[Ffmpeg-devel-irc] ffmpeg-devel.log.20161214

burek burek021 at gmail.com
Thu Dec 15 03:05:04 EET 2016


[00:58:53 CET] <cone-555> ffmpeg 03Muhammad Faiz 07master:2b0112d42d4d: swresample/resample: remove swri_resample function
[01:10:21 CET] <cone-555> ffmpeg 03Andreas Cadhalpun 07master:ad5807f8aa88: avformat: fix overflows during bit rate calculation
[03:22:07 CET] <cone-555> ffmpeg 03Martin Vignali 07master:90af962911bb: libavcodec : add decoder for Photoshop PSD image files
[03:22:08 CET] <cone-555> ffmpeg 03Martin Vignali 07master:a534862b278a: libavformat : add Photoshop PSD demuxer.
[03:44:37 CET] <KGB> [13FFV1] 15michaelni pushed 1 new commit to 06master: 02https://git.io/v11SA
[03:44:37 CET] <KGB> 13FFV1/06master 14f13bedd 15dericed: left-align some table columns
[04:56:53 CET] <KGB> [13FFV1] 15dericed opened pull request #29: add subsections (06master...06WIP-adding-narratives) 02https://git.io/v1175
[08:34:27 CET] <fy> I am trying the skip_initial_bytes option to skip the first 8 bytes of my video file, which is my own file header of size 8. but it seems not working.
[08:35:08 CET] <fy> sorry for posting in the wrong channel...
[11:34:30 CET] <michaelni> durandal_170, gbrap16be/le pixfmts-scale fails on qemu-mips it seems http://fate.ffmpeg.org/report.cgi?time=20161214093528&slot=mips-ubuntu-qemu-gcc-4.4
[11:35:46 CET] <durandal_170> michaelni: is that only big endian instance?
[11:37:29 CET] <michaelni> durandal_170, it does seem to work on loongson (mips64) http://fate.ffmpeg.org/report.cgi?time=20161214011218&slot=loongson-fedora-gcc-4.8
[11:37:36 CET] <michaelni> so i dont know what is happening
[11:39:03 CET] <michaelni> its also failing on http://fate.ffmpeg.org/report.cgi?time=20161214060811&slot=ppc64be-RHEL7.0-gcc-4.8.5-ibmcrl
[11:51:03 CET] <cone-706> ffmpeg 03Michael Niedermayer 07master:457e93391964: Bump for psd demuxer and decoder
[13:39:45 CET] <michaelni> Timothy_Gu, some pages on fatebeta always seem to timeout for example: http://fatebeta.ffmpeg.org/report/sparc-solaris-32bit-gcc-5.2/20161214061859
[14:41:01 CET] <durandal_170> is there gui for gdb of some kind to step through program with source code?
[14:42:18 CET] <BtbN> Any IDE i guess
[14:46:35 CET] <robUx4> qtcreator does the job
[15:08:20 CET] <J_Darnley> durandal_170: Do you want more than GDB's own tui?
[15:10:50 CET] <durandal_170> J_Darnley: KDbg using right now, seems fine
[17:34:14 CET] <cone-706> ffmpeg 03Diego Biurrun 07master:2d7aadf3d95c: Add missing #includes for standalone spherical-information-related headers
[19:42:23 CET] <cone-706> ffmpeg 03Andrey Utkin 07master:c4afd4abce58: doc/filters: drawtext: add example of printing texts on same baseline
[20:09:49 CET] <durandal_170> ubitux: some guy is trying to use 128 level Lut with ffmpeg
[20:25:06 CET] <ubitux> lut3d?
[20:27:08 CET] <durandal_170> yes
[20:27:48 CET] <ubitux> it's going to be a bit large, but that might be doable
[20:28:28 CET] <ubitux> does it really make a meaningful difference to change the MAX_LEVEL to 128?
[20:28:35 CET] <ubitux> (in the output)
[20:29:14 CET] <ubitux> 128 will create a 25M malloc afaict
[20:29:45 CET] <ubitux> we can dynamically allocate them though
[20:30:48 CET] <ubitux> (currently it's about 3M, which is already quite large)
[20:31:18 CET] <llogan> michaelni: how can i (user "llogan") get permissions to set a patch as "Accepted" in patchwork? apparently I forgot to use one of the magic words to auto-deal-with-it.
[20:35:04 CET] <michaelni> llogan, you should have the power now
[20:38:17 CET] <llogan> michaelni: thanks. https://www.youtube.com/watch?v=-dJolYw8tnk
[20:39:35 CET] <llogan> has patchwork been helpful for you?
[20:40:53 CET] <michaelni> its usefull to find (old) patches that have not been applied
[20:41:46 CET] <michaelni> or not reviewed
[20:43:24 CET] <michaelni> its less than fully automatic though, it needs someone to update patches which where rejected or otherwise changed their status in ways that cant be automatcially detected
[20:44:55 CET] <michaelni> ive also written a dumb script to do a bit more automatic updating:  https://github.com/michaelni/patchwork-update-bot
[20:45:17 CET] <michaelni> if you have ideas to improve this so more is auto updated, pull req or patch is welcome
[20:51:43 CET] <Compn> llogan : person asked webmaster@ about getting a samples repo on hd. you arent interested in that are you? i ask because you are one of the few usa devels...
[20:51:54 CET] <Compn> and the person asking was in n carolina i think
[21:02:12 CET] <durandal_1707> ubitux: simple increasing level doesn't help, there's some other issue with file
[21:11:21 CET] <sunny26_> <c_14> Are you present on this chat list also ?
[21:12:02 CET] <sunny26_> <durandal_1707> told me to post my queries here.
[21:13:45 CET] <durandal_1707> sunny26_: nicks are typed without <> 
[21:15:11 CET] <c_14> sunny26_: yeah, I'm here as well
[21:15:40 CET] <sunny26_> Hey man thanks, for so much help. make fate is now doing its work
[21:16:04 CET] <sunny26_> If any error comes will let you know. Thank you so much. 
[21:17:05 CET] <sunny26_> No body at my company is enthu to post updates to ffmpeg. Its me who is working overnight so that I can atleast contribute to this open source culture
[21:17:23 CET] <sunny26_> Thanks c_14 ....
[21:35:44 CET] <llogan> Compn: aren't you in US too? do how know how big samples is? internet service sucks in Alaska. Expensive and relatively slow.
[21:52:42 CET] <wm4> so there is this ffmpeg.c issue that using a hwaccel will trigger a format switch on the first frame
[21:52:52 CET] <wm4> avconv.c apparently had this fixed
[21:52:59 CET] <wm4> was the merged yet?
[21:53:25 CET] <nevcairiel> i skipped those changes in the merge because it requires huge refactoring of ffmpeg.c, outside of merge-size changes
[21:53:32 CET] <wm4> heh, ok
[21:56:30 CET] <nevcairiel> i tried to get it to work for a while too, but it quickly turned out that our filtering infrastructure is too different to just adapt it in the scope of a merge
[21:56:40 CET] <nevcairiel> filtering is just way different for us
[22:41:43 CET] <Timothy_Gu> michaelni: those pages have a lot errors, and to fix it would mean that I'll have to rewrite some parts of the parsing routine
[22:43:17 CET] <Timothy_Gu> Compn: I'm located in California if that helps
[22:51:50 CET] <cone-706> ffmpeg 03Paul B Mahol 07master:458fbee22154: avfilter/af_sofalizer: warn if user gives unknown channel names
[22:51:51 CET] <cone-706> ffmpeg 03Paul B Mahol 07master:e54b61a9ce19: doc/filters: fix channel names in sofalizer example
[22:55:13 CET] <Compn> Timothy_Gu : oh , nice to know. are you interested in making copies of the samples repo on hd and mailing them to people ?
[22:56:00 CET] <Compn> knowing full well that the repo contains clips of a pornographic nature and that we do not have the specific age and verified ID of the actressed to ensure full us title code compliance?
[22:57:13 CET] <Compn> llogan : yeah, but i'm not going to do it.
[22:57:36 CET] <Compn> llogan : ~100gb  last i remembered, i havent  checked as of late
[22:58:05 CET] <BtbN> 100GB still needs sending HDDs around?
[22:58:22 CET] <nevcairiel> some parts of the world are rather backwards
[22:58:30 CET] <nevcairiel> more interesting would be w hy people want the samples archive
[22:58:32 CET] <BtbN> But there are SD cards big enough for that.
[22:58:33 CET] <Compn> usually people just say that because they are scared of downloading or scared we wont give them permission
[22:59:05 CET] <Compn> i asked why he wants samples (when i gave him permission to download them), did not get reply back
[22:59:24 CET] <Compn> previous people wanted the samples include western digital who tested their hardware media player with the samples
[22:59:37 CET] <Compn> just to make sure it didnt crash, because they were not interested in supporting everything haha
[23:00:40 CET] <Compn> some guy @nla.gov.au asked for permission for the samples last year
[23:00:56 CET] <Compn> which is the national library of australia
[23:01:04 CET] <Compn> probably checking to see if they support all of our samples...
[23:01:09 CET] <Compn> for archiving formats
[23:01:32 CET] <nevcairiel> majority of the samples are butt-old formats without much relevance, or broken samples to some degree
[23:01:33 CET] <nevcairiel> so shrug
[23:02:13 CET] <Compn> heres an email we got this year, "I found the link, where there are lots of samples. I want to use them to test my transcoding program. It's inconvenient to download those samples by click each url link. Where and how to get those samples collection?"
[23:02:25 CET] <Compn> so this guy wanted to test his transcoding program
[23:02:44 CET] <JEEB> I think we have the way of getting those samples with rsync mentioned somewhere...
[23:02:56 CET] <JEEB> that person just didn't want to keep clicking http links I guess
[23:03:08 CET] <BtbN> Downloading 100GB isn't even that bad in terms of traffic cost anymore.
[23:03:20 CET] <durandal_1707> ban then from Internet nobody wants that ugly samples
[23:03:23 CET] <Compn> some software vendor wanted a copy of the samples repo. probably for testing youtube-like knockoff solution website
[23:05:19 CET] <Compn> guy at UTC aerospace systems wanted a copy of the repo too
[23:08:20 CET] <durandal_1707> Compn: who pays all that transactions?
[23:08:31 CET] <Compn> which transactions? 
[23:08:40 CET] <nevcairiel> if they want physical delivery, they get to pay that themself
[23:08:43 CET] <Compn> yes
[23:08:50 CET] <Compn> they offer to pay, but they want a price quote
[23:08:58 CET] <durandal_1707> server data transfer
[23:09:01 CET] <Compn> i havent found anyone who wants to do it , so i dont have price quote
[23:09:18 CET] <Compn> our host pays the bandwidth costs
[23:09:31 CET] <durandal_1707> price is 1000$
[23:09:45 CET] <Compn> you would do it for $1000 ? :)
[23:10:58 CET] <Compn> (its really not hard or a lot of work haha)
[23:11:03 CET] <durandal_1707> add more zeroes as you like
[23:11:08 CET] <BtbN> If it's a company or something, overcharging a bit and putting the excess money into the ffmpeg fund thing seems not too bad?
[23:11:22 CET] <Compn> sure
[23:11:27 CET] <Compn> i think most people just end up downloading it
[23:11:28 CET] <BtbN> But companies usually have good internet, so they can just download it.
[23:11:31 CET] <Compn> yea
[23:12:59 CET] <durandal_1707> Compn: why MPlayer development seems dead?
[23:13:50 CET] <rcombs> I mean, mpv is a thing
[23:15:40 CET] <Compn> durandal_1707 : developers got jobs, lives
[23:15:51 CET] <Compn> no one pays people to work on mplayer
[23:15:59 CET] <Compn> (except youtube pays someone to work on mencoder lol)
[23:16:10 CET] <Compn> but they dont share back with us
[23:16:37 CET] <durandal_1707> really? who?
[23:16:39 CET] <Compn> durandal_1707 : it could happen to any software project
[23:18:01 CET] <Compn> i dont know persons name . i just know that youtube used mencoder in the past for some binary codec loading
[23:18:18 CET] <Compn> so someone had to fix mencoder up a little...
[23:18:45 CET] <durandal_1707> **** ****
[23:18:59 CET] <Compn> durandal_1707 : how do you think some of these codecs are being decoded when there is no ffmpeg decoder ?
[23:19:52 CET] <Compn> nevcairiel : yes, the samples repo is mostly a museum piece at this moment in time
[23:20:12 CET] <durandal_1707> people should be shot immediately after attempting implementing codec
[23:20:13 CET] <Compn> and its kind of the only one , besides vlc's repo.
[23:20:44 CET] <Compn> durandal_1707 : how is pixlet ?
[23:20:52 CET] <Compn> durandal_1707 : will you work on vivo codecs next? 
[23:21:17 CET] <durandal_1707> pixlet is almost finished
[23:21:36 CET] <durandal_1707> vivo rip ;)
[23:22:02 CET] <Compn> i wish i could find vdo sample. all i came up were playlist files :\
[23:23:45 CET] <durandal_1707> there are other wavelets, besides qdmc need is too
[23:24:00 CET] <Compn> oh yea theres a bunch of wavelets
[23:24:24 CET] <Compn> i put a couple in mplayer binary codecs :)
[23:24:49 CET] <durandal_1707> I hope qdmc description on wiki is correct enough
[23:25:17 CET] <Compn> ancient quicktime format
[23:26:49 CET] <durandal_1707> I still haven't pushed bonk decoder
[23:27:18 CET] <durandal_1707> it's shitty like shorten
[23:27:58 CET] <Compn> is that like bink decoder ?
[23:28:16 CET] <Compn> oh no, lossless audio
[23:28:38 CET] <Compn> well, cant hurt to push it as experimental and let the users fix it :D
[23:28:54 CET] <Compn> assuming you fuzz it a little 
[23:29:21 CET] <Compn> better to push it than to let it rot
[23:29:31 CET] <llogan> Compn: i don't feel like being a samples mule either. I suggest our official policy is to only offer downloads.
[23:31:40 CET] <Compn> llogan : no problem at all. just figured i would ask around.
[23:32:33 CET] <Compn> i wonder if we should be doing anything since ffmpeg (or lavc) is now on billions of computers..
[23:33:02 CET] <Compn> such a weird thing to go from a hobby project to installed on all phones, tv and computers worldwide
[00:00:00 CET] --- Thu Dec 15 2016


More information about the Ffmpeg-devel-irc mailing list