[Ffmpeg-devel-irc] ffmpeg-devel.log.20130408

[burek]

[00:24] <ubitux> michaelni: i've bisected the ffplay issue
[00:24] <ubitux> i found two potential problems
[00:24] <ubitux> 612d9cdbe90a9cb0319b093efa54efa71bb1e4c4 and de81d8da4750c812ad7288e39c7fb410f9e0f0f6
[00:28] <ubitux> actually de81d8da4750c812ad7288e39c7fb410f9e0f0f6 looks like the real problem
[00:29] <ubitux> but that's not enough
[00:31] <ubitux> reverting it restore playback for me with -f lavfi mandelbrot, but it's still laggy
[00:32] <durandal_1707> you mean: its slow?, or it does not drop frames as it should
[00:33] <ubitux> it's framedropping
[00:33] <ubitux> like low fps
[00:33] <ubitux> (and "large" increments)
[00:34] <ubitux> gonna bisect properly now
[00:37] <ubitux> durandal_1707: how is the telecine filter going?
[00:38] <durandal_1707> ubitux: only pts thing, i dunno what is best way to fix it
[00:39] <durandal_1707> add int64_t pts and increase it each output frame
[00:39] <durandal_1707> and set outlink time_base
[00:39] <durandal_1707> but how that will work with seeking?
[00:41] <ubitux> it won't
[00:42] <ubitux> as simply as that :p
[00:42] <ubitux> just like any filter relying by some mean on the frame count
[00:42] <durandal_1707> so what is proper solution?
[00:42] <ubitux> do it anyway
[00:43] <ubitux> maybe introduce a non-seekable flag to notify it's not supported
[00:43] <ubitux> or inject frame id in the filtergraph (good luck with this)
[00:44] <durandal_1707> i could do what mp=telecine do, so using setpts filter becomes mandatory
[00:47] <ubitux> what does it do?
[00:47] <ubitux> and what setpts will be able to do?
[00:48] <durandal_1707> ubitux: without it mp=telecine is just setting AV_NOPTS to every frame
[00:48] <ubitux> if you rely on setpts, you'll likely use 'N', and that would just move the problem around
[00:51] <durandal_1707> that is obvious ....
[01:05] <Daemon404> http://git.videolan.org/?p=ffmpeg.git;a=commitdiff;h=HEAD;hp=4bcb75cc48be18341919f86ce26fa1f4db624f67
[01:05] <Daemon404> ^ fuck you durandal_1707 
[01:07] <durandal_1707> and that he goes out....
[01:07] <durandal_1707> what was actuall problem?
[01:09] <durandal_1707> perhaps he wanted that i send another version and than i get his explicit approval?
[01:18] <ubitux> durandal_1707: i think he wanted you to not ignore the problem he was pointed out
[01:18] <durandal_1707> ubitux: but i did not ignored it, i already had same line, that michaelni sent
[01:19] <durandal_1707> well it could appear that i did ingored it
[01:27] <ubitux> michaelni: ok so 0d6e5a17124d0452c3fb6482a3c6c9150ae9af68 introduced a slowdown
[01:28] <ubitux> significant enough here to produce some frame drop
[01:28] <ubitux> and then de81d8da4750c812ad7288e39c7fb410f9e0f0f6 broke the frame-dropping playback 
[01:28] <kierank> durandal_1707: you ignored his review totally
[01:29] <kierank> his response is totally justified
[01:29] <kierank> the bug is still present
[01:31] <durandal_1707> kierank: really? that whole code is buggy
[01:31] <kierank> and you added another bug
[01:31] <kierank> congratulations
[01:34] <durandal_1707> kierank: could you give some correct code so you can elighten me?
[01:47] <kierank> durandal_1707: 
[01:47] <kierank> 00:42:54 <Daemon404> it stool looks quite possible for frame_end - frame to call outside teh array itself.
[01:47] <kierank> 00:43:09 <Daemon404> still*
[01:47] <kierank> 00:43:46 <Daemon404> consider whe segment is (1<<31)-1 aand count and skip ar aways 255
[01:47] <kierank> 00:43:51 <Daemon404> always*
[02:37] <FanGrille> bt-e
[02:37] <cone-665> ffmpeg.git 03Christophe Gisquet 07master:e2946e5c3482: x86: sbrdsp: implement SSE qmf_deint_bfly
[02:41] <durandal_1707> michaelni: so now mmx stuff works with mp?
[02:42] <michaelni> it should
[03:55] <cone-665> ffmpeg.git 03James Almer 07release/0.10:0b8198346150: lavc/bink: Chech for malloc failure
[07:39] <oneal> join #webkit 
[10:05] <ubitux> michaelni: i get 7 ’ 35 fps with this patch http://pastie.org/7366993 with the default configuration
[10:13] <ubitux> oups patch is broken
[10:13] <ubitux> http://pastie.org/7367084 might be better
[10:19] <ubitux> well, patch sent
[10:31] <mateo`> Tjoppen: any final thoughts about the mxfdec patch ?
[10:34] <Tjoppen> lgtm
[11:04] <cone-840> ffmpeg.git 03Paul B Mahol 07master:1f97dfb77f27: lavfi/blend,noise: remove unedeed av_opt_free() call
[11:04] <cone-840> ffmpeg.git 03Paul B Mahol 07master:018cc6f0266a: lavfi/histogram: use standard options parsing
[11:20] <mateo`> michaelni: you can merge https://github.com/mbouron/FFmpeg.git (mxfdec: handle identification metadata)
[11:37] <durandal_1707> g+ become spammers heaven
[11:44] <av500> ?
[11:45] <durandal_1707> there is buch of different users that post same thing
[11:45] <durandal_1707> which is just an advert
[11:46] <av500> in your circles?
[11:46] <durandal_1707> no, when searching
[11:46] <av500> ah
[11:49] <durandal_1707> though to be honest i don't have you in my circles any more (but i read your posts sometimes)
[11:51] <av500> durandal_1707: I am not offended :)
[12:23] <cone-840> ffmpeg.git 03Martin Storsjö 07master:aa8d89536d35: bfin: Don't use the vp3 idct functions if bitexact behaviour is expected
[12:23] <cone-840> ffmpeg.git 03Michael Niedermayer 07master:1090f69386ab: Merge commit 'aa8d89536d35af0a0c8d8bac2b452ffe7b82cae5'
[12:45] <cone-840> ffmpeg.git 03Martin Storsjö 07master:0f5984570890: bfin: Make vp3 functions static
[12:45] <cone-840> ffmpeg.git 03Ronald S. Bultje 07master:610b18e2e3d8: x86: qpel: Move fullpel and l2 functions to a separate file
[12:45] <cone-840> ffmpeg.git 03Michael Niedermayer 07master:5f49e92eb934: Merge commit '610b18e2e3d8ef5eca3e78f33a0625689b8d2bb9'
[12:52] <cone-840> ffmpeg.git 03Ronald S. Bultje 07master:18df366a1857: vp3: Embed idct_permutation array directly in VP3DecoderContext
[12:52] <cone-840> ffmpeg.git 03Martin Storsjö 07master:9e0f14f16cfc: lavc: Make pointers to ff_cropTbl const
[12:52] <cone-840> ffmpeg.git 03Michael Niedermayer 07master:024a5f72eec5: Merge commit '9e0f14f16cfc9456a691655fda7d01090bffe47e'
[12:59] <cone-840> ffmpeg.git 03Ronald S. Bultje 07master:d2ec6ea6c6c0: lavc: Move ff_cropTbl and ff_zigzag_direct from dsputil to mathtables
[12:59] <cone-840> ffmpeg.git 03Michael Niedermayer 07master:0138fe56563b: Merge remote-tracking branch 'qatar/master'
[13:01] <michaelni> mateo`, which branch should i merge ?
[13:23] <mateo`> michaelni: i pushed the commit on master
[13:55] <cone-840> ffmpeg.git 03Michael Niedermayer 07master:ad6a50c0ed39: avformat: add writeout_count statistic
[13:55] <cone-840> ffmpeg.git 03Michael Niedermayer 07master:a286940da148: mux: also call avio_flush() when flushing the muxer
[13:56] <cone-840> ffmpeg.git 03Michael Niedermayer 07master:7cfaf727a195: flvenc: dont call avio_flush() unconditional
[14:00] <durandal_1707> michaelni: some members in AVIOContext have wrong identation
[14:08] <cone-840> ffmpeg.git 03Matthieu Bouron 07master:25b143aa7bd6: lavf/mxfdec: handle identification metadata
[14:08] <cone-840> ffmpeg.git 03Michael Niedermayer 07master:8a97f6acc889: Merge remote-tracking branch 'mbouron/master'
[14:08] <michaelni> mateo`, merged
[14:11] <cone-840> ffmpeg.git 03Paul B Mahol 07master:9fce2b8854c9: lavfi/il: use standard option parsing
[14:16] <cone-840> ffmpeg.git 03Michael Niedermayer 07master:c6ae7f64a09e: avio.h: fix indention
[14:19] <mateo`> michaelni: thx !
[16:17] <nevcairiel> something broke mxf fate tests
[16:17] <nevcairiel> at least on msvc
[17:12] <michaelni> nevcairiel, cant reproduce in mingw64 but i suspect mateo` s change could be related as it was what was changed last in mxf
[17:13] <nevcairiel> yeah my mingw test is still fine, its only the msvc ones that broke
[17:22] <Zeranoe> i just ran into this bug: https://ffmpeg.org/trac/ffmpeg/ticket/2417 is the attached patch (https://ffmpeg.org/trac/ffmpeg/attachment/ticket/2417/patchoddsize.diff) a fix?
[17:42] <michaelni> Zeranoe, possibly, it needs to be reviewed & tested
[17:57] <cone-840> ffmpeg.git 03Michael Niedermayer 07master:0678c388bac7: rtsp: add option to set the socket timeout of the lower protocol.
[18:11] <mateo`> nevcairiel: the diff looks like ffmpeg crashed ?
[18:18] <mateo`> Zeranoe: is this patch posted on the ml ? (it seems ok).
[18:20] <Zeranoe> mateo`: I'm not sure if it has been or not, I just stumbled on the bug report after I got the error.
[18:23] <schtinky> hi folks. quick question: Is "ffmpeg -f v4l2 -i /digitaldvbdevice0" supposed to work as "ffmpeg -i /dev/analogdevice0" does? I get unsupported operation and have been using "gnutv | ffmpeg" as a workaround
[18:24] <schtinky> just wondering if the devs can shed some light on the internal reason analog and digital are different animals in this respect
[18:24] <schtinky> (if they are, that is)
[18:26] <mateo`> Zeranoe: http://pastebin.com/hJLJU6mW seems more correct to me
[18:30] <michaelni> or next_tag_ofs = avio_tell(pb) + size + (size&1)
[18:30] <mateo`> michaelni: even better :)
[18:33] <mateo`> running fate and i'll send the patch
[19:05] <jojva> H-2
[19:07] <mateo`> michaelni: i updated the patch, so it will only affect next_tag_ofs before seeking
[19:09] <mateo`> next_tag_ofs++ would have been better than += 1 :/
[19:14] <ubitux> python got you
[19:14] Action: michaelni doesnt belive in a python eating a human
[19:15] <ubitux> it's killing him for pleasure only
[19:16] <ubitux> mateo`: you can be evil and do next_tag_ofs += (next_tag_ofs < INT64_MAX && next_tag_ofs & 1);
[19:16] Action: michaelni doesnt belive in a python killing for pleassure, they kill to eat
[19:17] Action: durandal_1707 you never know
[19:17] <mateo`> ubitux: love this one :)
[19:19] <mateo`> maybe the find_tag() should also be updated to skip padding byte
[19:19] <ubitux> michaelni: hard believer, aren't you? (bonus: https://www.youtube.com/watch?v=DDG8xqz7BIk)
[19:22] <ubitux> btw, accepted gsoc projects will be available soon here: http://www.google-melange.com/gsoc/accepted_orgs/google/gsoc2013
[19:23] <ubitux> (time is in UTC)
[19:23] <ubitux> so in about 1h½ afaict
[19:24] <saste> ubitux: splice is problematic
[19:24] <saste> that's why i'm going to keep enable for overlay
[19:25] <saste> (but as a separate patch)
[19:25] <ubitux> yeah i thought about it a little as well
[19:25] <saste> unless you can find a clean alternative solution
[19:25] <saste> splice can work only under certain assumption
[19:25] <ubitux> a splice filter is going to be really troublesome and not reliable
[19:25] <saste> so while it is in general useful, it should be used with care
[19:27] <ubitux> or we should go back to another solution :p
[19:28] <ubitux> saste: btw, the avfilter_graph_parse() prototype mismatch with the fork should be fixed
[19:28] <saste> still didn't look at that
[19:28] <saste> are you aware of a sane solution?
[19:31] <ubitux> overlay{between(n,30,40)}=... ; add a filter_disabled_frame() callback for special filters (multiple output, n counter, etc); default is to make the avfilter internals call ff_filter_frame() on the first outlink
[19:34] <ubitux> maybe you don't find this sane :)
[19:38] <ubitux> note that such syntax makes timeline creation somehow easy
[19:40] <saste> ubitux: you can create a common context, shared amongst filters, and set enable there
[19:40] <saste> also i'm not sure it is a good idea
[19:40] <saste> what is "enable" for overlay?
[19:40] <saste> what is "enable" for select?
[19:40] <saste> what is "enable" for drawtext?
[19:40] <ubitux> exceptions defined in filter_disabled_frame() callback
[19:41] <ubitux> enable means active filter, otherwise "context" passthrough
[19:42] <saste> anyway i see we may need more shared options
[19:42] <saste> for example to set a unique id for a filter
[19:42] <saste> that or we need to define more hacks to the syntax
[19:43] <saste> other common options: per filter loglevel 
[19:43] <saste> i don't know if we already have that for codec/formats etc.
[19:44] <ubitux> i can try to write a PoC for the lavfi enable thing
[19:44] <ubitux> do you think it's worth the effort or you want to try something else?
[19:45] <ubitux> (or you think it's definitely a bad idea)
[19:46] <saste> ubitux, what about shared options?
[19:46] <saste> that has its load of problems as well
[19:46] <saste> there is the usual namespace cluttering issue
[19:46] <saste> if we define common options, then we have to make sure that they don't clash with the ones defined in the private contexts
[19:47] <ubitux> mmh
[19:47] <ubitux> indeed that's not a bad idea
[19:47] <saste> also there is the possibility of more conflicts with l
[19:48] <ubitux> but i wonder how you will implement them
[19:48] <saste> in case they reinvent the wheel with a different color
[19:48] <saste> how me?
[19:48] <saste> no, i wonder how *you* will implement it ;-)
[19:48] <ubitux> :D
[19:48] <ubitux> hey :(
[19:48] <saste> anyway, about the enable in overlay is around 20 lines of code
[19:49] <ubitux> i'm ok with the enable overlay option
[19:49] <saste> we can get rid of them if we implement a general solution, but i don't know how much time it will take
[19:49] <ubitux> but please split it :p
[19:49] <ubitux> (from the x/y patch)
[19:49] <saste> already done, will send soon
[20:06] <saste> ffplay -f lavfi "testsrc=s=320x240[l]; life, scale=320x240[r]; [l][r] splice" -noframedrop
[20:06] <saste> cool stuff :-)
[20:08] <durandal_1707> that does what?
[20:08] <saste> durandal_1707, guess it
[20:09] <durandal_1707> it splices in random way frames [l] & [r]
[20:09] <saste> uhm why randomly?
[20:09] <saste> no the main problem with splice is that it tries to sort output frames
[20:09] <saste> depending on the input timestamps
[20:10] <durandal_1707> should it sort in order user specified?
[20:10] <saste> thus it needs to have all input queues filled, and can't work if a queue is not filled in time
[20:11] Action: durandal_1707 wonders why such esential filter is not already available
[20:23] <cone-840> ffmpeg.git 03Matthieu Bouron 07master:1f2ce32825f7: lavf/wavdec: skip padding byte
[20:30] <llogan> negative aspect: http://alan.appredeemdev.com/wtf.mov
[20:30] <llogan> any hints?
[20:30] <durandal_1707> it could be nice if range_{min,max} is actually used for AV_OPT_TYPE_STRING
[20:30] <cone-840> ffmpeg.git 03Michael Niedermayer 07release/1.0:aef81fa75df2: vf_idet: fix free after use
[20:37] <cone-840> ffmpeg.git 03Clément BSsch 07master:51bcd5cd65df: lavfi/mandelbrot: fix speedloss with default config after morphing introduction.
[20:41] <cone-840> ffmpeg.git 03Clément BSsch 07master:e82f562fa5ef: lavfi/mandelbrot: reindent after 51bcd5cd.
[20:44] Action: durandal_1707 brace yourself
[20:45] <durandal_1707> i want to K&R vsrc_mandelbrot
[20:47] <llogan> make sure to break it as per tradition.
[21:02] <nevcairiel> a short list so far =p
[21:02] Action: durandal_1707 really do not want to waste night/day starring at that page
[21:06] <llogan> Thank you for submitting "FFmpeg" organization application to Google Summer of Code 2013. Unfortunately, we were unable to accept your organization's application at this time.
[21:07] <jojva> damn
[21:07] <llogan> yeah
[21:07] <jojva> :(
[21:07] <nevcairiel> do they at least tell you why?
[21:08] <llogan> no. only, "If you would like some general feedback on why your organization was not accepted, please consider attending the IRC meeting in #gsoc on Freenode on Friday, April 19, 2013 at 16:00 UTC."
[21:08] <jojva> I don't understand... The list of projects is fine...
[21:09] <llogan> yes, it was an improvement over last years in my opinion
[21:09] <jojva> still because of the fork situation maybe
[21:09] <llogan> i'm not sure if they got accepted.
[21:09] <llogan> i assume gsoc will simply say, "we are trying to make room for new organizations".
[21:10] <jojva> mvc decoder won't be for this summer i guess
[21:11] <llogan> if anyone is in their channel i'd be interested if they got accepted/rejected
[21:11] <nevcairiel> no word yet
[21:11] <nevcairiel> i dont know who is their admin and if he is even there
[21:12] <nevcairiel> but also not int he project list
[21:12] <jojva> they're not even talking about it
[21:12] <llogan> i'll mentioned this on the list after i go move some furniture...
[21:13] <llogan> *mention
[21:13] <ungureanuvladvic> so no ffmpeg in gsoc this year :(
[21:13] <durandal_1707> llogan: mailing list?
[21:13] <jojva> i just asked libav
[21:15] <jojva> it's negative on their side, as well as vlc
[21:17] <kurosu> michaelni / Skyler_ / pengvado: what about adding xora and anda (or whatever) resolving to xorps/pxor and andps/pand? There are probably others to consider
[21:17] <kurosu> (might also ask this in #x264dev)
[21:20] <mark4o> Is http://www.google-melange.com/gsoc/accepted_orgs/google/gsoc2013 the complete list?  only 25 projects there
[21:21] <durandal_1707> that is filled slowly, rejected orgs get mail
[21:26] <mark4o> I see, now they have 43 listed
[21:27] <llogan> durandal_1707: yes
[21:27] <jojva> Do you think there's something to do with MPEG-LA/Google relationships ?
[21:28] <nevcairiel> nah, dont believe any troll you read
[21:29] <gnafu> Well, at least LibreOffice got excepted: they have a "VLC integration" project listed as a possibility ;D.
[21:42] <michaelni> jojva, ungureanuvladvic, others, if anyone wants to work on their gsoc projects anyway, that should be no problem, only problem is there wont be any payment ...
[21:43] <michaelni> unless some rich madman or company comes forth and offers money ...
[21:43] <jojva> michaelni, i'm still very enthusiastic with doing mvc stuff, but it won't happen during this summer, or very slowly
[21:43] <durandal11707> what about donations? what about ff***h?
[21:44] <jojva> ff***h ?
[21:45] <durandal11707> secret agency
[21:50] <ubitux> so we're not accepted as gsoc?
[21:50] <nevcairiel> no
[21:50] <ubitux> :/
[21:51] <nevcairiel> You'll have to fix the subtitles yourself
[21:51] <ubitux> :(
[21:53] <jojva> life is mean
[21:54] <llogan> who is that guy? ^
[21:55] <durandal11707> ask him
[21:55] <durandal11707> lavfi-pp fails here
[21:58] <llogan> phpmyadmin got accepted with an ideas page like this: http://wiki.phpmyadmin.net/pma/GSoC_2013_Ideas_List
[21:58] <cone-840> ffmpeg.git 03Paul B Mahol 07master:ed8373e7db14: lavfi: always check return value of ff_get_{audio,video}_buffer()
[21:59] <jojva> llogan: it's not about the ideas page, the one you had was great. 
[22:00] <durandal11707> its about that airplane
[22:00] <jojva> lo
[22:00] <jojva> lol*
[22:00] <jojva> i can confirm, durandal11707 did say the plane was childish
[22:01] <llogan> you should have told me
[22:02] <nevcairiel> i like the plane :P
[22:02] Action: ubitux imagine the gsoc guys telling us "well your page is great, but we couldn't deal with that logo"
[22:02] <llogan> i tohught it was good google bait
[22:03] <jojva> There is one tag "ffmpeg" in XBMC that doesn't have anything to do there (i checked their ideas page), what about that ?
[22:04] <nevcairiel> its an essential part of xbmc, may just have put it there for good measure
[22:04] <durandal11707> llogan: tell lithosza to that the only way his hq audio wiki page will get accepted is if it send it to hydrogenaudio guys for review
[22:04] <llogan> go for it
[22:05] <llogan> good number of projects i've never heard of
[22:07] <llogan> jojva: GSoC claims the ideas page is the most important part
[22:08] <ubitux> let's put some <banner> and <blink> next time
[22:10] <llogan> ubitux: or "A better VP9 encoder".
[22:11] <nevcairiel> dont offend the google encoder!
[22:11] <durandal11707> there is list of ffmpeg's libavcodec decoders chromium uses
[22:14] <ubitux> i guess i'll have to leave out libavfilter a little and go back to the evil subtitles now...
[22:14] <ubitux> life is a bitch
[22:14] <llogan> i'll make it up with some ffmpeg stickers
[22:16] <durandal11707> priority list:
[22:17] <llogan> Cinepak ASM
[22:17] <ubitux> :D
[22:18] <durandal11707> 0: libavfilter & libavformat must become popular as libavcodec
[22:18] <ubitux> (lol @Derek discussing FFmpeg project on the fork devel channel)
[22:18] <durandal11707> he also commented vivtc patch on libav ml by mistake
[22:18] <ubitux> durandal11707: i wonder how that was possible, but it was funny :))
[22:33] <kierank> llogan: xvp9
[22:34] Action: gnafu still wants xvp8 :-(.
[22:44] <durandal11707> michaelni: 54524518 decicycles in mpnoise, 512 runs, 0 skips vs 54996903 decicycles in noise, 512 runs, 0 skips
[22:44] <durandal11707> probably because of for loop
[22:45] <durandal11707> i think such difference is irrelevant
[22:46] <michaelni> durandal11707, the removial of the low quality mode should also have produced a slight slight speedloss so yes should be irrelevant
[22:53] <durandal11707> michaelni: can sse(2) & avx version be done in inline?
[22:54] <cone-840> ffmpeg.git 03Paul B Mahol 07master:693424441450: lavfi/noise: copy inline asm from mp=noise
[22:55] <michaelni> durandal11707, should be no problem
[23:06] <cone-840> ffmpeg.git 03Paul B Mahol 07master:9dc88ac3676e: lavfi: remove noise libmpcodecs wrapper
[23:10] <durandal11707> ubitux: what mp filters you gonna remove?
[23:12] <ubitux> ideally, mp={detc,divtc,filmdint,pullup,ivtc}
[23:12] <ubitux> but i expect some (justified?) resistance for pullup filter
[23:12] <durandal11707> divVerent said pullup and filmdint are good
[23:12] <durandal11707> others are crap
[23:13] <ubitux> filmdint isn't supported anyway by mp afaik
[23:13] <ubitux> pullup is likely a good enough real-time ivtc filter
[23:13] <ubitux> but i won't port it
[23:16] <durandal11707> wtf is softpulldown? another teleciner?
[23:16] <ubitux> no idea
[23:16] <durandal11707> mpeg2 soft 3:2 pulldown"
[23:16] <ubitux> yeah it's a teleciner
[23:17] <ubitux> it uses the tff flag in mpeg2 streams
[23:17] <ubitux> which is kind of a nice feature
[23:17] <durandal11707> does it actually work with ffmpeg?
[23:17] <ubitux> dunno
[23:18] <durandal11707> (i could add this to telecine)
[23:18] <ubitux> but it could be added as an option in telecine
[23:18] <ubitux> durandal11707: btw, ideally dvd should contain progressive material with the video stream flagged where the player is supposed to interlace itself when the tv is analog; but having it progressive allow to render on digital as well
[23:18] <ubitux> which is nice.
[23:20] <durandal11707> analog tv is dead
[23:21] <durandal11707> and i wonder why people f* still use this
[23:22] <ubitux> broadcasting doesn't consider all analog tv dead, so they continue to stream daemons all over the world
[23:22] <ubitux> and we're here to help them
[23:25] <kierank> durandal11707: analogue tv is dead but there's hundreds of years of old content
[23:30] <ubitux> kierank: we're talking about creating some right now
[23:31] <ubitux> basically, if it's relevant to have a telecining filter using mpeg rff
[23:31] <kierank> i am not 100% sure of the use case because everything would be hard telecined
[23:34] <ubitux> arg fuck
[23:34] <ubitux> using -codec as input option sets codec only for mapped streams
[23:35] <ubitux> :(
[23:42] <saste> too bad
[23:46] <durandal11707> saste: cheer up
[23:49] <ubitux> !@#$ subtitles
[23:49] <ubitux> :(
[23:49] <ubitux> the ass/ssa hell drives me crazy
[23:53] <Compn> it drove previous sub dev mad too
[23:53] <Compn> dont let it! :)
[23:54] <saste> durandal11707, always look on the bright side of life
[23:54] <saste> http://www.youtube.com/watch?v=L2Wx230gYJw
[23:55] <saste> on the bright side, more time for us, for coding and for having fun
[23:56] <saste> then we should consider to run our promotion program and enhance our fundraising capabilities (which are pretty non-existent right now)
[23:57] <saste> Compn, who was the previous sub dev?
[23:57] Action: saste saste waits for ubitux to review overlay x/y expressions
[23:58] Action: ubitux ubitux too
[23:58] <ubitux> saste: how long are you staying?
[23:58] <ubitux> (so i can continue to fight a little more with subtitles tonight before review)
[23:59] <saste> 1 hour
[23:59] <ubitux> mmh ok
[23:59] <ubitux> subtitles for 15 minutes then
[00:00] --- Tue Apr  9 2013