[Ffmpeg-devel-irc] ffmpeg-devel.log.20130608

[burek]

[01:40] <saste> i'm going to propose an irc meeting on the next saturday
[01:41] <saste> i'm going to propose 16 UTC but feel free to suggest a better time
[01:42] <ubitux> fine with me, what topic?
[01:42] <saste> ubitux, various
[01:43] <saste> i'd like to discuss things related to donations, crowdfunding and stuff
[01:43] <ubitux> ok
[01:43] <saste> but it's been a long time since the last meeting so there could be much more to discuss
[01:44] <ubitux> will it be public, here?
[01:45] <saste> ubitux, what about a dedicated channel?
[01:46] <ubitux> sure whatever
[01:46] <saste> so we don't spam ffmpeg-devel, and preserve privacy
[01:55] <llogan> saste: about when was the last meeting?
[01:56] <saste> llogan, i remember one back in the 2011, but maybe it was an fftech thing
[01:56] <llogan> i haven't heard that word in a while. is fftech still in existance?
[01:56] <saste> ffmtech thing
[01:56] <saste> i don't know
[01:57] <saste> i think it is formally live, but it is not active
[01:58] <llogan> i will try to be there.
[01:58] <michaelni> the meeting should be public
[01:58] <ubitux> saste: maybe it could be worth mentioning global infrastructure (servers & stuff); i guess arpi would be welcome to discuss as well
[01:58] <ubitux> maybe ETOOMUCHTOPICS though
[01:59] <saste> ubitux, sure
[01:59] <saste> michaelni, i want to keep a good signal/noise ratio
[02:01] <saste> what happens is that often such meetings end up as a wishlist session from people not really involved with development, i want to avoid that
[02:01] <llogan> we could silence the bots and just do it here. that way it can be in our standard irc logs
[02:01] <saste> also it could be good to discuss things privately, with no public log and preserve a bit of privacy
[02:02] <saste> that said, i won't fight if you insist to keep it public
[02:02] Action: ubitux doesn't care either way
[02:02] <llogan> saste: i don't foresee wishlisting happening, but then maybe i'd be wrong and owe you a beer.
[02:02] <ubitux> though, a public dedicated muted channel where we voice contributors & dev might be more appropriate
[02:03] <michaelni> saste, didnt you had "more transparency" or something like that on your ffmtech director agenda ?
[02:06] <saste> michaelni, i'm no more part of the ffmtech board
[02:07] <saste> also even in that case the few meetings were private, and we provided a public report
[02:07] Action: michaelni doesnt understand the call for "privacy" for discussions about a free / open sourtce project
[02:08] <saste> michaelni, free / open source doesn't imply that all must be public at all costs
[02:09] <saste> and sometimes you have more freedom if not everything that you say/write is publically logged
[02:23] <cone-250> ffmpeg.git 03Michael Niedermayer 07master:56abad0ee4a3: avcodec/smvjpegdec: copy palette too
[02:27] <michaelni> saste, noone stops you or anyone else from discussing anything privately with anyone. But IMHO a (once on 2 years) meeting about ffmpeg should be public
[02:36] <Compn> llogan : i'd probably try to wishlist it, so maybe i should stay out :D
[02:36] <Compn> ehe
[02:37] <Compn> michaelni : we can always have a private and a public meeting :P
[02:38] <Compn> i can probably guess what the private meeting topic is
[02:39] <Compn> i have a pretty good guess , that is
[02:39] <Compn> and i agree it shouldnt be public, for various stupid stupid reasons
[02:40] <llogan> anyone got another contact for herve? i tried 'herve at whynotflores' and 'herve.flores at free fr'.
[02:44] <Compn> hmm
[02:44] <Compn> whats his website ? where he hosted all of the logos he created ?
[02:46] <llogan> i just wanted to know what typeface he used for the logo. even whatthefont didn't know.
[02:48] <Compn> which logo ?
[02:49] <llogan> the current one on the site
[02:49] <llogan> i'm making stickers.
[02:50] <Compn> did he post a file that contained all of the required stuff ? 
[02:50] <Compn> i forgot what image format , svg ?
[02:51] <llogan> yes, svg
[02:51] <llogan> IIRC
[02:51] <llogan> but i don't just want text outlines.
[02:56] <Compn> llogan : http://www.movie50p60p.movieconverter-studio.com/
[02:56] <Compn> i think herve runs that , try that email at the bottom ?
[02:56] <Compn> maybe its just a guy with same name :D
[02:57] <Compn> looks like hes on facebook too
[03:10] <Compn> hey highgod , still working hard? :)
[03:16] <Compn> yes, that movieconverter-studio is his site, its linked to his whynotflores site
[03:16] <Compn> llogan : in case you were wondering.
[03:19] <highgod> Hi, Compn, yes very busy these days
[11:13] <cone-388> ffmpeg.git 03Luca Barbato 07master:92b3caa1089b: vorbis: return meaningful errors
[11:13] <cone-388> ffmpeg.git 03Luca Barbato 07master:f8a4d5e970f3: h264_parser: K&R formatting cosmetics
[11:13] <cone-388> ffmpeg.git 03Michael Niedermayer 07master:e2d4bcd7b844: Merge commit 'f8a4d5e970f32f8cc747f9fa7bd975ee4a060ea1'
[11:28] <cone-388> ffmpeg.git 03Luca Barbato 07master:103937681d7e: tiff: refactor deflate support in a separate function
[11:28] <cone-388> ffmpeg.git 03Michael Niedermayer 07master:a804632b155d: Merge commit '103937681d7e6afd3e8caf8937272575e450fb55'
[11:39] <cone-388> ffmpeg.git 03Luca Barbato 07master:016c5b066de0: tiff: refactor fax support in a separate function
[11:39] <cone-388> ffmpeg.git 03Michael Niedermayer 07master:1fec361f353c: Merge commit '016c5b066de08a93a5f6b5beb0ef377356b35cde'
[12:12] <cone-388> ffmpeg.git 03Luca Barbato 07master:999ccd2d0a43: tiff: cleanup the error paths
[12:12] <cone-388> ffmpeg.git 03Michael Niedermayer 07master:dc318505d010: Merge commit '999ccd2d0a43640921088578f138c874f6cc0f8a'
[12:27] <cone-388> ffmpeg.git 03Luca Barbato 07master:9c2216976907: tiff: do not overread the source buffer
[12:27] <cone-388> ffmpeg.git 03Michael Niedermayer 07master:d0b27eab925c: Merge commit '9c2216976907336dfae0e8e38a4d70ca2465a92c'
[12:37] <cone-388> ffmpeg.git 03Luca Barbato 07master:f32aefcf349f: faxcompr: K&R formatting cosmetics
[12:37] <cone-388> ffmpeg.git 03Michael Niedermayer 07master:a5a9bcc8760a: Merge commit 'f32aefcf349f2f1367f0143782397643416d18f3'
[12:43] <cone-388> ffmpeg.git 03Luca Barbato 07master:a3b2b83f01a4: faxcompr: return meaningful errors
[12:43] <cone-388> ffmpeg.git 03Michael Niedermayer 07master:f973f1741649: Merge remote-tracking branch 'qatar/master'
[14:06] <cone-388> ffmpeg.git 03Timothy Gu 07master:ba7cb4807f05: doc/encoders: Add libvo-aacenc doc
[14:16] <highgod> Hi, I want to reference the solution of ffmpeg process mpeg2 audio, but I did't find it, does anyone konw which file is it in? thanks
[14:20] <cone-388> ffmpeg.git 03Michael Niedermayer 07master:f320b00541fc: doc/codecs: Mention relation of the experimental setting and codecs
[14:21] <highgod> we use the packet directly but we found that the payload is different than the value we get from ffmpeg payload got form av_read_frame(is->ic, pkt);
[16:05] <cone-388> ffmpeg.git 03Lukasz Marek 07master:23a76b71deea: ftp: reconnect on tcp read error
[16:05] <cone-388> ffmpeg.git 03Lukasz Marek 07master:43eda88200a7: ftp: fix flush control connection input
[16:05] <cone-388> ffmpeg.git 03Lukasz Marek 07master:bc29acdc76fd: ftp: explicit return code checks
[16:05] <cone-388> ffmpeg.git 03Lukasz Marek 07master:4d617715c93d: ftp: abort function optimalization
[16:05] <cone-388> ffmpeg.git 03Lukasz Marek 07master:eeedca4b7f4b: ftp: fix seeking beyond file size
[16:05] <cone-388> ffmpeg.git 03Lukasz Marek 07master:89b4800eef7b: ftp: probe seek capability
[16:05] <cone-388> ffmpeg.git 03Lukasz Marek 07master:ff35c7cdfac3: ftp: add invalid code for RETR operation
[16:05] <cone-388> ffmpeg.git 03Michael Niedermayer 07master:33fb22f800dc: Merge remote-tracking branch 'lukaszmluki/master'
[19:35] <ubitux> michaelni: unless i'm missing something, the case with qp tables have h=1 can not work in spp (because of the h accesses in filter())
[19:44] <cone-310> ffmpeg.git 03Paul B Mahol 07master:9644fc95befc: matroskadec: move ALAC extradata creation bellow AAC one
[19:44] <cone-310> ffmpeg.git 03Paul B Mahol 07master:b94f045b18f1: wavpack: add missing .\n to entropy error message
[19:47] <ubitux> i can't really test it though, since the qp table are not exported
[19:48] <ubitux> (with mp=spp)
[19:48] <durandal11707> ahh, yet another filter then does not work at all
[19:48] <ubitux> it does
[19:48] <ubitux> you can use a constant qp
[19:49] <ubitux> it just doesn't honor the qp tables from the codecs
[19:56] <ubitux> btw, it would be nice to port qp filter
[19:56] <durandal11707> EBORING
[19:57] <ubitux> i guess the filter could be useful in certain situation
[19:57] <ubitux> such as keeping most details in the center
[19:58] <ubitux> still no one to write the gamma/contrast/brightness/saturation filter?
[20:01] <ubitux> durandal11707: btw, weren't you motivated to work on mpegts a while ago?
[20:01] <iive> ubitux: you mean, port mp_eq2
[20:01] <ubitux> iive: saste raised some objections about the port
[20:01] <ubitux> because GPL
[20:02] <ubitux> (and that filter is relatively "essential")
[20:02] <ubitux> while we could have the optim in GPL, i agree with him about the main feature
[20:02] <iive> and, what is the problem with relicensing?
[20:02] <wm4> mpegts really needs work
[20:02] <ubitux> iive: ETOOMUCHCONTRIBS
[20:03] Action: iive facepalms
[20:03] <ubitux> wm4: yes, demuxer and muxer :(
[20:03] <wm4> can't vlc's libdvbpsi be used?
[20:03] <ubitux> project name is ffmpeg
[20:03] <wm4> I don't know enough about this stuff to tell
[20:04] <ubitux> it would be a shame to rely on external libraries for... mpeg :p
[20:04] <wm4> uh what
[20:04] <wm4> that doesn't warrant (very bad) NIH
[20:04] <durandal11707> ubitux: nope, i was looking for someone to do it
[20:04] <wm4> when in doubt you still could copy all its code
[20:04] <ubitux> durandal11707: didn't you find yourself since then?
[20:04] <wm4> (which seems to be ok ;)
[20:04] <durandal11707> EDOINGSOMETHINGELSEIACTUALLYREALLYCAREABOUT
[20:04] <ubitux> wm4: well, i believe the best course here is just to have improve our code
[20:05] <ubitux> durandal11707: ff related?
[20:05] <wm4> well, I was just hinting that VLC has a very good implementation
[20:05] <wm4> so maybe it could be used to reduce waste of effort
[20:05] <ubitux> wm4: ok, that's good to know
[20:05] <ubitux> but i'm not planing to work on mpegts anyway
[20:05] <ubitux> i would just love to see the number of tickets on http://trac.ffmpeg.org/query?status=!closed&keywords=~mpegts reduced :p
[20:07] <cone-310> ffmpeg.git 03Michael Niedermayer 07master:9e9ec5ad2ab9: smvjpegdec: No frame at all is an error
[20:07] <cone-310> ffmpeg.git 03Michael Niedermayer 07master:2a91038e13a0: avformat/wavdec: Dont rescale timestamps but use exact comparission
[21:39] <ubitux> oh the height is actually multiplied by the qp stride... ok.
[21:59] Action: durandal11707 finally something recognisable can be decoded
[22:41] <ubitux> i get a lot of "BUG in FFmpeg, draw_slice called with NULL pointer!" when trying to seek with MPlayer in matrixbench
[22:42] <ubitux> any idea if it's actually a ffmpeg bug?
[22:44] <wm4> mplayer has incredibly broken use of ffmpeg by now
[22:44] <ubitux> since TEP?
[22:44] <wm4> even before that
[22:45] <wm4> you can try to disable that code, maybe -noslices or something? (I don't remember much, I deleted all that stuff)
[22:45] <ubitux> well, it works
[22:45] <ubitux> i just get my stderr flood with that red message
[22:46] <ubitux> i'm just wondering if the problem really is in ffmpeg
[22:46] <ubitux> if yes then we should do something about it
[22:46] <ubitux> otherwise, that message should be dropped :p
[22:46] <wm4> it sounds like it uses the slices drawing callback
[22:46] <wm4> which probably nothing else but mplayer does
[22:48] <ubitux> mplayer2 & mpv dropped that code?
[22:48] <wm4> mpv did, mplayer2 has an old broken version of that code, but which is probably inactive in almost all cases
[22:49] <wm4> the only thing the slices callback is still used in mpv/mplayer2 is vdpau hw decoding
[23:24] <cone-565> ffmpeg.git 03Clément BSsch 07master:b77e58aad331: lavfi/pp: set out frame size.
[00:00] --- Sun Jun  9 2013