[Ffmpeg-devel-irc] ffmpeg-devel.log.20171230

[burek]

[01:06:28 CET] <chance> Does anyone have any idea why ffmpeg audio filters impose an audio skew/drift?  Is there a way around it?
[01:07:38 CET] <Compn> chance : you would be better off creating a trac ticket, with specific command lines, example input and output samples, and 3rd party verification of said skew/drift
[01:07:48 CET] <Compn> then we can reproduce it , and possibly fix it :)
[01:08:05 CET] <Compn> i'm not sure if our audio filter guys are on irc right at the moment
[01:08:15 CET] <Compn> so documenting the bug would be important first of all 
[01:08:18 CET] <chance> @Compn - I integrate with FFmpeg at the library level
[01:08:23 CET] <Compn> http://trac.ffmpeg.org
[01:09:00 CET] <Compn> well this channel is still for development of ffmpeg itself and not 3rd party library integrations
[01:09:17 CET] <Compn> can you reproduce with ffmpeg binary or no ?
[01:09:49 CET] <Compn> its either a bug with us, or a bug with your code, which is why these questions needs t obe asked
[01:10:18 CET] <chance> @Compn good question - I will attempt to reproduce it on the command line
[01:10:50 CET] <Compn> because then we have to debug your stuff.... and not going to do tat heh
[01:10:53 CET] Action: Compn afk
[01:11:08 CET] <chance> Probably not my stuff
[01:11:09 CET] <chance> ;)
[01:13:15 CET] <tmm1> well filters certainly do not impose audio skew/drift. plenty of people use ffmpeg and libavfilter and don't experience that problem
[01:14:58 CET] <chance> Yeah - I see the same skew at the command linke
[01:15:02 CET] <chance> command line
[01:16:10 CET] <chance> Im actually wondering if it has to do the the aac encoder in ffmpeg
[01:16:32 CET] <chance> I recently switched to using this (from faac).  Noticing not as good compression efficiency also
[01:17:06 CET] <BtbN> compared to faac? faac is terrible
[01:19:23 CET] <kierank> probably an issue at source
[01:19:40 CET] <kierank> bad timestamps, edit list, the usual tropes
[01:20:05 CET] <chance> all my audio samples are larger with the same encode settings using the ffmpeg aac encoder
[01:20:16 CET] <chance> maybe the audio quality is arguably better
[01:20:41 CET] <BtbN> well, if the quality is better, I wouldn't complain about the size
[01:20:51 CET] <BtbN> compare the quality at cbr to have a fair comparison at that
[01:21:09 CET] <chance> @kierank I reproduced the issue at the command line
[01:21:25 CET] <BtbN> open a ticket on track then
[01:22:15 CET] <chance> ffmpeg -i <some aac mpeg.ts> -af "volume=volume=0.9" -acodec aac -strict -2 ffmpeg_output.ts
[01:22:18 CET] <atomnuker> chance: the native encoder uses different vbr settings
[01:22:24 CET] <atomnuker> *range
[01:22:36 CET] <BtbN> If you need -strict, your ffmpeg version must be old.
[01:22:44 CET] <chance> hang on
[01:22:48 CET] <chance> I will try it with 3.3
[01:22:55 CET] <BtbN> try master
[01:23:03 CET] <BtbN> anything else won't be a valid ticket anyway
[01:26:35 CET] <chance> hmm.. same audio drift in 3.3.  I will try master
[01:26:40 CET] <chance> thanks for the help guys
[01:26:58 CET] <chance> I dont want to make an issue of this unless it is a real issue
[01:28:24 CET] <BtbN> how do you even get audio drift with just audio?
[01:28:32 CET] <BtbN> Or is there video in that file as well?
[01:29:22 CET] <chance> @BtbN : I take a small chunk of audio, normalize it whatever, then extract it as PCM and import it in audacity and compare it to a raw dump of the source
[01:29:48 CET] <kierank> er, that totally ignores timestamping
[01:29:55 CET] <BtbN> and what do you define as "drift"?
[01:31:04 CET] <chance> What I call "drift" or "skew" is import 'X' number of samples the attack and decay should be in phase
[01:31:44 CET] <chance> in this case the ffmpeg filtered output is always shifted forward
[01:36:12 CET] <atomnuker> chance: ffmpeg.c does not use the preskip value for anything
[01:36:27 CET] <atomnuker> so everything outputted as raw pcm is going to be 1024 samples longer in the case of aac
[01:37:09 CET] <atomnuker> patches welcome
[01:37:17 CET] <kierank> atomnuker: i doubt many decoders on phones and whatever do preskip properly
[01:37:38 CET] <atomnuker> yeah, they're not ideal for gapless playback either
[01:37:51 CET] <chance> @atomnuker - thank you!
[01:38:17 CET] <atomnuker> it'll be longer at the start, mind you, not the end
[01:38:34 CET] <kierank> ~20ms is by my calculation the delay you have 6.8m away from a tv screen so getting close to margin of error territory
[01:38:40 CET] <atomnuker> the end is properly trimmed according to the number of samples or side data
[01:38:45 CET] <chance> afk
[01:39:17 CET] <chance> ill keep you posted on my findings after building master etc
[02:15:45 CET] <Luigi12> Hi, I was wondering if it would be possible to get one last 2.4.x update to fix outstanding security issues, so I could do one last security update for Mageia 5 before the EOL.  Thanks.
[06:06:09 CET] <philipl> argh
[08:21:36 CET] <Tzimmo> Compn: No, a single vob file dumped from DVD source using mplayer -dumpstream -dumpfile out.vob
[08:23:28 CET] <Tzimmo> Compn: I have automated these steps to make these from DVDs because I have multiple TVs unfortunately supporting different formats and even more unfortunately having different bugs making some combinations break on some of them. In this case, unless the subtitle is text-based, in some cases I'm forced to burn it into the video to make it work on some of the TVs.
[11:25:53 CET] <durandal_1707> atomnuker: if you do not write atrac9 decoder in next 24h i will
[11:56:44 CET] <atomnuker> durandal_1707: that would just duplicate work
[12:20:39 CET] <mistym> Is there a way to specify contexts in which a parser should be used?
[12:21:14 CET] <mistym> I'm updating a patch which uses AVCodecParser to extract ATRAC3+ audio from a particular video format. After looking at solo .at3 files, it seems like the parser is inappropriate for standalone ATRAC3+ files.
[12:31:17 CET] <rcombs> mistym: need_parsing on the AVStream
[12:31:21 CET] <rcombs> set by the demuxer
[12:34:20 CET] <mistym> rcombs: Thanks! I'll take a look at updating RIFF/WAV to ignore it for ATRAC3+ then...
[12:34:28 CET] <mistym> s/ignore it/set it to false/
[12:44:42 CET] <atomnuker> durandal_1707: go and write a multiband compressior filter to keep yoursefl busy, someone was requesting that
[12:46:14 CET] <durandal_1707> atomnuker: -af mcompand
[12:58:17 CET] <atomnuker> wow, we have that already?
[13:34:19 CET] <c3-Win> atomnuker: Didn't you say "good"? ;-)
[13:34:32 CET] <c3-Win> (and yes I'm just joking)
[14:09:10 CET] <durandal_1707> c3-Win: good what?
[15:11:56 CET] <Compn> Tzimmo : why not use mencoder -vobsubout ?
[15:12:04 CET] <Compn> or mplayer -vobsubout whichever it requires
[15:12:32 CET] <Compn> i mean as long as you are using mplayer... :)
[15:14:38 CET] <durandal_1707>  /ignore Compn
[15:35:13 CET] <Tzimmo> Compn: What do I do with that vobsub then? Multiplex it back to a single vob file using ffmpeg?
[15:36:53 CET] <Tzimmo> I could try that out. I was kind of hoping to use ffmpeg for conversions but just use mplayer to dump dvd://N to a vob file. I don't know if I could split DVD titles easily using ffmpeg...?
[15:37:44 CET] <Tzimmo> TV needs a single file to show video, audio and subtitles properly. Yes, it could use an external textbased subtitle as well in some cases.
[15:38:52 CET] <Tzimmo> vob/mpg, unlike mkv, mp4, avi or m2ts, is supported by all my TVs
[15:39:20 CET] <Tzimmo> That's why I prefer those for SD content
[15:47:03 CET] <Compn> oh wow never seen a tv support vobsub
[15:47:15 CET] <Compn> it must be based on ffmpeg/mplayer :D
[16:16:59 CET] <durandal_1707> is new utvideo works only with qt?
[17:13:54 CET] <wm4> I suspect the recent EOF change makes lavf eat 100% CPU if a TCP connection dies, can anyone confirm?
[17:16:02 CET] <wm4> ah yes, tcp_read() returns 0 on EOF
[17:16:04 CET] <wm4> fucking well done
[17:25:34 CET] <BtbN> so, make it return EOF?
[17:42:12 CET] <Chloe> how can I move ff_register_rtp_dynamic_payload_handlers() (and rdt one) out of allformats.c?
[17:42:32 CET] <Chloe> oh, I can just remove linked lists therer
[17:42:56 CET] <Chloe> can you have rtp without rdt?
[17:45:05 CET] <wm4> BtbN: you bet someone finds a reason why that's wrong
[17:48:19 CET] <Luigi12> michaelni: Hi, I was wondering if it would be possible to get one last 2.4.x update to fix outstanding security issues, so I could do one last security update for Mageia 5 before the EOL. Thanks. (sorry for asking so late)
[18:01:01 CET] <michaelni> Luigi12, ill see what i can do
[18:01:09 CET] <Luigi12> thanks!
[18:06:02 CET] <cone-427> ffmpeg 03Rostislav Pehlivanov 07master:51027d0b8b28: opus: merge encoder and decoder bitallocation functions into one
[18:27:50 CET] <BtbN> wm4, what does recv do if there is actually no data?
[18:27:57 CET] <BtbN> Just block or return EAGAIN in non-block?
[18:29:49 CET] <wm4> should error and set EAGAIN
[18:30:33 CET] <wm4> why was the UDP thing not just changed to return EAGAIN instead of changing the semantics and breaking a lot of shit
[18:31:26 CET] <durandal_1707> you always have nicolas to veto patches
[18:34:52 CET] <BtbN> Because AGAIN is not EOF
[18:37:39 CET] <wm4> the thing with UDP is that there are 0 sized packets, so I'd have said they should just be ignored or so
[18:37:54 CET] <wm4> like if(ret==0)return AVERROR(EAGAIN);
[18:39:02 CET] <BtbN> Can't it return actual EAGAIN in non-blocking mode?
[18:39:14 CET] <wm4> sure
[18:39:33 CET] <BtbN> So if some weird ass codec relies on zero-sized packages for whatever messed up reason, you'd have a problem
[18:41:50 CET] <wm4> is there one?
[18:42:21 CET] <nevcairiel> zero-sized packets would run against barries all over the place, i doubt anyone would be silly enough
[18:42:28 CET] <nevcairiel> barriers*
[18:42:39 CET] <BtbN> Wasn't that the whole reason this whole effort was started? Because some streaming protocol used them?
[18:45:31 CET] <wm4> BtbN: I thought it was because lavf reacted with EOF to 0 sized UDP packets
[18:45:41 CET] <wm4> but discarding the packets would have been fine
[18:46:56 CET] <BtbN> Yeah, because some protocol commonly sends them
[18:47:07 CET] <BtbN> Or some broken but common implementation of it or something
[18:57:36 CET] <Chloe> the most linked lists i remove, the more i find
[18:58:39 CET] <Chloe> more*
[18:59:29 CET] <cone-427> ffmpeg 03Aman Gupta 07master:c0b08ef94f03: avformat/http: return EINVAL if ff_http_do_new_request is called with non-http URLContext
[19:06:22 CET] <wm4> Chloe: like which?
[19:06:38 CET] <Chloe> RTPDynamicProtocolHandler
[19:06:56 CET] <wm4> ew
[19:07:21 CET] <Chloe> that needs to be removed so lavf can be free'd from its init func
[19:07:35 CET] <wm4> tmm1: didn't I NACK this patch
[19:11:05 CET] <tmm1> wm4: i didnt commit the crypto one
[19:25:28 CET] <wm4> tmm1: now you might as well push the proper fix
[19:56:43 CET] <TD-Linux> wm4, theora used to have 0 sized packet = show existing frame, but I think no real file used it because ffmpeg couldn't handle it
[19:59:58 CET] <TD-Linux> also 0 sized packets are sometimes used to say "there was supposed to be a packet to be here but it was lost", maybe ffmpeg has a special way to signal that
[20:00:55 CET] <Compn> avi had that feature too i think
[20:00:55 CET] <kierank> i have seen such packets sent to signal 7 NULL TS packets that didn't want to be sent over the link
[20:00:58 CET] <Compn> dup frame marker
[20:01:14 CET] <TD-Linux> you could probably use AV_PKT_FLAG_CORRUPT for that too but it's not documented as such
[20:01:15 CET] <kierank> actually not 0
[20:01:20 CET] <kierank> just 0 rtp payload
[20:02:03 CET] <TD-Linux> (you can also use rtp sequence numbers to infer such a packet)
[20:02:05 CET] <DHE> iirc AVI back in the day used the 0 byte payload hack as a means of doing variable FPS
[20:09:25 CET] <kierank> durandal_1707: what new interesting codecs are left?
[20:26:28 CET] <durandal_1707> kierank: Dolby Vision or whatever its called, and xvc
[20:26:44 CET] <kierank> xvc lol
[20:27:24 CET] <durandal_1707> and atrac9, lazy sloooow atomnuker 
[20:32:02 CET] <kierank> michaelni: can you look at my patch please
[20:32:06 CET] <kierank> mpeg4video one
[20:33:00 CET] <Compn> kierank : waiting for vivo to be finished... 
[20:33:02 CET] <Compn> ehe
[20:33:19 CET] <Compn> still lots of binary codecs in mplayer that arent RE'd
[20:35:07 CET] <durandal_1707> dead codecs and formats nobody cares about except mencoder lusers
[20:36:36 CET] <Compn> i was never able to find a vdo sample :(
[20:36:53 CET] <Compn> the other vdo , not vdow/vdo3
[20:37:19 CET] <Compn> durandal_1707 : what about bayer redcode stuff ?
[20:37:21 CET] <durandal_1707> good, it should not exist at first place
[20:38:24 CET] <durandal_1707> Compn, kierank yes red stuff or generally whatever pro users use
[20:39:26 CET] <Compn> kierank : is there any pro /broadcast / film codecs that need to be added to ffmpeg ?
[20:39:33 CET] <Compn> s/is/are
[21:03:11 CET] <cone-427> ffmpeg 03Michael Niedermayer 07master:903be5e4f662: avcodec/exr: Check buf_size more completely
[21:32:39 CET] <atomnuker> durandal_1707: got a link to the windows atrac9 decoder/encoders?
[21:42:51 CET] <atomnuker> Chloe: ping
[21:52:10 CET] <Chloe> atomnuker: pong
[21:52:31 CET] <atomnuker> planning to submit a new version of the lavc patch?
[21:52:49 CET] <Chloe> would you prefer I submit it before I finish the lavf/lavd one?
[21:53:13 CET] <Chloe> cause I have a new version (though not entirely sure it's complete) of the lavc one
[21:56:16 CET] <Chloe> atomnuker: ^
[21:58:48 CET] <atomnuker> oh, ok
[21:58:57 CET] <atomnuker> if you want to submit it anyway
[22:51:55 CET] <atomnuker> why is ff_codec_wav_tags populated with 2-byte tags
[22:52:07 CET] <atomnuker> considering the compare function uses 4 bytes of the rag
[22:52:10 CET] <atomnuker> *tag
[22:53:34 CET] <nevcairiel> because riff audio tags are TwoCC
[22:54:24 CET] <atomnuker> then how come if I put a 4 byte tag it works but if I cut it down to 2 bytes it doesn't?
[22:57:25 CET] <atomnuker> hm, I guess its only in the case of when a GUID is needed
[22:59:22 CET] <atomnuker> why is 0xFFFE special?
[22:59:42 CET] <nevcairiel> because its defined to be
[23:00:25 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:b33d3021954c: avcodec/jpeglsdec: Check ilv for being a supported value
[23:00:26 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:3d6ffa2bb50a: avcodec/jpeglsdec: Check for end of bitstream in ls_decode_line()
[23:00:27 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:18fbf2622cd5: avcodec/aacdec_fixed: Fix integer overflow in predict()
[23:00:28 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:02612c3e3eb5: avcodec/aacdec_fixed: Fix integer overflow in apply_dependent_coupling_fixed()
[23:00:29 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:faa84a0c0667: avcodec/xan: Improve overlapping check
[23:00:30 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:6fb7e324fee1: avcodec/h264idct_template: Fix integer overflows in ff_h264_idct8_add()
[23:00:31 CET] <cone-427> ffmpeg 03Luca Barbato 07release/3.3:912448efc110: avformat: Free the internal codec context at the end
[23:00:32 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:4a47195d2a88: avcodec/exr: fix undefined shift in pxr24_uncompress()
[23:00:33 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:ffa2d60ac5af: avcodec/xan: Check for bitstream end in xan_huffman_decode()
[23:00:34 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:d3264c496a0e: avcodec/h264idct_template: Fix integer overflows in ff_h264_idct8_add()
[23:00:35 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:56a56c0cb564: avutil/softfloat: Add FLOAT_MIN
[23:00:36 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:67208cf992ef: avcodec/aacsbr_fixed: Fix division by zero in sbr_gain_calc()
[23:00:37 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:19fb467fcbbf: avcodec/sbrdsp_fixed: Fix integer overflow in shift in sbr_hf_g_filt_c()
[23:00:38 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:286e3bf17429: avcodec/cngdec: Fix integer clipping
[23:00:39 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:c8027878d024: avcodec/snowdec: Fix integer overflow in header parsing
[23:00:40 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:c1d31ccfac48: avcodec/mdct_*: Fix integer overflow in addition in RESCALE()
[23:00:41 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:8be48f1c9a4e: avcodec/aacdec_fixed: Fix undefined shift
[23:00:42 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:b3067f95c980: avcodec/aacpsdsp_template: Fix integer overflows in ps_decorrelate_c()
[23:00:43 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:96fe37a3390a: avcodec/x86/mpegvideodsp: Fix signedness bug in need_emu
[23:00:44 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:aac7ca7a36da: avcodec/h264dec: Fix potential array overread
[23:00:45 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:066c65737682: avcodec/vc2enc: Clear coef_buf on allocation
[23:00:46 CET] <cone-427> ffmpeg 03Fredrik Hubinette 07release/3.3:74104d2dc05d: avformat/mov: Check size of STSC allocation
[23:00:47 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:488c2e8487e5: avcodec/snowdec: Check intra block dc differences.
[23:00:48 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:01439fe1e139: avcodec/snowdec: Check for remaining bitstream in decode_blocks()
[23:00:49 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:4a412dc6ad19: avcodec/wmv2dec: Check end of bitstream in parse_mb_skip() and ff_wmv2_decode_mb()
[23:00:50 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:78a0356fae83: avcodec/dirac_dwt: Fix integer overflow in COMPOSE_DD137iL0()
[23:00:51 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:1c9af4d7a888: avcodec/zmbv: Check that the buffer is large enough for mvec
[23:00:52 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:f4e25620a1fc: avcodec/mlpdsp: Fix undefined shift ff_mlp_pack_output()
[23:00:53 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:fa29141e34c9: avcodec/hevcdsp_template: Fix invalid shift in put_hevc_epel_bi_w_v()
[23:00:54 CET] <cone-427> ffmpeg 03Jacob Trimble 07release/3.3:a0eccf673cda: avformat/mov: Propagate errors in mov_switch_root.
[23:00:55 CET] <cone-427> ffmpeg 03Dale Curtis 07release/3.3:362967fec6a4: Fix leak of frame_duration_buffer in mov_fix_index().
[23:00:56 CET] <cone-427> ffmpeg 03Dale Curtis 07release/3.3:edd0cd21f41e: Use ff_thread_once for fixed, float table init.
[23:00:57 CET] <cone-427> ffmpeg 03Dale Curtis 07release/3.3:11a940adbcab: Fix undefined shift on assumed 8-bit input.
[23:00:58 CET] <cone-427> ffmpeg 03Dale Curtis 07release/3.3:2de4eb6fec18: Close ogg stream upon error when using AV_EF_EXPLODE.
[23:00:59 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:70dc266342ee: avcodec/mpeg4videodec: Check also for negative versions in the validity check
[23:01:00 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:7bc064d46165: avcodec/dirac_dwt: Fix integer overflow in COMPOSE_FIDELITYi*
[23:01:01 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:d5bdcd8a2702: avcodec/kgv1dec: Check that there is enough input for maximum RLE compression
[23:01:02 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:cead6c94c502: avcodec/mlpdsp: Fix signed integer overflow, 2nd try
[23:01:03 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:c8bbddf057e6: avcodec/hevcdsp_template: Fix undefined shift in put_hevc_epel_bi_w_h()
[23:01:04 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:79ec6381151c: avcodec/j2kenc: Fix out of array access in encode_cblk()
[23:01:05 CET] <cone-427> ffmpeg 03Dale Curtis 07release/3.3:85ea121684a7: avformat/utils: Prevent undefined shift with wrap_bits > 64.
[23:01:06 CET] <cone-427> ffmpeg 03Dale Curtis 07release/3.3:9bc2f44c27a3: avcodec/vorbis: 1 << 31 > int32_t::max(), so use 1u << 31 instead.
[23:01:07 CET] <cone-427> ffmpeg 03Dale Curtis 07release/3.3:272a9687a73c: Don't manipulate duration when it's AV_NOPTS_VALUE.
[23:01:08 CET] <cone-427> ffmpeg 03Dale Curtis 07release/3.3:39db2f95145f: avcodec/vorbis: Fix another 1 << 31 > int32_t::max() with 1u.
[23:01:09 CET] <cone-427> ffmpeg 03Nikolas Bowe 07release/3.3:01ab4117dc03: avcodec/extract_extradata_bsf: Fix leak discovered via fuzzing
[23:01:10 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:454a2405ce80: avcodec/dirac_dwt: Fix integer overflows in COMPOSE_DAUB97*
[23:01:11 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:054188db1087: avcodec/diracdsp: Fix integer overflow in PUT_SIGNED_RECT_CLAMPED()
[23:01:12 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:3d297038a9ab: avcodec/amrwbdec: Fix division by 0 in voice_factor()
[23:01:13 CET] <cone-427> ffmpeg 03Jun Zhao 07release/3.3:603845225cb3: avfilter/formats: fix wrong function name in error message
[23:01:14 CET] <cone-427> ffmpeg 03Kelly Ledford 07release/3.3:b7c9f27ad6e8: libavfilter/af_dcshift.c: Fixed repeated spelling error
[23:01:15 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:badca11741ea: avcodec/hevcdsp_template: Fix undefined shift in put_hevc_qpel_bi_w_hv()
[23:01:16 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:bdba0f6786d5: avcodec/hevc_sei: Fix integer overflows in decode_nal_sei_message()
[23:01:17 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:7a5c73896394: tests/audiomatch: Add missing return code at the end of main()
[23:01:18 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:41a706b9125c: avcodec/hevc_cabac: Fix integer overflow in ff_hevc_cu_qp_delta_abs()
[23:01:19 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:1d9830cba30d: avcodec/dirac_dwt: Fix integer overflow in COMPOSE_DD97iH0() and COMPOSE_DD137iL0()
[23:01:20 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:b66e3e321f64: avcodec/hevcdsp_template.c: Fix undefined shift in FUNC(dequant)
[23:01:21 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:0da741ba6b30: avcodec/flacdec: avoid undefined shift
[23:01:22 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:2cde8dc055c0: avcodec/hevcdsp_template: Fix Invalid shifts in put_hevc_qpel_bi_w_h() and put_hevc_qpel_bi_w_w()
[23:01:23 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:1bc06771d83e: avcodec/flacdec: Fix overflow in multiplication in decode_subframe_fixed()
[23:01:24 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:f2b83f4aba2b: avcodec/exr: Check buf_size more completely
[23:01:25 CET] <cone-427> ffmpeg 03Michael Niedermayer 07release/3.3:54897d74663f: Update for 3.3.6
[23:01:40 CET] <atomnuker> nevcairiel: what does it mean? use guid instead of 2cc tag?
[23:13:11 CET] <atomnuker> apparently it does
[23:25:02 CET] <atomnuker> if there's one thing matroska got right it was codec id handling
[23:25:15 CET] <atomnuker> here's a string, it maps to a codec, that's it
[23:25:42 CET] <atomnuker> 1 line to add support for a new mapping with full muxing and demuxing
[23:25:50 CET] <nevcairiel> except their codec id mappings are the wild west
[23:26:08 CET] <nevcairiel> there is no proper official registration list
[23:26:38 CET] <atomnuker> well wav has one and look at the list of rogue format ccs
[23:26:52 CET] <atomnuker> the wild west is nice, as long as nothing overlaps
[23:39:05 CET] <cone-427> ffmpeg 03Luca Barbato 07release/3.3:96e8400553ae: x264: Support version 153
[23:39:06 CET] <cone-427> ffmpeg 03James Almer 07release/3.3:03292829aa2e: changelog: update with previous commit
[23:39:57 CET] <jamrial_> michaelni: make sure to include those two in 3.3.6, please
[00:00:00 CET] --- Sun Dec 31 2017