[02:25:33 CET] <cone-016> ffmpeg 03Michael Niedermayer 07master:e94288c4932e: avcodec/nuv: add FF_CODEC_CAP_INIT_CLEANUP [02:25:33 CET] <cone-016> ffmpeg 03Linjie Fu 07master:5764c9528b97: lavc/hevc_parser: report detailed log when missing picture occurs [09:40:04 CET] <cone-817> ffmpeg 03Peter Ross 07master:ad0d5d7516dc: avcodec/dstdec: use appropriate alignment [13:18:24 CET] <pross> bit-exact vp6 achieved. now time to do something actually productive. [13:18:35 CET] <JEEB> najs [13:19:19 CET] <JEEB> yea, I got back from my holidays and I'll probably start poking at stuff I want to work with again [13:20:32 CET] <pross> classic On 2 style code reuse. VP 3/4/5/6 codecs all use the same loop filter design. [13:20:40 CET] <pross> cool! [13:20:54 CET] <durandal_1707> lies, lies and lies [13:21:23 CET] <JEEB> https://github.com/jeeb/ffmpeg/commits/mpegts_arib_stuff [13:21:27 CET] <JEEB> rebased this while on flight :D [13:23:47 CET] <Compn> pross : fixes all vp6 artifact bugs? :) [13:24:18 CET] <Compn> hmm http://trac.ffmpeg.org/ might be down ? [13:24:41 CET] <Compn> pross : http://trac.ffmpeg.org/ticket/6149 [13:26:25 CET] <Compn> interlaced vp6 http://trac.ffmpeg.org/ticket/5581 [13:28:17 CET] <Compn> this was the old bug i was thinking of though http://trac.ffmpeg.org/ticket/1282 [13:28:38 CET] <Compn> thought there was another one too hmm [13:28:57 CET] <Compn> oh you commented on that one :D [13:29:08 CET] Action: Compn goes back under his rock to wake up [13:30:38 CET] <pross> Compn: yes, artifacts all gone! interlaced support, hell no. [13:32:01 CET] <durandal_1707> pross: have you split codec? introducing new id? [13:32:27 CET] <JEEB> &41 [13:33:02 CET] <Compn> JEEB : the subtitle people will love you :D [13:33:12 CET] <Compn> or is it caption heh [13:33:44 CET] <durandal_1707> Compn: you are getting older and older [13:35:56 CET] <pross> durandal_1707: updated patch not ready yet. i wanted to understand the vp5/6 loop filter first [13:38:47 CET] <Compn> durandal_1707 : its just my buffer is full. need to flush cache [13:42:07 CET] <Compn> durandal_1707 : guy replied to your question http://trac.ffmpeg.org/ticket/959 [13:51:57 CET] <JEEB> Compn: same thing, basically :) text stuff is generally in broadcast called captions, I guess? (see: DVB subs being pictures) [13:52:21 CET] <JEEB> although teh funny thing with ARIB captions is that not only you have styling and positioning, but also you have in-band images [13:52:32 CET] <JEEB> not sure if I want to go that far, but I've seen OSS implementations of it :) [13:54:23 CET] <BtbN> Did the coverity tool download URL change or something? [13:57:22 CET] <atomnuker> lol, vp6 supported interlacing? [14:00:46 CET] <durandal_1707> no more coverity crap mails - finally freedom! [15:24:19 CET] <BtbN> https://community.synopsys.com/s/article/Coverity-Scan-Update oh, it's actually dead for now [15:44:42 CET] <nevcairiel> again? [15:45:01 CET] <BtbN> Their hoster has vanished from the face of earth or something [16:17:31 CET] <cone-614> ffmpeg 03Paul B Mahol 07master:f4fd8ef725a8: avfilter/af_anlmdn: rework how denoising strength is used [16:45:55 CET] <cone-614> ffmpeg 03Jerome Borsboom 07master:96740ec69f27: avcodec/vc1: shuffle calculation of MV predictor candidates [16:45:56 CET] <cone-614> ffmpeg 03Jerome Borsboom 07master:b6402278322d: avcodec/vc1: fix decoding of old WMV3 format [16:47:31 CET] <cone-614> ffmpeg 03Michael Niedermayer 07master:f9728feaf90e: avcodec/exr: Check for duplicate channel index [16:47:32 CET] <cone-614> ffmpeg 03Michael Niedermayer 07master:433d2ae4353f: avcodec/exr: set layer_match in all branches [16:47:33 CET] <cone-614> ffmpeg 03Michael Niedermayer 07master:77e56d74f972: avcodec/h264_slice: Fix integer overflow in implicit_weight_table() [17:18:05 CET] <cone-614> ffmpeg 03Carl Eugen Hoyos 07master:261e4cf03f9a: lavc/psd: Support CMYK images. [17:30:29 CET] <cone-614> ffmpeg 03Carl Eugen Hoyos 07master:90ab9a58ff69: lavc/tiff: Support CMYK images. [17:33:37 CET] <cone-614> ffmpeg 03Carl Eugen Hoyos 07master:2e8b0446c679: lavc: Allow very high bitrates in AVCPBProperties after next version bump. [18:07:01 CET] <JEEB> 9 [18:11:21 CET] <BtbN> Unless someone gets me one of those cards, there won't ever be much I can do in those extreme edge cases... [18:11:35 CET] <BtbN> 40 parallel streams. [18:12:27 CET] <BtbN> Also no idea what "ffmpeg was work unstable" is supposed to mean. [18:12:34 CET] <durandal_1707> https://whoislewys.com/2018/10/18/audio_norm_guides_are_evil_and_immoral/ ----> what i should write there in comments? [18:25:39 CET] <BtbN> what? [18:25:55 CET] <ePirat> o_o [18:43:26 CET] <kurosu> pross: so you checked the old vp6 ticket, and confirmed it is the in-loop filter? [18:43:50 CET] <kurosu> pross: also, bitexact, meaning you checked against the binary decoding (using eg mplayer loader) ? [19:52:18 CET] <cone-614> ffmpeg 03Paul B Mahol 07master:1ea5529dd260: avfilter: add maskfun filter [20:07:11 CET] <durandal_1707> atomnuker: when you finish fft/dct ? [20:23:09 CET] <Compn> ehe kurosu also does not believe bit exact vp6 :D [20:23:12 CET] <Compn> same as me [20:26:30 CET] <durandal_1707> Compn: why are you still connected? [20:46:09 CET] <Compn> durandal_1707 : mplayer isnt dead yet :D [20:47:52 CET] <durandal_1707> Compn: it's dead, and you can't do shit [21:02:46 CET] <kurosu> Compn, I'm not saying this or believing that [21:03:02 CET] <kurosu> just that if he is saying this, he has tested something, and it improved [21:03:08 CET] <kurosu> (besides vp4) [21:03:54 CET] <kurosu> bye [23:14:02 CET] <j-b> durandal_1707: do you know if AAC-ELD is supported? [23:17:50 CET] <JEEB> there was some type of AAC with a very specific sample count which wasn't supported in the lavc decoder [23:18:44 CET] <nevcairiel> AAC ELD should be supported for decoding [23:19:04 CET] <JEEB> ok [23:19:13 CET] <nevcairiel> although perhaps not necessarily all bitstream features [23:19:57 CET] <nevcairiel> http://git.videolan.org/?p=ffmpeg.git;a=commitdiff;h=b3be41ca82529d60f90107d... [23:20:10 CET] <nevcairiel> the 960 transform was since added [23:20:13 CET] <nevcairiel> not sure about the others [00:00:00 CET] --- Sun Jan 13 2019
participants (1)
-
burek