame*)s-picture;
+
+return 0;
+}
+
AVCodec ff_tiff_encoder = {
tiff,
AVMEDIA_TYPE_VIDEO,
CODEC_ID_TIFF,
sizeof(TiffEncoderContext),
-NULL,
+encode_init,
encode_frame,
NULL,
NULL,
[issue2674] segmentation fault when accessing flv files that use avc1/h264 video
Justin Ruggles justin.rugg...@gmail.com added the comment: reproduced. backtrace says it all really. -- priority: normal - important status: new - open substatus: new - analyzed topic: +avformat __ Libav issue tracker iss...@roundup.libav.org https://roundup.libav.org/issue2674 __
[issue2674] segmentation fault when accessing flv files that use avc1/h264 video
Justin Ruggles justin.rugg...@gmail.com added the comment: fixed in e3c007b -- status: open - closed substatus: analyzed - fixed __ Libav issue tracker iss...@roundup.libav.org https://roundup.libav.org/issue2674 __
[issue585] Disabling and then re-enabling sound changes audio position
Justin Ruggles justin.rugg...@gmail.com added the comment: please provide a sample so we can either reproduce or close the issue -- nosy: -metamorphosis substatus: new - needs_more_info topic: +ffplay __ Libav issue tracker iss...@roundup.libav.org https://roundup.libav.org/issue585 __
[issue598] New AAC decoder underdocumented?
Justin Ruggles justin.rugg...@gmail.com added the comment: . -- nosy: -jackjansen, superdump priority: normal - minor status: new - open substatus: new - open topic: +documentation type: bug - feature_request __ Libav issue tracker iss...@roundup.libav.org https://roundup.libav.org/issue598 __
[issue598] New AAC decoder underdocumented?
Justin Ruggles justin.rugg...@gmail.com added the comment: weird. nobody showed up on the nosy list, but roundup says i removed jackjansen and superdump. so i'm adding them back... -- nosy: +jackjansen, superdump topic: -documentation __ Libav issue tracker iss...@roundup.libav.org https://roundup.libav.org/issue598 __
[issue2616] 0 channels and channel element 1.0 is not allocated
Justin Ruggles justin.rugg...@gmail.com added the comment: confirmed that the behavior is the same with current libav git. unless this is a regression, it is just a normal bug, not critical. -- priority: critical - normal status: new - open substatus: new - reproduced __ Libav issue tracker iss...@roundup.libav.org https://roundup.libav.org/issue2616 __
[issue2647] AC-3 files with a 16-byte header
Justin Ruggles justin.rugg...@gmail.com added the comment: all except bytes 8 to 11 seem to be the same for every frame. 01 10 00 00 00 00 00 00 XX XX XX XX 00 14 22 06 bytes 8 to 11 could be some sort of timestamp. first two frames are 00 00 00 00 subsequent frames increment by FF C0 any other samples with a different bitrate/channels/sample rate so we can compare? -- status: new - open substatus: new - open type: bug - feature_request FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2647
[issue2647] AC-3 files with a 16-byte header
Justin Ruggles justin.rugg...@gmail.com added the comment: found a forum post mentioning a Soft Encode commandline option: -TimeStamp=Off - use the frame type SMPTE (16 bytes before the normal header). FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2647
[issue2647] AC-3 files with a 16-byte header
Justin Ruggles justin.rugg...@gmail.com added the comment: seems to be some binary decimal format. the increment is not fixed in binary. FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2647
[issue2639] frame copy of flac audio produces broken file
Justin Ruggles justin.rugg...@gmail.com added the comment: All I can confirm is that it does work fine with vlc and ffplay. Maybe there is something about the mkv files that mpc-hc (or whatever mkv demuxer is being used with it) doesn't like for some reason. -- status: new - open substatus: new - open topic: -avcodec FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2639
[issue2639] frame copy of flac audio produces broken file
Justin Ruggles justin.rugg...@gmail.com added the comment: attached: mkvinfo of file generated by ffmpeg File 'ffmkv.txt' not attached - you can download it from https://roundup.ffmpeg.org/file1358. FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2639
[issue2639] frame copy of flac audio produces broken file
Justin Ruggles justin.rugg...@gmail.com added the comment: attached: mkvinfo of file produced by mkvmerge File 'mkvmerge.txt' not attached - you can download it from https://roundup.ffmpeg.org/file1359. FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2639
[issue2640] ff_fmt_convert_init_ppc not defined for powerpc only defined for altivec
Justin Ruggles justin.rugg...@gmail.com added the comment: working on a fix. thanks for the bug report. broken build is priority:important -- assignedto: - jbr nosy: +jbr priority: normal - important status: new - open substatus: new - analyzed topic: +avcodec, build system FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2640
[issue2635] symbol av_get_sample_fmt_string not defined
Justin Ruggles justin.rugg...@gmail.com added the comment: Changing priority back to normal until verified. -- priority: critical - normal topic: -git FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2635
[issue2570] configure incorrectly enables #define HAVE_AMD3DNOW 1 in config.h
Justin Ruggles justin.rugg...@gmail.com added the comment: unlike in libavcodec and libavutil, runtime cpu detection for libswscale is optional. maybe this should be changed to at least default to 'on' instead of 'off' to avoid unwanted surprises like this. -- status: new - open substatus: new - open topic: +build system, swscaler FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2570
[issue2622] Recent change to BBC HD encoder causes un-playable TS file
Justin Ruggles justin.rugg...@gmail.com added the comment: Missing required information for a bug report. Please see http://www.ffmpeg.org/bugreports.html The video plays fine for me with current ffplay. -- substatus: new - needs_more_info FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2622
[issue2617] ffmpeg stalls and finally fails on handbrake-generated mp4
Justin Ruggles justin.rugg...@gmail.com added the comment: confirmed that this issue exists. it's not a demuxer/decoder problem since transcoding to ffv1/nut works just fine with this sample. probably an issue with ffmpeg, libx264 wrapper, or mp4 muxer. -- status: new - open substatus: new - reproduced FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2617
[issue2576] AVCHD lite MTS file conversion loses first seconds
Justin Ruggles justin.rugg...@gmail.com added the comment: changing priority to normal. needs more info and/or reproducing to determine what the cause of the problem is. -- priority: important - normal status: new - open substatus: new - needs_more_info FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2576
[issue421] Support for WMA lossless
Justin Ruggles justin.rugg...@gmail.com added the comment: feature requests are not priority:important except maybe in some extreme circumstances. maybe we can add a topic tag for bounty? -- priority: important - minor FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue421
[issue2618] alac files created with ffmpeg do not play through iTunes sharing
Justin Ruggles justin.rugg...@gmail.com added the comment: Thanks for the bug report. Can you please provide a sample of the Apple-generated file before and after passing it through ffmpeg with -acodec copy? That will allow us to compare the mp4 structures to see where the problem might be. -- status: new - open substatus: new - open topic: +avformat FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2618
[issue2604] ffmpeg crashes when muxing video file with smaller audio file
Justin Ruggles justin.rugg...@gmail.com added the comment: missing necessary information. please see http://www.ffmpeg.org/bugreports.html -- substatus: new - needs_more_info FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2604
[issue2605] WARNING: Library configuration mismatch
Justin Ruggles justin.rugg...@gmail.com added the comment: I don't consider this a bug. It seems like a completely valid warning to me, and pretty obvious since it prints out the configure string for each library that doesn't match the configure string of the ffmpeg program. In the example case the warning can be ignored because the differences in the configuration don't have any effect, but that's not always true. If questions about this warning are common, adding an FAQ entry about it would be fine I think. I'm changing issue type to feature request. -- status: new - open substatus: new - open type: bug - feature_request FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2605
[issue2585] Error when execute ffmpeg
Justin Ruggles justin.rugg...@gmail.com added the comment: Please don't change to important until the issue is verified opened. -- priority: important - normal FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2585
[issue2556] Can't detect AC3 stream properly in WTV
Justin Ruggles justin.rugg...@gmail.com added the comment: fixed by Peter Ross in e4f85b849913794395bb03dfc09546cd41b10882 -- status: open - closed substatus: analyzed - fixed FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2556
[issue2587] AC3 encoder has been changed to take float but accepts S16
Justin Ruggles justin.rugg...@gmail.com added the comment: Many of the audio encoders do not check sample format explicitly. Instead they have a list of supported sample formats. The application is supposed to check AVCodecContext-codec-sample_fmts. -- status: new - closed substatus: new - invalid FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2587
[issue2556] Can't detect AC3 stream properly in WTV
Justin Ruggles justin.rugg...@gmail.com added the comment: Our wtv demuxer isn't handling something correctly. There seem to be 2 streams with the same stream id of 267. The first is AC-3, and it uses a stream2 GUID. The second is MP2, and it uses a stream GUID. The 2nd seems to be overriding the first, which then is possibly filtering out the data based on the index of the MP2 stream? or maybe there is a data2 chuck type we're not detecting? I don't know... but there is definitely AC-3 in the file. -- status: new - open substatus: new - reproduced topic: +avformat FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2556
[issue2556] Can't detect AC3 stream properly in WTV
Justin Ruggles justin.rugg...@gmail.com added the comment: It seems that the mp2 and ac3 are interleaved in the same stream, with each packet having an mp2 frame concatenated with an ac3 frame. Our demuxer skips the ac3 frames. The data for ac3 is using stream id 266, but that is not detected in the stream setup. Both are detected as 267 for some reason... -- substatus: reproduced - analyzed FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2556
[issue2556] Can't detect AC3 stream properly in WTV
Justin Ruggles justin.rugg...@gmail.com added the comment: I mean... they're not interleaved in the same stream. they have 2 separate stream id's (266 267) but the setup is not detecting the AC3 as a separate stream id. they're both detected as 267 and the mp2 info overrides the ac3 info. the ac3 data is skipped due to the data stream id (266) not matching anything from the stream setup. Hope that explains it... now to find the fix... FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2556
[issue2556] Can't detect AC3 stream properly in WTV
Justin Ruggles justin.rugg...@gmail.com added the comment: I have a fix. I'll send it to ffmpeg-devel after a bit more testing. Basically, stream2_guid chunks can initialize new streams as well. Existing demuxer only parses stream2_guid info if the stream already exists but no data has been read for it. FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2556
[issue2556] Can't detect AC3 stream properly in WTV
Justin Ruggles justin.rugg...@gmail.com added the comment: Out of the 3 samples streams (this one + 2 in mphq samples), all of them have extra streams using other stream ids that are skipped because they're initialized with stream2_guid, not stream_guid. But this is the only sample of the 3 that actually sends data chunks for one of the other sid's. It seems pointless to initialize the others and never have them used, especially since they may come before the primary stream(s). Maybe there is a flag somewhere that says whether there will be data sent for this stream? Or maybe it's just setting up multiple TV channels then switching between them by sending data chunks with those sids? I can't think of a good solution. Sometimes there are even data chunks sent with silence. For example, the Law+and+Order-+SVU_Five_2008_08_12_23_36_55.wtv sample has many sid's but data is only sent for 4: subtitle, stereo audio, video, and mono audio. The mono audio is valid mp2, but it's just silence... The silent one is initialized with stream2_guid. But in the case of this AC3 sample, it is also initialized with stream2_guid, data chunks are sent for it, and it doesn't contain silence, but rather a 5.1 version of the primary stereo mp2 audio stream. FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2556
[issue2556] Can't detect AC3 stream properly in WTV
Justin Ruggles justin.rugg...@gmail.com added the comment: Actually the stereo mp2 stream seems to be descriptive audio for the visually impaired, while the ac3 stream is the main program. FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2556
[issue2475] ffmpeg fails assertion on audio files with invalid sample rates
Justin Ruggles justin.rugg...@gmail.com added the comment: patch sent to ffmpeg-devel FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2475
[issue2581] Sound fragments after seeking
New submission from Justin Ruggles justin.rugg...@gmail.com:
copy/paste of attached SoundFragmentsAfterSeeking.txt:
Hi, while trying to use FFmpeg libs as a developer I found the following:
When decoding sound from some compressed audio formats (namely i tested some mp4
files with aac audio codec and some wmv files with wma audio codec).
After seeking the beginning of the first decoded packet is buggy.
Looks like some leftovers from internal buffers - but I am using
'avcodec_flush_buffers'!
That of course leads to nasty sound clicks.
See sourcecode below.
Basically it shows the first few bytes of the first decoded sound packet then
seeks back and shows the first packets decoding again.
If you use it with some wmv or mp4 file with sound you should see see effect
that both outputs are different but should be equal.
Just make sure the sound is not only silence at the first 10 packets.
Or am I missing some other flush...whatever function here?
I am developing under windows with the r26400 32bit developer compilation from
http://ffmpeg.arrozcru.org/autobuilds/ffmpeg/mingw32/dev/shared/
It happens also with other releases and also the older 64 bit version.
// *** Sourcecode from here ***
AVFormatContext *pFormatCtx = NULL;
AVPacketpacket;
AVStream*pSoundStream = NULL;
AVCodec *pSoundCodec = NULL;
unsigned char *pSoundDecodeBuff;
const int SoundDecodeBuffSize = (AVCODEC_MAX_AUDIO_FRAME_SIZE * 3) / 2;
// Decode next sound packet, print some data and return timestamp
int64_t DecodeSoundPacket() {
while (av_read_frame(pFormatCtx, packet) = 0) {
if (packet.stream_index == pSoundStream-index) {
int nBuff = SoundDecodeBuffSize;
int used = avcodec_decode_audio3(pSoundStream-codec,
(int16_t*)pSoundDecodeBuff, nBuff, packet);
if (used != packet.size) printf(Problem while decoding!\n); //
This also indicates multi frame packets
av_free_packet(packet);
break;
}
av_free_packet(packet);
}
printf(Packet at %.2f:, (float)packet.dts * pSoundStream-time_base.num /
pSoundStream-time_base.den);
for (int i = 0; i 20; i++) printf( %.2x, (int)pSoundDecodeBuff[i]);
printf(\n);
return packet.dts;
}
int main(int argc, char* argv[]) {
// Open media file
av_register_all();
if(av_open_input_file(pFormatCtx,
D:\\Projekte\\VideoTest\\VideoTest_1280x720_29p_h264_Audio48k.mp4, NULL, 0,
NULL) != 0) return -1;
if(av_find_stream_info(pFormatCtx) 0) return -1;
// Find the first sound stream
for(int i = 0; i (int)pFormatCtx-nb_streams; i++)
if(pFormatCtx-streams[i]-codec-codec_type == CODEC_TYPE_AUDIO) {
pSoundStream = pFormatCtx-streams[i];
break;
}
if(pSoundStream == NULL) return -1;
// Open the decoder for the sound stream
pSoundCodec = avcodec_find_decoder(pSoundStream-codec-codec_id);
if (pSoundCodec==NULL) return -1;
if (avcodec_open(pSoundStream-codec, pSoundCodec) 0) return -1;
pSoundDecodeBuff = (unsigned char*)av_malloc(SoundDecodeBuffSize);
// Decode first packet and some others
int64_t SeekTo = DecodeSoundPacket();
for (int i = 0; i 10; i++) DecodeSoundPacket();
// Seek back
av_seek_frame(pFormatCtx, pSoundStream-index, SeekTo,
AVSEEK_FLAG_BACKWARD);
avcodec_flush_buffers(pSoundStream-codec);
// Enable this and the bug is gone
//avcodec_close(pSoundStream-codec);
//avcodec_open(pSoundStream-codec, pSoundCodec);
// Decode first packet again
printf(\n);
DecodeSoundPacket();
// No cleanup - I am s dirty
}
--
topic: -avcodec, avfilter, avformat
FFmpeg issue tracker iss...@roundup.ffmpeg.org
https://roundup.ffmpeg.org/issue2581
[issue442] parameter for getting information about files
Justin Ruggles justin.rugg...@gmail.com added the comment: Should this issue be closed now that ffprobe is in FFmpeg? FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue442
[issue2483] Mix Multiple Audio Inputs
Justin Ruggles justin.rugg...@gmail.com added the comment: It's not clear what HoboZero wants to do by that commandline, but the mail referred to is regarding mixing multiple mono streams into a single multichannel stream. That seems like a valid feature request and probably a good candidate for an audio filter. -- priority: normal - wish substatus: needs_more_info - open FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2483
[issue2580] Complete your registration to FFmpeg issue tracker -- key Ps4rBm7HXyP9fz9ADVr0WmBjAE9kM6H9
Justin Ruggles justin.rugg...@gmail.com added the comment: not an issue -- status: new - closed substatus: new - invalid FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2580
[issue1153] ac3dec 5.1 material is 3x softer than 2.0 material
Justin Ruggles justin.rugg...@gmail.com added the comment: changing to feature request. implementation of heavy compression is needed, as well as exporting of dialogue level. -- substatus: needs_changes - open topic: +avcodec type: patch - feature_request FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue1153
[issue670] Audio at end of clip is cutoff with libfaac encoder
Justin Ruggles justin.rugg...@gmail.com added the comment: fixed in git 243f8241dbf4a451e1197661ccd387c519ae3349 -- status: open - closed substatus: reproduced - fixed FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue670
[issue670] Audio at end of clip is cutoff with libfaac encoder
Justin Ruggles justin.rugg...@gmail.com added the comment: file size samples original92160008 23039991 ffmpeg libfaac 92147756 23036928 ffmpeg aacenc 92160044 2304 faac92164140 23041024 What the FFmpeg native AAC encoder does seems to be pretty sensible, and it's closest in length to the original. It appears to have a 1024-sample delay, so the first encode() call outputs nothing, and the extra call due to CODEC_CAP_DELAY encodes the remaining 1024 samples in the buffer. I'm not sure if libfaac works in a similar way. -- status: new - open substatus: new - reproduced title: Audio at end of clip is cutoff - Audio at end of clip is cutoff with libfaac encoder topic: +avcodec FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue670
[issue2562] Common ID3 tags set via -metadata don't show in Windows Explorer
Justin Ruggles justin.rugg...@gmail.com added the comment: Changing this to a feature request. It might be useful to have a per-muxer AVOption for the mp3 muxer to write (or not write) different versions of ID3 tags. -- priority: important - wish status: new - open substatus: new - open topic: +avformat -ffmpeg type: bug - feature_request FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2562
[issue2560] voc regression
Justin Ruggles justin.rugg...@gmail.com added the comment: On 01/17/2011 03:19 PM, ami_stuff wrote: New submission from ami_stuff ami_st...@o2.pl: the latest voc modification broken decoding of http://samples.mplayerhq.hu/voc/pcm_s16_2/nem.voc Author: cehoyos Date: Tue Jan 11 15:08:45 2011 New Revision: 26309 Log: Do not crash for illegal sample size, fixes issue 2502. Patch by Daniel Kang, daniel.d.kang at gmail @@ -113,7 +117,11 @@ voc_get_packet(AVFormatContext *s, AVPac dec-sample_rate = get_le32(pb); dec-bits_per_coded_sample = get_byte(pb); dec-channels = get_byte(pb); -dec-codec_id = ff_codec_get_id(ff_voc_codec_tags, get_le16(pb)); +tmp_codec = ff_codec_get_id(ff_voc_codec_tags, get_byte(pb)); +if (dec-codec_id == CODEC_ID_NONE) +dec-codec_id = tmp_codec; +else if (dec-codec_id != tmp_codec) +av_log(s, AV_LOG_WARNING, Ignoring mid-stream change in audio codec\n); There is the problem. get_le16() was changed to get_byte(). I'll send a patch to ffmpeg-devel shortly. FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2560
[issue2505] ffmpeg crashes on ts files with invalid headers
Justin Ruggles justin.rugg...@gmail.com added the comment: Fixed in 1360f07 -- status: open - closed substatus: reproduced - fixed FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2505
[issue2560] voc regression
Justin Ruggles justin.rugg...@gmail.com added the comment: fixed in 8ca8e82 -- status: open - closed substatus: reproduced - fixed FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2560
[issue2502] ffmpeg crashes for pcm audio with invalid sample_size
Justin Ruggles justin.rugg...@gmail.com added the comment: The heart of the issue seems to be that voc_get_packet() changes the codec_id when reading each packet based on input stream data. If a file is damaged, a random value will most likely make it CODEC_ID_NONE since there are only 8 valid codec_id's for it to choose from. voc_get_packet() should be fixed to check for an invalid codec_id, and the pcm decoder should be fixed to at least not crash when encountering such changes. FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2502
[issue2502] ffmpeg crashes for pcm audio with invalid sample_size
Justin Ruggles justin.rugg...@gmail.com added the comment: change status -- substatus: open - analyzed FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2502
[issue810] FLAC files with large blocksize do not decode with ffmpeg
Justin Ruggles justin.rugg...@gmail.com added the comment: Well, the solution I had in mind back in July isn't the best solution. This will be fixed after the audio decoders use get/release_buffer(). Which is something I have been working on. FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue810
[issue1505] Decode FLAC but get AV_NOPTS_VALUE in pts
Justin Ruggles justin.rugg...@gmail.com added the comment: The example code is wrong. I verified that ffmpeg.c now gets raw FLAC packets with correct sizes and pts since the FLAC parser was added. -- status: open - closed substatus: open - fixed FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue1505
[issue2387] MPEG-4 ALS decoder MCC decoding bug (non-lossless output)
Justin Ruggles justin.rugg...@gmail.com added the comment: Sample moved to samples/A-codecs/lossless/als trying to keep it organized. FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2387
[issue810] FLAC files with large blocksize do not decode with ffmpeg
Justin Ruggles justin.rugg...@gmail.com added the comment: This should be fixable for FLAC once the FLAC parser makes it to SVN. The packet duration will be set by the demuxer. Then ffmpeg.c could be modified to allocate enough memory based on sample format, sample rate, channels, and duration. The current samples_size calculation makes no sense to me anyway... FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue810
[issue2086] libavformat does not recoginize secondary DTS audio in BDAV
Justin Ruggles justin.rugg...@gmail.com added the comment: Carl Eugen Hoyos wrote: Carl Eugen Hoyos ceho...@rainbow.studorg.tuwien.ac.at added the comment: Secondary E-AC-3 stream works fine (transf.ts in samples). Where is transf.ts? It's not listed in allsamples.txt. -- title: libavformat does not recoginize secondary DTS audio in BDAV - libavformat does not recoginize secondary DTS audio inBDAV FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2086
[issue2086] libavformat does not recoginize secondary DTS audio in BDAV
Justin Ruggles justin.rugg...@gmail.com added the comment: Justin Ruggles wrote: Justin Ruggles justin.rugg...@gmail.com added the comment: Carl Eugen Hoyos wrote: Carl Eugen Hoyos ceho...@rainbow.studorg.tuwien.ac.at added the comment: Secondary E-AC-3 stream works fine (transf.ts in samples). Where is transf.ts? It's not listed in allsamples.txt. Oops nevermind. Found it. allsamples.txt just hasn't been updated yet. FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2086
[issue2071] AC3 audio is not detected in MPEG TS stream
Justin Ruggles justin.rugg...@gmail.com added the comment: Baptiste Coudurier wrote: Baptiste Coudurier baptiste.coudur...@gmail.com added the comment: On 7/5/10 1:20 AM, Carl Eugen Hoyos wrote: Carl Eugen Hoyosceho...@rainbow.studorg.tuwien.ac.at added the comment: [mpegts @ 0x11bf470] probing stream 1 Last message repeated 75 times [mpegts @ 0x11bf470] Probe buffer size limit 500 reached [mpegts @ 0x11bf470] probing stream 1 Last message repeated 3 times Input #0, mpegts, from 'mpegts_ac3_not_detected.ts': Duration: 00:00:07.85, start: 51368.546678, bitrate: 14076 kb/s Stream #0.0[0x31], 165, 1/9: Video: mpeg2video, yuv420p, 1280x720 [PAR 1:1 DAR 16:9], 1001/12, 14950 kb/s, 66.49 fps, 59.94 tbr, 90k tbn, 119.88 tbc Stream #0.1[0x34], 85, 1/9: Data: [0][0][0][0] / 0x At least one output file must be specified What happens during the probing ? Stream 1 doesn't get probbed as ac3, is there a conflict between probes (ie multiple codec returning the same probe score) ? The packets apparently do not start exactly on an AC-3 frame boundary, so the best probe score is AVPROBE_SCORE_MAX/4. This happens to be right at the threshold for detecting a codec via CODEC_ID_PROBE. I'm not sure what the best solution would be, raise the score, lower the threshold, or something else... FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2071
[issue2022] Dependent substream eac3 samples are not autodetected as E-AC-3
Justin Ruggles justin.rugg...@gmail.com added the comment: fixed in r24103 -- status: open - closed substatus: open - fixed FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2022
[issue2022] Dependent substream eac3 samples are not autodetected as E-AC-3
Justin Ruggles justin.rugg...@gmail.com added the comment: Carl Eugen Hoyos wrote: New submission from Carl Eugen Hoyos ceho...@rainbow.studorg.tuwien.ac.at: Both dependent substream eac3 samples on incoming (blu-ray_eac3 and Dependent_substream_decoding_sample.m2ts) are auto-detected as ac3, making HDMI remuxing difficult (works fine with -acodec eac3 -i sample). ffmpeg -analyzeduration 10 -i Dependent_substream_decoding_sample.m2ts FFmpeg version SVN-r23646, Copyright (c) 2000-2010 the FFmpeg developers built on Jun 19 2010 13:25:53 with gcc 4.4.4 configuration: --cc=/usr/local/gcc-4.4.4/bin/gcc libavutil 50.19. 0 / 50.19. 0 libavcodec52.76. 0 / 52.76. 0 libavformat 52.68. 0 / 52.68. 0 libavdevice 52. 2. 0 / 52. 2. 0 libavfilter1.20. 0 / 1.20. 0 libswscale 0.11. 0 / 0.11. 0 Input #0, mpegts, from 'Dependent_substream_decoding_sample.m2ts': Duration: 00:00:11.11, start: 599.966644, bitrate: 8245 kb/s Program 1 Stream #0.0[0x1011]: Video: h264, yuv420p, 1920x1080 [PAR 1:1 DAR 16:9], 29.97 fps, 29.97 tbr, 90k tbn, 59.94 tbc Stream #0.1[0x1100]: Audio: ac3, 48000 Hz, 5.1, s16, 640 kb/s At least one output file must be specified -- messages: 10893 priority: normal status: open substatus: open title: Dependent substream eac3 samples are not autodetected as E-AC-3 type: bug I'll look into it, but IIRC the issue is the probe buffer size. The dependent E-AC-3 frame doesn't occur until after the AC-3 frame. Try increasing the size to see if that helps. -- title: Dependent substream eac3 samples are not autodetected as E-AC-3 - Dependent substream eac3 samples are not autodetected as E-AC-3 FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue2022
[issue1921] crash in flac encoding with large frame_size
Justin Ruggles justin.rugg...@gmail.com added the comment: Fixed in r23570. The FLAC encoder still won't encode large frames, but at least it won't crash anymore. -- status: open - closed substatus: reproduced - fixed FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue1921
[issue1921] crash in flac encoding with large frame_size
Justin Ruggles justin.rugg...@gmail.com added the comment: I patched this within the last couple weeks for the ALS encoder that Thilo and I are working on. I just haven't had a chance to submit it for review. FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue1921
[issue1921] crash in flac encoding with large frame_size
Justin Ruggles justin.rugg...@gmail.com added the comment: I'm not sure how to create a patch for an old git commit... so I'll post the URL here. http://github.com/justinruggles/FFmpeg-alsenc/commit/011c17464f89b188c01627b9c177af665d65b159 FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue1921
[issue1709] ac3_decode_frame does not return err on error
Justin Ruggles justin.rugg...@gmail.com added the comment: Joakim Plate wrote: Joakim Plate elu...@ecce.se added the comment: I think what he meant was it returns frame_size even if the supplied amount of data was less than that. Atleast going by the check he added in our codebase. That can't be intended behaviour can it? That can indeed be fixed (and will be shortly). But that's not what this issue is about. He clearly refers to the value of err not being returned when an error occurs. He says nothing about the return value exceeding the input data size. FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue1709
[issue1850] AC3 transcode (eg to WAV) - bad audio
Justin Ruggles justin.rugg...@gmail.com added the comment: So right now the application can know that the number of channels changes because avctx-channels changes. But I guess it wouldn't hurt to log a warning in the decoder when the number of channels changes and the decoder goes from matching request_channels to not matching it. Or just log it either way. Same probably goes for sample rate too. FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue1850
[issue1850] AC3 transcode (eg to WAV) - bad audio
Justin Ruggles justin.rugg...@gmail.com added the comment: So what happens is that when the stream changes to stereo, the decoder will start outputting stereo. The audio seems correct though. But since the output format is wav, it expects everything to be the same channel layout. So when it changes to stereo, it throws off the channel order. So after that point the channels are interpreted as Rs, LFE, L, C, R, Ls instead of L, C, R, Ls, Rs, LFE. The next question is, should FFmpeg do anything about this? When the mid-stream switch is the other way (stereo-5.1-stereo) we handle it by downmixing the 5.1 content to match. But in this case we don't upmix the stereo content to match the 5.1 content. We just change AVCodecContext.channels to 2 and start outputting stereo. FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/issue1850
[issue246] support for CONFIG_MPEGAUDIO_HP in mpegaudioenc.c
Justin Ruggles justin.rugg...@gmail.com added the comment: USE_HIGHPRECISION went away in r16594 in favor of using CONFIG_MPEGAUDIO_HP directly. Part of the issue you mentioned was fixed, in that other files that include mpegaudio.h no longer need an extra definition, only mpegaudioenc.c. -- title: support for USE_HIGPRECISION in mpegaudioenc.c - support for CONFIG_MPEGAUDIO_HP in mpegaudioenc.c FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/roundup/ffmpeg/issue246
[issue491] g726 encoder generates invalid files
Justin Ruggles justin.rugg...@gmail.com added the comment: That is not exactly true. I have been able to generate files that play with the acm codec that are not 32kbps. ffmpeg -i test.wav -acodec g726 -ac 1 -ab 128k -ar 32000 test_g726.wav mplayer -ac g726 test_g726.wav But FFmpeg can generate many bitrate/samplerate combinations that end up with noisy output when played with the acm codec. Some more testing is needed to determine which combinations are actually valid. -- title: g726 encoder only support 32k - g726 encoder generates invalid files topic: +avcodec FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/roundup/ffmpeg/issue491
[issue668] ffmpeg cannot remux ac3 from mkv into mp4
Justin Ruggles justin.rugg...@gmail.com added the comment: The data needs to be run through the AC-3 parser to get the required codec parameters. Patch attached. -- topic: +avformat FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/roundup/ffmpeg/issue668 mkv_ac3_stream_copy.patch Description: Binary data
[issue1720] ffmpeg does not store metadata in converting to ogg, when using vorbis encoder
Justin Ruggles justin.rugg...@gmail.com added the comment: This seems to me to be a feature request to add support for writing of VorbisComment metadata. -- topic: +avformat _ FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/roundup/ffmpeg/issue1720 _
[issue1720] ffmpeg does not store metadata in converting to ogg, when using vorbis encoder
Justin Ruggles justin.rugg...@gmail.com added the comment: change type -- type: bug - feature_request _ FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/roundup/ffmpeg/issue1720 _
[issue1709] ac3_decode_frame does not return err on error
Justin Ruggles justin.rugg...@gmail.com added the comment: This is intended behavior. The only time it actually returns an error is when there is a frame sync loss. Partial or invalid frames will trigger error concealment, which repeats the last known good block until the next valid frame is decoded. -- status: new - closed substatus: new - invalid _ FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/roundup/ffmpeg/issue1709 _
[issue1709] ac3_decode_frame does not return err on error
Justin Ruggles justin.rugg...@gmail.com added the comment: Yes, it would be welcome. Reopening as a feature request. -- priority: normal - minor status: closed - open substatus: invalid - open type: bug - feature_request _ FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/roundup/ffmpeg/issue1709 _
[issue1692] JetAudio infringes the GPL
Justin Ruggles justin.rugg...@gmail.com added the comment: Vitor wrote: New submission from Vitor vitor1...@gmail.com: Webiste at http://www.jetaudio.com, no source code offer. Binary at http://dn2.cowon.com/JetAudioInc/jetAudio/JAD8002_BASIC.exe . Moreover it looks like they hacked the configure script to mask that they've compiled with --enable-gpl --enable-non-free... That makes me wonder if they also use GPL software in the firmware for their PMPs. I bought an A2 for its format support, but it would be a shame if it violates the GPL. On the other hand, it would be wonderful to have access to the source code so I could try to fix some of the bugs myself. _ FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/roundup/ffmpeg/issue1692 _
[issue1595] FLAC decoder fails with large silent chunks.
Justin Ruggles justin.rugg...@gmail.com added the comment: I don't think anything is wrong here. You're getting weird audio output because you're using --sign=unsigned. In unsigned audio, 0 is full negative amplitude. flac -d foo.flac -o foo_flac.wav ffmpeg -i foo.flac foo_lavc.wav cmp foo_flac.wav foo_lavc.wav The results are identical. No issue. -- status: open - closed substatus: needs_more_info - invalid _ FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/roundup/ffmpeg/issue1595 _
[issue1595] FLAC decoder fails with large silent chunks.
Justin Ruggles justin.rugg...@gmail.com added the comment: Deiz wrote: Deiz ffm...@pwnly.com added the comment: I didn't realize --unsigned was what was causing the output issue (As it turns out, that's dmix mixing the 0 dB sample with other output.) but that's not the underlying issue. I have a copyrighted file (which I'm not going to upload) which is prepended by 10 minutes of silence, and then contains roughly a minute of audio. When I play the FLAC with ffplay or mplayer, bad things happen. For example, seeking with ffplay results in Seek to 43% (-247034:-20:-58) of total duration (-577014:-32:-22). Seek errors seem to be common with all files, but if I play a WAV of the same track, it outputs a sane amount: Seek to 38% ( 0:04:14) of total duration ( 0:11:01) Does ffmpeg differentiate between decoding to disk and decoding for use by audio devices? The fact that no FFmpeg-using application is able to play the FLAC tells me something's wrong. Can't play or can't seek? There is a big difference. If it can't play/decode, what errors are you getting? Is the decoded output different from the input? _ FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/roundup/ffmpeg/issue1595 _
[issue1595] FLAC decoder fails with large silent chunks.
Justin Ruggles justin.rugg...@gmail.com added the comment: needs more info. as Carl said, please supply complete output of the ffmpeg command that leads to this failure. -- substatus: - needs_more_info _ FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/roundup/ffmpeg/issue1595 _
[issue1595] FLAC decoder fails with large silent chunks.
Justin Ruggles justin.rugg...@gmail.com added the comment: Well, it's hard to tell anything without a sample. Can you at least attach the foo.ana file output from flac -a foo.flac and the output of metaflac --list foo.flac? _ FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/roundup/ffmpeg/issue1595 _
[issue1595] FLAC decoder fails with large silent chunks.
Justin Ruggles justin.rugg...@gmail.com added the comment: That foo.ana file indicates that your source is all silence. So I recreated the file and could not reproduce any issues with ffmpeg or ffplay from SVN. Update to latest SVN and see if you can reproduce any issues with ffmpeg or ffplay. -- substatus: - open _ FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/roundup/ffmpeg/issue1595 _
[issue1595] FLAC decoder fails with large silent chunks.
Justin Ruggles justin.rugg...@gmail.com added the comment: there is already an open issue about flac seeking. closing this issue. -- status: open - closed substatus: open - invalid _ FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/roundup/ffmpeg/issue1595 _
[issue850] ffmpeg should parse the audio stream as well as the container
Justin Ruggles justin.rugg...@gmail.com added the comment: Setting avctx-channels for each frame decode fixes the issue. That is what the AC-3 decoder does. In this case, ffmpeg still reports it as 5.0 due to the container, but it is changed to 5.1 by the decoder and decodes properly. FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/roundup/ffmpeg/issue850 Index: libavcodec/dca.c === --- libavcodec/dca.c (revision 20544) +++ libavcodec/dca.c (working copy) @@ -1260,7 +1260,7 @@ unset. Ideally during the first probe for channels the crc should be checked and only set avctx-channels when the crc is ok. Right now the decoder could set the channels based on a broken first frame.*/ -if (!avctx-channels) +//if (!avctx-channels) avctx-channels = channels; if(*data_size (s-sample_blocks / 8) * 256 * sizeof(int16_t) * channels)
[issue1511] aac encoding limited to 152kBits
Justin Ruggles justin.rugg...@gmail.com added the comment: It is libfaac. Try the faac commandline encoder and see what you get. faac -b 192 test.wav -o test.aac Freeware Advanced Audio Coder FAAC 1.26.1 (Aug 16 2008) UNSTABLE Average bitrate: 152 kbps Quantization quality: 100 Bandwidth: 16000 Hz Object type: Low Complexity(MPEG-2) + M/S Container format: Transport Stream (ADTS) Encoding test.wav to test.aac frame | bitrate | elapsed/estim | play/CPU | ETA 13314/13314 (100%)| 152.0 | 12.9/12.9 | 23.98x | 0.0 _ FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/roundup/ffmpeg/issue1511 _
[issue1505] Decode FLAC but get AV_NOPTS_VALUE in pts
Justin Ruggles justin.rugg...@gmail.com added the comment: This is the same issue as 1150. A FLAC parser should fix the pts and also allow generic seeking. -- assignedto: - jbr dependson: +FFplay can't seek *.flac file nosy: +jbr status: new - open substatus: new - open superseder: +FFplay can't seek *.flac file topic: +avformat -avcodec _ FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/roundup/ffmpeg/issue1505 _
[issue1290] aac encoder - weird behaviour regarding global_quality
Justin Ruggles justin.rugg...@gmail.com added the comment: I have noticed the same thing. The rate control for aacenc just seems strange. I can't quite figure it out. It should either be intuitive or it should be documented. Also, when bit_rate is high the encoder will hang or crash. I'll try to come up with a good bug report for some crashing/hanging examples. -- status: new - open substatus: new - reproduced _ FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/roundup/ffmpeg/issue1290 _
[issue1153] ac3dec 5.1 material is 3x softer than 2.0 material
Justin Ruggles justin.rugg...@gmail.com added the comment: I couldn't find the actual bulletin, but I did find the Dolby press release[1] that describes it, and it seems to be saying that for DVB, using RF Mode on the receiver will make the loudness of (E)AC-3 closer to that of MP2. The spec has information about implementing Heavy Compression, which is one part of RF Mode. The Dolby Professional Encoding Guidelines[2] give even more information about the RF Mode itself. libavcodec does not use Heavy Compression, which is used in place of Dynamic Range Compression for RF Mode. Its purpose is to limit peaks to prevent overmodulation since the signal is boosted by 11dB. I have no objection to adding this option, which would apply a gain of 11dB and apply heavy compression. Dialogue Normalization would still need to be done outside of libavcodec though, since it is relative to actual output loudness in dBFS LAeq, which libavcodec has no control over. [1] http://www.dolby.com/about/news_events/dtvaudio_update/dau_june2008_vol9no2.html [2] http://www.dolby.com/uploadedFiles/zz-_Shared_Assets/English_PDFs/Professional/46_DDEncodingGuidelines.pdf -- nosy: -jbr _ FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/roundup/ffmpeg/issue1153 _
[issue1480] Attempting to seek FLAC breaks avf
Justin Ruggles justin.rugg...@gmail.com added the comment: duplicate of Issue1150 though this issue focuses more on lavf rather than just ffplay. -- assignedto: - jbr nosy: +jbr status: new - open substatus: new - duplicate topic: +avformat _ FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/roundup/ffmpeg/issue1480 _
[issue1455] MP3 from AIF puts noise to the end of the file
Justin Ruggles justin.rugg...@gmail.com added the comment:
Baptiste Coudurier wrote:
Baptiste Coudurier baptiste.coudur...@gmail.com added the comment:
On 10/11/09 1:02 PM, Justin Ruggles wrote:
Justin Rugglesjustin.rugg...@gmail.com added the comment:
trying to replace the file with new file...
--
nosy: -jbr
_
FFmpeg issue trackeriss...@roundup.ffmpeg.org
https://roundup.ffmpeg.org/roundup/ffmpeg/issue1455
_
aiff_data_size_2.patch
diff --git a/libavformat/aiff.c b/libavformat/aiff.c
index 570e05d..9501a7b 100644
--- a/libavformat/aiff.c
+++ b/libavformat/aiff.c
@@ -46,6 +46,10 @@ static const AVCodecTag codec_aiff_tags[] = {
#define AIFF0
#define AIFF_C_VERSION1 0xA2805140
+typedef struct {
+int64_t data_end;
+} AIFFInputContext;
+
AIFFContext
There is already an AIFFOutputContext in the same file. I chose to name
the new struct AIFFInputContext to distinguish between the two. I guess
I could also merge them, even though the fields used are mutually
exclusive between the muxer and demuxer.
Well, really it would be good to split the file, but that's a different
issue...
_
FFmpeg issue tracker iss...@roundup.ffmpeg.org
https://roundup.ffmpeg.org/roundup/ffmpeg/issue1455
_
[issue1455] MP3 from AIF puts noise to the end of the file
Justin Ruggles justin.rugg...@gmail.com added the comment: fixed in r20219. -- status: open - closed substatus: analyzed - fixed _ FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/roundup/ffmpeg/issue1455 _
[issue1455] MP3 from AIF puts noise to the end of the file
Justin Ruggles justin.rugg...@gmail.com added the comment:
trying to replace the file with new file...
--
nosy: -jbr
_
FFmpeg issue tracker iss...@roundup.ffmpeg.org
https://roundup.ffmpeg.org/roundup/ffmpeg/issue1455
_diff --git a/libavformat/aiff.c b/libavformat/aiff.c
index 570e05d..9501a7b 100644
--- a/libavformat/aiff.c
+++ b/libavformat/aiff.c
@@ -46,6 +46,10 @@ static const AVCodecTag codec_aiff_tags[] = {
#define AIFF0
#define AIFF_C_VERSION1 0xA2805140
+typedef struct {
+int64_t data_end;
+} AIFFInputContext;
+
static enum CodecID aiff_codec_get_id(int bps)
{
if (bps = 8)
@@ -314,6 +318,7 @@ static int aiff_read_header(AVFormatContext *s,
unsigned version = AIFF_C_VERSION1;
ByteIOContext *pb = s-pb;
AVStream * st;
+AIFFInputContext *ctx = s-priv_data;
/* check FORM header */
filesize = get_tag(pb, tag);
@@ -366,6 +371,7 @@ static int aiff_read_header(AVFormatContext *s,
get_meta(s, comment , size);
break;
case MKTAG('S', 'S', 'N', 'D'): /* Sampled sound chunk */
+ctx-data_end = url_ftell(pb) + size;
offset = get_be32(pb); /* Offset of sound data */
get_be32(pb); /* BlockSize... don't care */
offset += url_ftell(pb);/* Compute absolute data offset */
@@ -420,10 +426,18 @@ static int aiff_read_packet(AVFormatContext *s,
AVPacket *pkt)
{
AVStream *st = s-streams[0];
+AIFFInputContext *ctx = s-priv_data;
+int64_t max_size;
int res;
+/* calculate size of remaining data */
+max_size = ctx-data_end - url_ftell(s-pb);
+if (max_size = 0)
+return AVERROR_EOF;
+
/* Now for that packet */
-res = av_get_packet(s-pb, pkt, (MAX_SIZE / st-codec-block_align) * st-codec-block_align);
+max_size = FFMIN(max_size, (MAX_SIZE / st-codec-block_align) * st-codec-block_align);
+res = av_get_packet(s-pb, pkt, max_size);
if (res 0)
return res;
@@ -436,7 +450,7 @@ static int aiff_read_packet(AVFormatContext *s,
AVInputFormat aiff_demuxer = {
aiff,
NULL_IF_CONFIG_SMALL(Audio IFF),
-0,
+sizeof(AIFFInputContext),
aiff_probe,
aiff_read_header,
aiff_read_packet,
[issue589] Incorrect channel routing when decoding multichannel audio
Justin Ruggles justin.rugg...@gmail.com added the comment: The decoders mentioned in this bug report have all been fixed. Closing the issue. -- status: open - closed substatus: reproduced - fixed FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/roundup/ffmpeg/issue589
[issue1455] MP3 from AIF puts noise to the end of the file
Justin Ruggles justin.rugg...@gmail.com added the comment:
The issue is that the SSND chunk does not have to be the last chunk in the AIFF
file, but the AIFF demuxer reads audio data until the end of the file. The
attached patch determines the audio data size so that aiff_read_packet() does
not return data past the end of the SSND chunk.
--
nosy: +jbr
substatus: open - analyzed
_
FFmpeg issue tracker iss...@roundup.ffmpeg.org
https://roundup.ffmpeg.org/roundup/ffmpeg/issue1455
_diff --git a/libavformat/aiff.c b/libavformat/aiff.c
index 570e05d..66088da 100644
--- a/libavformat/aiff.c
+++ b/libavformat/aiff.c
@@ -46,6 +46,12 @@ static const AVCodecTag codec_aiff_tags[] = {
#define AIFF0
#define AIFF_C_VERSION1 0xA2805140
+typedef struct {
+int64_t data_offset;
+int64_t data_size;
+int max_packet_size;
+} AIFFInputContext;
+
static enum CodecID aiff_codec_get_id(int bps)
{
if (bps = 8)
@@ -314,6 +320,7 @@ static int aiff_read_header(AVFormatContext *s,
unsigned version = AIFF_C_VERSION1;
ByteIOContext *pb = s-pb;
AVStream * st;
+AIFFInputContext *ctx = s-priv_data;
/* check FORM header */
filesize = get_tag(pb, tag);
@@ -368,7 +375,9 @@ static int aiff_read_header(AVFormatContext *s,
case MKTAG('S', 'S', 'N', 'D'): /* Sampled sound chunk */
offset = get_be32(pb); /* Offset of sound data */
get_be32(pb); /* BlockSize... don't care */
+ctx-data_size = size - (8 + offset);
offset += url_ftell(pb);/* Compute absolute data offset */
+ctx-data_offset = offset;
if (st-codec-block_align)/* Assume COMM already parsed */
goto got_sound;
if (url_is_streamed(pb)) {
@@ -420,10 +429,23 @@ static int aiff_read_packet(AVFormatContext *s,
AVPacket *pkt)
{
AVStream *st = s-streams[0];
+AIFFInputContext *ctx = s-priv_data;
+int64_t max_size, pos;
int res;
+/* calculate size of remaining data */
+pos = url_ftell(s-pb);
+if (pos ctx-data_offset) {
+url_fseek(s-pb, ctx-data_offset, SEEK_SET);
+pos = ctx-data_offset;
+}
+max_size = ctx-data_size - (pos - ctx-data_offset);
+if (max_size = 0)
+return AVERROR_EOF;
+
/* Now for that packet */
-res = av_get_packet(s-pb, pkt, (MAX_SIZE / st-codec-block_align) * st-codec-block_align);
+max_size = FFMIN(max_size, (MAX_SIZE / st-codec-block_align) * st-codec-block_align);
+res = av_get_packet(s-pb, pkt, max_size);
if (res 0)
return res;
@@ -436,7 +458,7 @@ static int aiff_read_packet(AVFormatContext *s,
AVInputFormat aiff_demuxer = {
aiff,
NULL_IF_CONFIG_SMALL(Audio IFF),
-0,
+sizeof(AIFFInputContext),
aiff_probe,
aiff_read_header,
aiff_read_packet,
[issue1455] AIFF demuxer puts noise to the end of the file
Justin Ruggles justin.rugg...@gmail.com added the comment: change title. update topic list. -- title: MP3 from AIF puts noise to the end of the file - AIFF demuxer puts noise to the end of the file topic: +avformat _ FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/roundup/ffmpeg/issue1455 _
[issue1426] AC3 don't want to decode
Justin Ruggles justin.rugg...@gmail.com added the comment: Applied the patch in r20110. -- status: open - closed substatus: analyzed - fixed _ FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/roundup/ffmpeg/issue1426 _
[issue773] AAC/AC3/EAC3/DTS should support channel_layout field
Justin Ruggles justin.rugg...@gmail.com added the comment: I think this is done now for the decoders listed here. FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/roundup/ffmpeg/issue773
[issue1426] AC3 don't want to decode
Justin Ruggles justin.rugg...@gmail.com added the comment: Benjamin Larsson wrote: Benjamin Larsson ba...@ludd.ltu.se added the comment: Justin Ruggles wrote: Justin Ruggles justin.rugg...@gmail.com added the comment: The file is broken. The AC-3 spec very clearly states that there must be a new rematrixing strategy defined in block 0. There is no default specified. The attached patch is a workaround that allows the file to play. I'm undecided on whether this should be applied or not. Try eac3too and see if it is able to decode the files. If it is I suggest we do that also. eac3to can use either Nero or lavc for AC-3 decoding. It uses lavc by default if you don't have Nero installed. But yes, it would be nice to see if it decodes with Nero. _ FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/roundup/ffmpeg/issue1426 _
[issue786] Support for WMA Voice
Justin Ruggles justin.rugg...@gmail.com added the comment: this is a valid feature request. change status to open/open. -- status: new - open substatus: new - open FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/roundup/ffmpeg/issue786
[issue1325] Can't transcode from 5.1 to vorbis
Justin Ruggles justin.rugg...@gmail.com added the comment:
Patch attached to allow multichannel input to the libvorbis encoder. Assigning
the issue to myself since there is no libvorbis.c maintainer.
--
assignedto: - jbr
topic: +avcodec
_
FFmpeg issue tracker iss...@roundup.ffmpeg.org
https://roundup.ffmpeg.org/roundup/ffmpeg/issue1325
_Index: libavcodec/libvorbis.c
===
--- libavcodec/libvorbis.c (revision 19890)
+++ libavcodec/libvorbis.c (working copy)
@@ -142,19 +142,14 @@
int l;
if(data) {
+int i, ch;
int samples = OGGVORBIS_FRAME_SIZE;
float **buffer ;
buffer = vorbis_analysis_buffer(context-vd, samples) ;
-if(context-vi.channels == 1) {
-for(l = 0 ; l samples ; l++)
-buffer[0][l]=audio[l]/32768.f;
-} else {
-for(l = 0 ; l samples ; l++){
-buffer[0][l]=audio[l*2]/32768.f;
-buffer[1][l]=audio[l*2+1]/32768.f;
-}
-}
+for(i = 0, l = 0; l samples; l++)
+for(ch = 0; ch avccontext-channels; ch++)
+buffer[ch][l]=audio[i++]/32768.f;
vorbis_analysis_wrote(context-vd, samples) ;
} else {
if(!context-eof)
[issue1391] Alias Wavefront - misidentified as adpcm_ea_xas
Justin Ruggles justin.rugg...@gmail.com added the comment: We do not support Alias Wavefront files. With the addition of the low probe score warning message, I would consider this issue fixed. -- status: new - closed substatus: new - fixed _ FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/roundup/ffmpeg/issue1391 _
[issue1384] TwinVQ segfault during ff_sine_window_init
Justin Ruggles justin.rugg...@gmail.com added the comment:
Yes, your 2nd patch fixes the issue for me. Attaching it here for historical
purposes.
_
FFmpeg issue tracker iss...@roundup.ffmpeg.org
https://roundup.ffmpeg.org/roundup/ffmpeg/issue1384
_Index: libavcodec/wma.c
===
--- libavcodec/wma.c (revision 19838)
+++ libavcodec/wma.c (working copy)
@@ -341,8 +341,8 @@
for (i = 0; i s-nb_block_sizes; i++) {
int n;
n = 1 (s-frame_len_bits - i);
-ff_sine_window_init(ff_sine_windows[s-frame_len_bits - i - 7], n);
-s-windows[i] = ff_sine_windows[s-frame_len_bits - i - 7];
+ff_sine_window_init(ff_sine_windows[s-frame_len_bits - i], n);
+s-windows[i] = ff_sine_windows[s-frame_len_bits - i];
}
s-reset_block_lengths = 1;
Index: libavcodec/mdct.c
===
--- libavcodec/mdct.c (revision 19838)
+++ libavcodec/mdct.c (working copy)
@@ -48,13 +48,17 @@
window[i] = sqrt(local_window[i] / sum);
}
+DECLARE_ALIGNED(16, float, ff_sine_32 [ 32]);
+DECLARE_ALIGNED(16, float, ff_sine_64 [ 64]);
DECLARE_ALIGNED(16, float, ff_sine_128 [ 128]);
DECLARE_ALIGNED(16, float, ff_sine_256 [ 256]);
DECLARE_ALIGNED(16, float, ff_sine_512 [ 512]);
DECLARE_ALIGNED(16, float, ff_sine_1024[1024]);
DECLARE_ALIGNED(16, float, ff_sine_2048[2048]);
DECLARE_ALIGNED(16, float, ff_sine_4096[4096]);
-float * const ff_sine_windows[6] = {
+float * const ff_sine_windows[] = {
+NULL, NULL, NULL, NULL, NULL, // unused
+ff_sine_32 , ff_sine_64 ,
ff_sine_128, ff_sine_256, ff_sine_512, ff_sine_1024, ff_sine_2048, ff_sine_4096
};
Index: libavcodec/twinvq.c
===
--- libavcodec/twinvq.c (revision 19838)
+++ libavcodec/twinvq.c (working copy)
@@ -648,7 +648,7 @@
tctx-dsp.vector_fmul_window(out2,
prev_buf + (bsize-wsize)/2,
buf1 + bsize*j,
- ff_sine_windows[av_log2(wsize) - 7],
+ ff_sine_windows[av_log2(wsize)],
0.0,
wsize/2);
out2 += wsize;
@@ -903,9 +903,9 @@
}
-ff_sine_window_init(ff_sine_windows[av_log2(size_m) - 7], size_m );
-ff_sine_window_init(ff_sine_windows[av_log2(size_s/2) - 7], size_s/2);
-ff_sine_window_init(ff_sine_windows[av_log2(mtab-size) - 7], mtab-size);
+ff_sine_window_init(ff_sine_windows[av_log2(size_m)], size_m );
+ff_sine_window_init(ff_sine_windows[av_log2(size_s/2) ], size_s/2);
+ff_sine_window_init(ff_sine_windows[av_log2(mtab-size)], mtab-size);
}
/**
Index: libavcodec/dsputil.h
===
--- libavcodec/dsputil.h (revision 19838)
+++ libavcodec/dsputil.h (working copy)
@@ -760,13 +760,15 @@
* @param n size of half window
*/
void ff_sine_window_init(float *window, int n);
+extern float ff_sine_32 [ 32];
+extern float ff_sine_64 [ 64];
extern float ff_sine_128 [ 128];
extern float ff_sine_256 [ 256];
extern float ff_sine_512 [ 512];
extern float ff_sine_1024[1024];
extern float ff_sine_2048[2048];
extern float ff_sine_4096[4096];
-extern float * const ff_sine_windows[6];
+extern float * const ff_sine_windows[13];
int ff_mdct_init(MDCTContext *s, int nbits, int inverse, double scale);
void ff_imdct_calc_c(MDCTContext *s, FFTSample *output, const FFTSample *input);
Index: libavcodec/wmaprodec.c
===
--- libavcodec/wmaprodec.c (revision 19838)
+++ libavcodec/wmaprodec.c (working copy)
@@ -426,7 +426,7 @@
/** init MDCT windows: simple sinus window */
for (i = 0; i WMAPRO_BLOCK_SIZES; i++) {
const int n = 1 (WMAPRO_BLOCK_MAX_BITS - i);
-const int win_idx = WMAPRO_BLOCK_MAX_BITS - i - 7;
+const int win_idx = WMAPRO_BLOCK_MAX_BITS - i;
ff_sine_window_init(ff_sine_windows[win_idx], n);
s-windows[WMAPRO_BLOCK_SIZES - i - 1] = ff_sine_windows[win_idx];
}
[issue1325] Can't transcode from 5.1 to vorbis
Justin Ruggles justin.rugg...@gmail.com added the comment: It seems the libvorbis wrapper is written to only support mono or stereo, but it doesn't fail otherwise. It needs to be fixed to either support multi-channel input or fail when the input is not mono or stereo. Changing title since this has nothing to do with AC-3, but rather multi-channel input to the libvorbis encoder. -- nosy: +jbr status: new - open substatus: new - reproduced title: Can't transcode from ac3 to vorbis - Can't transcode from 5.1 to vorbis _ FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/roundup/ffmpeg/issue1325 _
[issue1301] support SWF codecs other than MP3
Justin Ruggles justin.rugg...@gmail.com added the comment: add myself to nosy list -- nosy: +jbr _ FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/roundup/ffmpeg/issue1301 _
[issue1301] support SWF codecs other than MP3
Justin Ruggles justin.rugg...@gmail.com added the comment: Do you have SWF samples with Speex? Full Speex support in FLV is not far off I hope, but I don't know how it is used in SWF. _ FFmpeg issue tracker iss...@roundup.ffmpeg.org https://roundup.ffmpeg.org/roundup/ffmpeg/issue1301 _
