[Ffmpeg-devel-irc] ffmpeg-devel.log.20141217

[burek]

[02:14] <cone-916> ffmpeg.git 03Michael Niedermayer 07master:f3b5b139ad85: avcodec/h264: make the first field of H264Context an AVClass
[02:43] <cone-916> ffmpeg.git 03Supraja Meedinti 07master:abbcc68e5cef: libavutil: cosmetic: inserted spaces wherever necessary
[03:04] <timothy_gu> fflogger: why are you so confused
[04:30] <cone-916> ffmpeg.git 03Reimar Döffinger 07master:254da44bf97e: flacenc: calculate lower sum levels in-place.
[04:56] <jamrial> is cehoyos acting alone based on his own beliefs/knowledge regarding licensing with this digimetrics ticket?
[05:11] <timothy_gu> jamrial: do you think it's time for a third-party to intervene? cehoyos never said he was a lawyer
[05:11] <timothy_gu> and I don't like where the current situation is going
[05:20] Action: rcombs points vaguely in the direction of http://firecore.com/infuse
[05:23] <cone-916> ffmpeg.git 03Michael Niedermayer 07master:3305acdc92fa: avcodec/indeo3: use signed variables to avoid underflow
[05:23] <cone-916> ffmpeg.git 03Michael Niedermayer 07master:ddd7dac7ae05: avcodec/utils: check the private context class
[05:23] <cone-916> ffmpeg.git 03Michael Niedermayer 07master:3ba105029279: avcodec/imc: dont read bits beyond the end
[05:40] <jamrial> if he's not acting based on a general consensus but instead on his own beliefs, then i personally don't think we should let him talk for the entire project in legal matters just like that
[07:33] <fffan__> hi, fellows , which codec id should I use if I wanna encode MPEG-4 Part 2 video.
[08:12] <timothy_gu> fffan__: AV_CODEC_ID_MPEG4
[10:01] <fffan__> hi, timo, I am writing h.262 video to mpegts container, but the file can't be opened, could you help me? http://pastebin.com/3wQ4Rt2W
[10:13] <fffan__> fflogger, could you take a look at my pastebin, I tried 4 hours, nothing done.  http://pastebin.com/3wQ4Rt2W
[10:27] <anshul_mahe> I was implimenting scte message queue in ffmpeg, but I am having difficulty in parsing scte data from mpegts packet
[10:28] <anshul_mahe> handle_packet function of mpegts always return 0, and files end it return EOF
[10:30] <anshul_mahe> I think problem is since scte data is not contained in pes packet, it is loose in ts packet
[10:32] <wm4> does ffmpeg.c still use a shared AVCodecContext for demuxing and decoding?
[10:33] <anshul_mahe> dont know, was that question related to mine
[10:34] <wm4> no
[10:46] <anshul_mahe> does any know of some other codec which do not comes in pes packet but are there in ts file
[10:47] <nevcairiel> everything in ts is in pes
[10:48] <nevcairiel> unless its some special metadata magic
[10:51] <nevcairiel> carl sure has some weird understanding of how lgpl works
[10:52] <nevcairiel> it almost sounds like he thinks there is someone on our side that actually gives out licenses :p
[11:02] <anshul_mahe> nevcairiel: I am working on scte 35 messgae cue, it comes in stream without PES, do you have some suggestion, how should I handle without breaking any thing
[11:02] <nevcairiel> probably shouldnt
[11:02] <nevcairiel> the mepgts parser is rather strict about getting everything in pes
[11:03] <nevcairiel> how can you even have mpegts without pes, the mpegts part around the pes is not very big and cannot contain much information :d
[11:06] <anshul_mahe> Its not my stream, but they are broadcasted like that, would you like to see that video. I have also bought specification of those codec.
[11:06] <anshul_mahe> its all about https://trac.ffmpeg.org/ticket/3356
[11:10] <anshul_mahe> should I raise this issue in mailing list, since  now its matter of opinion that we want to handle data which is not in PES
[11:10] <nevcairiel> as i understand the spec from a quick glance, the PMT has a descriptor which tells you which PID is used for the scte stuff, so there is your answer, remember this information and do appropriate parsing
[11:11] <nevcairiel> there is already a check in handle_packet to differentiate between types, right now it has section and pes, so you add a new one
[11:12] <anshul_mahe> nevcairiel: thanks for hint 
[11:14] <anshul_mahe> If I understand correctly, is section type means  parsing pmt pat nit sdp etc
[11:34] <fffan__> hi, fflogger.
[11:57] <cone-41> ffmpeg.git 03Martin Storsjö 07master:2f628d5943c1: dashenc: Write segment timelines properly if the timeline has gaps
[11:57] <cone-41> ffmpeg.git 03Michael Niedermayer 07master:b32cb08b7f43: Merge commit '2f628d5943c12389c07d652d23d3916997f9f0f6'
[14:04] <cone-41> ffmpeg.git 03Martin Storsjö 07master:456e93bfdd4c: dashenc: Adjust the start time of a segment to the end of the previous segment
[14:04] <cone-41> ffmpeg.git 03Michael Niedermayer 07master:973684a4fba9: Merge commit '456e93bfdd4cbc5e995dea415019abd0703d0e16'
[15:45] <anshul_mahe> I am able to copy scte data from ts stream if I use map 0:0, but I am unable to copy scte stream without that
[15:46] <anshul_mahe> Any hints what I mihgt need to implement to copy data stream
[16:06] <cone-41> ffmpeg.git 03Michael Niedermayer 07master:7d593495e42e: avcodec/dcadec: Check that the added xch channel isnt already there
[17:00] <paradoxical> Hi there. I have a small question regarding encoding raw video (avi) in rgb2 color space to huffyuv avi in rgb24. I specify output with -pix_fmt rgb24 - but my output file is in rgb32 color-space? FFMPEG supports rgb24, so i am kinda lost why..
[17:02] <kepstin-laptop> there's no real difference between rgb24 and 32; the "32" format just adds an extra byte of padding in each pixel to make things line up nicer in memory.
[17:04] <paradoxical> that's a good thing. My main issue is that i want bit by bit identical files, this is for a research project (echo imaging) and it is ahrd to sell this to my boss ;)
[17:05] <wm4> paradoxical: maybe you need to force the raw encoder
[17:05] <wm4> probably using a fourcc
[17:06] <paradoxical> hmm I will look into that
[17:06] <nevcairiel> huffyuv supports both, so your pix fmt option was probably just in the wrong spot
[17:06] <paradoxical> ffprobe says the raw avi is in rgb24
[17:06] <paradoxical> thanks for these pointers
[17:07] <paradoxical> I will look into this tonight!
[17:08] <nevcairiel> looks like the huffyuv decoder in ffmpeg is unable to output to rgb24, so you can encode it like that, but the decoder will always pad it to rgb32
[17:08] <nevcairiel> probably mostly because 3-byte pixel formats are stupid to SIMD
[17:09] <paradoxical> Well my my go raw avi > huffyuv, so with ffmeg i am only encoding
[17:09] <nevcairiel> how did you verify what the huffyuv contained then?
[17:09] <paradoxical> and indeed, when i play in VLC it says RGB32!
[17:10] <nevcairiel> vlc uses the ffmpeg decoder, so yeah.
[17:10] <kepstin-laptop> if you want rgb24 output, you can always convert back from rgb32 to rgb24 without any loss.
[17:10] <paradoxical> I get what you mean now!
[17:10] <paradoxical> Thanks alot
[17:12] <paradoxical> I'll report in later, should be fine now!. Thanks again
[19:14] <kierank> wm4: are you gjdfgh
[19:15] <Daemon404> he is
[19:15] <Daemon404> or maybe not
[19:15] <Daemon404> i cant remember
[19:16] <wm4> yes
[19:30] <rcombs> mashed the middle of home row on the nick field?
[19:30] <wm4> yes
[19:31] <Daemon404> probably was meant as a one off
[19:31] <wm4> yes
[19:32] <Compn> yes but we kept trolling him , and now hes stuck responding on the trac :)
[19:32] <Daemon404> i dont use trac
[19:32] <Daemon404> because carl
[19:32] <gnafu> wm4: Any relation to Gjdfgh Fdgjdf?
[19:32] <gnafu> Or Fhsdfgd Gjdfgh?
[19:32] <wm4> no
[19:32] <Daemon404> it's his cousim
[19:33] <Daemon404> cousin
[19:33] <Compn> Daemon404 : why not use libav's bugzilla ?
[19:33] <gnafu> (I was surprised a Google search yielded anything other than Trac tickets.)
[19:33] <wm4> Compn: because it's dead?
[19:33] <Compn> then just paste those bugs here? :)\
[19:33] <Daemon404> Compn, why dont i just throw my bugs down into a pit?
[19:33] <Daemon404> same effect
[19:33] <Compn> wm4 : yeah but carl isnt on there (i dont think)
[19:33] <Compn> lol
[19:33] <wm4> although, didn't they recently add IRC notification to libav bugzilla
[19:33] <wm4> maybe that makes it slightly less dead
[19:33] <JEEBcz> yeah
[19:33] <JEEBcz> let's see
[19:34] <wm4> since now someone will actually see it
[19:34] <Daemon404> also: it's bugzilla
[19:34] <Daemon404> i hate bugzilla
[19:34] <Compn> everyone hates bugzilla
[19:34] <Compn> and trac
[19:34] <Compn> and roundup
[19:34] <rcombs> and all other bug trackers
[19:34] <Daemon404> zarro usability found
[19:34] <rcombs> (github issues have their share of problems)
[19:35] <wm4> you may hate bugzilla, but that doesn't mean there isn't worse
[19:35] <Compn> github requires javascript
[19:35] <Daemon404> wm4, lets switch to clearcase
[19:35] <Compn> google code just plain terrible 
[19:35] <nevcairiel> personally i like mantis if its configured decently
[19:35] <Daemon404> or was it clearquest
[19:35] <rcombs> even google doesn't use google code anymore
[19:35] <Daemon404> rcombs, google code is perfect for google
[19:35] <Daemon404> it enables them to ignore the community more efficiently
[19:36] <Compn> they love ignoring google products forum haha
[19:36] <nevcairiel> i should shut down my googlecode page, i dont use it anymore for anything, except ignoring user reports
[19:36] <Compn> i wonder when google is going to make homepage into a portal like yahoo /msn
[19:36] <rcombs> you mean iGoogle
[19:36] <rcombs> the thing they shut down a while back?
[19:36] <wm4> Compn: not doing so was exactly what made google popular
[19:36] <Compn> widgets
[19:37] <Compn> all the widgets
[19:37] <rcombs> also TIL Compn can't into JS
[19:37] <Compn> has a widget ever been useful?
[19:37] <Compn> javascript is mine enemy.
[19:37] <Daemon404> compn only recently upgarded off win2k
[19:37] <Compn> that and flash
[19:37] <Compn> hehe
[19:37] <rcombs> Compn: yes, if you mean UI widgets
[19:37] <wm4> Daemon404: to XP?
[19:37] <rcombs> like text boxes and buttons and shits
[19:37] <Compn> yes to xp
[19:37] <Daemon404> rcombs, Compnball?
[19:37] <Compn> it was a mistake, xp couldnt handle it either
[19:37] Action: Compn wants to go back to 2k
[19:38] <Daemon404> we're at 4k now, it's the future.
[19:38] <gnafu> Daemon404: 8K, actually.
[19:38] <gnafu> Get with the times, bro.
[19:38] Action: rcombs buys a 5K iMac
[19:38] <Compn> 16K should be enough for everyone
[19:38] <rcombs> (but actually not because I have no use for a desktop)
[19:39] <Compn> we should invent a 4d byte packing (or is that alpha?) rgb128, 32x32x32x32
[19:39] <Compn> to go with these 8k videos
[19:39] <Compn> "our codec is best, it has extra bits!"
[19:41] <wm4> lol 32 bits... you don't even get 8 bits
[20:00] <cone-41> ffmpeg.git 03Michael Niedermayer 07master:8aa8d1255486: avcodec/hevc: clear filter_slice_edges() on allocation
[20:01] <ubitux> can anyone confirm the -show_region in x11grab still works?
[20:01] <ubitux> when i'm using it a region is drawn but the content is not visible
[20:01] <ubitux> like, it's keeping some kind of "screenshot" of the area
[20:06] <ubitux> also, the framerate looks totally broken
[20:11] <ubitux> maybe that's just 2.5* for the frame rate
[20:14] <j-b> kierank: http://forum.doom9.org/showthread.php?t=171543
[20:30] <Compn> oh hohohoho
[20:31] <Compn> hows x262 coming ?
[20:59] <kierank> j-b: yeah i know that guy
[21:02] <kierank> Compn: it works ok
[22:14] <cone-41> ffmpeg.git 03Michael Niedermayer 07master:e8714f6f93d1: avcodec/h264: Clear delayed_pic on deallocation
[22:59] <wm4> that's... interesting
[23:01] <jamrial> yeah, fflogger has been acting up
[23:02] <wm4> the statement about the shittymetrics situation too
[23:06] <Daemon404> thats not a nice name to call carl, wm4
[23:07] <wm4> u
[00:00] --- Thu Dec 18 2014