[00:18:29 CET] <ubitux> mmh it seems there is a subtitles format that look just like microdvd but uses [] instead of {} [00:18:35 CET] <ubitux> i wonder what's that format [00:20:51 CET] <wm4> vplayer? [00:22:25 CET] <ubitux> don't think so [00:23:56 CET] <ubitux> it really is similar to microdvd [00:23:57 CET] <ubitux> [3579][3607]Nie... piraci... [00:23:59 CET] <ubitux> [3672][3708]E.J. Lotterman|Redaktor naczelny [00:24:01 CET] <ubitux> [3728][3753]Nie teraz. [00:25:02 CET] <cone-869> ffmpeg 03Paul B Mahol 07master:4a7c705fde1a: avfilter/vf_vectorscope: make it possible to override colorspace [00:27:02 CET] <ubitux> they look quite common actually [00:27:48 CET] <ubitux> oh we actually support them, it's simply mpl2, i forgot i did this one as well [00:29:32 CET] <wm4> lol [00:30:14 CET] <wm4> if you make stats, I'd be interested what types of tags happen in which formats [00:30:24 CET] <wm4> ass, microdvd, html [00:49:47 CET] <ubitux> yep [00:52:08 CET] <Zeranoe> Lots of bugs for h264_qsv.... Another one was just posted on my forum [00:53:39 CET] <Compn> looks like mpl :P [00:53:50 CET] <Compn> surprised anyone uses it to be honest [00:54:02 CET] <Zeranoe> I am too [00:54:59 CET] <michaelni> jamrial, thx for e71096bb3ab4310999731dc53988cdcbd8aa4a72 [00:55:00 CET] <Zeranoe> Based on a recent email I got from Intel http://links.mail3.spopessentials3.com/servlet/MailView?ms=MjQ5MzY2NjQS1&r=MjI0NzEwNTQ4ODcyS0 they're pushing it pretty hard, so that might be part of it. [00:55:17 CET] <Compn> i was talking about ubitux's subtitle ehehe [00:55:25 CET] <Compn> but yeah people want to use intel encoder [00:56:15 CET] <Zeranoe> From what I've found it produces worse quality than x264, but it does it faster [00:56:38 CET] <ubitux> mmh so the uncompressed dump of subtitles is 195G [00:56:44 CET] <Compn> nice [00:58:09 CET] <ubitux> which makes an average size of ~54kB per subtitles file [00:58:26 CET] <ubitux> this was the useless stat of the day [01:02:20 CET] <Compn> how many files total [01:02:33 CET] <Compn> must be thousands [01:02:39 CET] <Compn> :P [01:02:56 CET] <Compn> millions [01:06:06 CET] <ubitux> about 3.8 millions [01:08:16 CET] <Compn> i hope your harddrive fs can handle it [01:12:03 CET] <ubitux> well i'm not running on fat32 or worse (like hfs+) so it should be fine [01:12:31 CET] <Compn> are you going to write up a blog post detailing the great subtitle sample ? [01:12:36 CET] <Compn> i like reading such blog posts [01:14:43 CET] <ubitux> probably, but not anytime soon [01:15:02 CET] <ubitux> there are a lot of things to analyze, and probably a lot of things to fix in ffmpeg first :) [01:19:52 CET] <cone-869> ffmpeg 03Michael Niedermayer 07master:52e7f6b17eed: Revert "avfilter/vf_decimate: Check that input parameters match" [01:19:53 CET] <cone-869> ffmpeg 03Michael Niedermayer 07master:58cb1fb1fe92: avfilter/vf_decimate: Use the correct frame in difference calculation [01:33:27 CET] <ubitux> 1 [01:33:28 CET] <ubitux> 00:00:19,740 -> 00:00:22,740 [01:33:30 CET] <ubitux> omfg [01:33:42 CET] <ubitux> (supposed to be "-->", whole file like this) [01:34:04 CET] <Compn> haha [01:35:01 CET] <Compn> i think i've ran into a sub with that problem before [01:35:10 CET] <ubitux> 1 [01:35:11 CET] <ubitux> 00:00:-13,000 --> 00:00:04,881 [01:35:13 CET] <ubitux> lol [01:35:40 CET] <Compn> how do you know the video DOESNT start with -13 timestamp ? [01:35:42 CET] <Compn> how do you KNOW [01:36:13 CET] <ubitux> dammit this archive is a fucking gold [01:36:27 CET] <ubitux> 00:00:46,913-->00:00:49,006 [01:36:44 CET] <Compn> no spaces [01:36:46 CET] <Compn> i kind of want a copy. [01:37:01 CET] <ubitux> i can't share it [01:37:13 CET] <Compn> ah dang [01:37:18 CET] <ubitux> A [01:37:19 CET] <ubitux> 00:00:01,011 --> 00:00:04,074 [01:37:21 CET] <ubitux> omfg @!# [01:37:33 CET] <ubitux> 'A' [01:37:37 CET] <Compn> the next line isnt B is it ? [01:37:40 CET] <ubitux> no... [01:37:43 CET] <ubitux> next are 2, 3, 4... [01:37:48 CET] <ubitux> but the first one is a fucking A [01:37:48 CET] <Compn> how are you finding the non standard files btw ? [01:37:50 CET] <Compn> lol [01:37:58 CET] <ubitux> i'm feeding them to ffprobe [01:38:04 CET] <ubitux> and look at the failing ones [01:38:06 CET] <Compn> edited ffmpeg to output all non standard lines [01:38:06 CET] <Compn> ah [01:38:21 CET] <ubitux> <br /> [01:38:23 CET] <ubitux> <b>Warning</b>: filesize() [<a href='function.filesize'>function.filesize</a>]: stat failed for /24.401-med.srt in <b>C:\Szerver\xampp\htdocs\downloadsubs.php</b> on line <b>12</b><br /> [01:38:25 CET] <ubitux> lol [01:39:27 CET] <ubitux> omg, some guy edited a file with wordpad [01:39:37 CET] <Compn> wordpad added cruft ? [01:39:39 CET] <Compn> ehe [01:39:41 CET] <ubitux> yeah [01:39:56 CET] <Compn> yep gonna be a lot of crlf vs lf no doubt [01:40:23 CET] <ubitux> well i think it's wordpad [01:40:37 CET] <ubitux> http://b.pkh.me/b44c42f400d6e45850d75543c3eb93a5 [01:41:09 CET] <ubitux> damn i really need to push my fix for vplayer [01:41:50 CET] <ubitux> i see a non negligible amount of srt files where the first chapter is simply absent [01:41:55 CET] <ubitux> and it starts right away with a timing [01:42:44 CET] <Compn> so many different subtitle softwares out there, you're going to see some weird stuff [01:43:07 CET] <Compn> country specific stuff. [01:43:24 CET] <ubitux> heh there are some webvtt files [01:44:52 CET] <Compn> any real text .rt ? [01:45:30 CET] <ubitux> Compn: none so far, but it's going to take a while to probe everything [01:45:32 CET] <ubitux> -1 [01:45:34 CET] <ubitux> 00:00:03,320 --> 00:00:39,845 [01:45:37 CET] <ubitux> omg this is a joke [01:45:38 CET] <Compn> haha [01:45:46 CET] <Compn> does it go to 0 or just 1 ? [01:45:57 CET] <ubitux> -1 0 1 2 ... [01:46:08 CET] <Compn> better than A > 1 then [01:46:16 CET] <Compn> at least its some kind of pattern [01:46:48 CET] <Compn> er A > 2 [01:47:16 CET] <ubitux> they really like negative stuff [01:47:18 CET] <ubitux> 00:00:-5,-601 --> 00:00:03,157 [01:47:32 CET] <ubitux> i have a lot of them like this [01:48:12 CET] <ubitux> nice, a file starting with microdvd subtitles [01:48:20 CET] <ubitux> and then suddenly it becomes srt [01:48:31 CET] <Compn> not at the 1cd / 2cd mark? [01:50:15 CET] <ubitux> no, just one line like this: http://b.pkh.me/555de38a7c4bddfe8c3c964948a9495b [01:51:23 CET] <ubitux> oh nice, twice the utf-8 bom [01:52:28 CET] <ubitux> some files are so broken [01:52:32 CET] <ubitux> they are detected as AAC [01:52:40 CET] <Compn> good for fixing aac probe then [01:53:02 CET] <Compn> are we allowed to add filesizes to the probes (when available) ? [01:53:20 CET] <ubitux> probing is only buffer based [01:53:31 CET] <ubitux> like, content only, that's all you get [01:54:00 CET] <wm4> more like aac probing is so broken it reacts to subtitle files [01:54:11 CET] <wm4> at least it's not mp3 [01:55:33 CET] <Compn> right [02:06:10 CET] <ubitux> http://b.pkh.me/3c127e35bcce41698fb9665e32df55ce lol [02:08:01 CET] <ubitux> http://b.pkh.me/3a034b410c18041904976930324f6562 [02:08:53 CET] <Compn> mixing html / ass type and srt type [02:08:56 CET] <Compn> hehe [02:10:11 CET] <ubitux> http://b.pkh.me/6f990b214cb514f35ff4de20c636a49a [02:10:23 CET] <ubitux> some people are basically thinking these files are free editable zones [02:11:08 CET] <ubitux> i actually see this pattern a lot... [02:11:19 CET] <ubitux> it's generally 2 or 3 lines [02:17:22 CET] <ubitux> i think i saw enough for now [04:47:03 CET] <alexandre_> Good night, I'm interested in participating gsoc and work on ffmpeg. Looking out for info with a possible mentor. Thanks in advance. [05:01:10 CET] <Compn> alexandre_ : did you see the ffmpeg gsoc wiki ? [05:01:31 CET] <Compn> alexandre_ : read this page https://trac.ffmpeg.org/wiki/SponsoringPrograms/GSoC/2016 [05:01:45 CET] <Compn> anyone want to write a gsoc/opw news entry for ffmpeg.org ? [05:01:50 CET] <Compn> we need one badly... [05:05:26 CET] <alexandre_> thanks for answering, i saw the wiki. I will try to reach a mentor directly and sorry for spamming [05:05:46 CET] <Compn> its ok, feel free to ask any questions here [05:05:57 CET] Action: Compn goes afk [08:15:58 CET] <ubitux> http://b.pkh.me/bb8bc4a6e0e845517943fb49be5465b6 [08:16:02 CET] <ubitux> i don't get it. [08:22:53 CET] <wm4> hilarious [08:23:36 CET] <wm4> just wtf must have been going on in this user' shead [08:24:58 CET] <wm4> and what program parses this... oh well the world is a bad place [09:06:08 CET] <JEEB> wow [09:42:36 CET] <cone-921> ffmpeg 03Carl Eugen Hoyos 07master:db4786f1cf03: lavc/hevc_ps: Support GBR decoding. [09:42:36 CET] <cone-921> ffmpeg 03Carl Eugen Hoyos 07master:d11d78facb1b: lavc/libx265: Support GBR encoding. [09:50:13 CET] <durandal_170> dev can be only mentor for one project [10:15:35 CET] <cone-921> ffmpeg 03Thomas Mundt 07master:5024a82e9548: avfilter/vf_bwdif: add x86 SIMD [11:29:56 CET] <cone-921> ffmpeg 03Paul B Mahol 07master:912fec3e54ae: avfilter/vf_waveform: add acolor filter [13:13:31 CET] <cone-921> ffmpeg 03Mats Peterson 07master:f680c8e41a9b: tests/ref/vsynth: Remove unused file [13:13:32 CET] <cone-921> ffmpeg 03Mats Peterson 07master:ed5d295a6983: lavf/riffenc: Improve spec compliance; Fix WMP playback of AVI with xxpc chunks [13:16:17 CET] <cone-921> ffmpeg 03Carl Eugen Hoyos 07master:4405ff2cae50: lavc/hevc_ps: Fix default display window offsets for yuv422 and yuv444. [16:06:53 CET] <cone-921> ffmpeg 03Mats Peterson 07master:9ee5b4096476: avformat/avienc: factor out avi_add_ientry() [16:06:54 CET] <cone-921> ffmpeg 03Mats Peterson 07master:77bc3e1944a9: lavf/avienc: Add xxpc entries to index [16:47:39 CET] <cone-921> ffmpeg 03Mats Peterson 07master:bdc798cc8d00: lavf/avienc: Clear whole tag in avi_add_ientry() [17:41:53 CET] <cone-921> ffmpeg 03Ganesh Ajjanagadde 07master:0c33ceda91dd: lavu/lfg-test: add simple sample mean, stddev checks [18:20:26 CET] <cone-921> ffmpeg 03Reimar Döffinger 07master:7c93f2c0b940: Move cbrt tables to separate cbrt_data(_fixed).c files. [19:27:27 CET] <omerjerk> I have a question regarding libavcodec. [19:28:12 CET] <omerjerk> Support there's a decoder (being specific to alsdec.c), will it always convert to .wav only ? [19:31:48 CET] <BtbN> decoders decode to raw audio, that's their job. [19:31:56 CET] <nevcairiel> it decodes to raw pcm [19:32:04 CET] <JEEB> libavcodec either decodes in some way encoded things into one of libavcodec's raw format [19:32:12 CET] <nevcairiel> which can then be processed into .wav, or re-encoded into any other format [19:32:12 CET] <JEEB> or encodes raw formats to some coded format [19:33:44 CET] <omerjerk> okay, so my qualification task is to add floating point support in the alsdec. [19:34:05 CET] <omerjerk> I know how to generate the floating number from the bitstream. [19:35:04 CET] <omerjerk> but, I'm not able to understand how to exactly return that sampled float data to ffmpeg for it to write. [19:36:03 CET] <omerjerk> there's this ALSBloackData struct - https://github.com/FFmpeg/FFmpeg/blob/master/libavcodec/alsdec.c#L244 [19:40:27 CET] <omerjerk> so my question is, if I write float value to *dest, will everything work ? - https://github.com/FFmpeg/FFmpeg/blob/master/libavcodec/alsdec.c#L1554 [19:40:44 CET] <omerjerk> or is there some other place too where I'll need to make changes. [19:40:51 CET] <omerjerk> Just asking in a general sense. [19:41:08 CET] <omerjerk> A general answer would suffice. [19:46:17 CET] <wm4> yes, but you have to set the sample format to a float one too [19:48:26 CET] <omerjerk> hey. I'm really sorry. My internet got disconnected. [19:48:42 CET] <omerjerk> I didn't see any messages if anyone replied after my last message. [19:48:49 CET] <fritsch> 19:46 <+wm4> yes, but you have to set the sample format to a float one too [19:48:53 CET] <fritsch> ^^ was the only answer [19:49:31 CET] <omerjerk> Thanks a lot. [19:49:35 CET] <omerjerk> Where do we have to set that ? [19:49:46 CET] <fritsch> :-) if it's made too easy [19:49:51 CET] <fritsch> you don't learn anything [19:49:54 CET] <fritsch> while implementing it [19:51:14 CET] <omerjerk> Okay. I'll see. Hopefully I'll be able to find it. :) [19:51:16 CET] <omerjerk> Thanks. [20:12:43 CET] <cone-921> ffmpeg 03Paul B Mahol 07master:80974559294f: avfilter/vf_vectorscope: copy frame props, mainly for bench filter [20:12:44 CET] <cone-921> ffmpeg 03Paul B Mahol 07master:55c872146d45: avfilter/vf_vectorscope: process alpha at end [20:12:45 CET] <cone-921> ffmpeg 03Paul B Mahol 07master:370cecc1e87d: avfilter/vf_vectorscope: less aggressive memory allocation [00:00:00 CET] --- Mon Mar 14 2016