[Ffmpeg-devel-irc] ffmpeg-devel.log.20130620

burek burek021 at gmail.com
Fri Jun 21 02:05:02 CEST 2013


[01:08] <Daemon404> thats a lot of attachments.
[03:08] <cone-902> ffmpeg.git 03Timothy Gu 07master:c16496c3770f: doc/decoders: Document libgsm decoder
[03:08] <cone-902> ffmpeg.git 03Andrey Utkin 07master:f9739a34ad53: Document "crypto" protocol
[09:52] <saste> uhm so, no more socis mentors?
[11:24] <cone-963> ffmpeg.git 03Alex Smith 07master:e0d6e2fc67b8: configure: Remove unneeded icl inline compatibility definition
[11:24] <cone-963> ffmpeg.git 03Michael Niedermayer 07master:99dc90ea59f6: Merge remote-tracking branch 'qatar/master'
[15:24] <cone-963> ffmpeg.git 03Andrey Utkin 07master:2d70f11ef5d4: Document "sctp" protocol
[15:24] <cone-963> ffmpeg.git 03Michael Niedermayer 07master:c605adbf562f: avformat/libmodplug: Reduce the probe score for small input
[15:26] <durandal_1707> ok, i think encoder is now ready for testing
[15:38] <BBB> Daemon404: spec is in the works
[16:47] <durandal_1707> michaelni: fine to rename files? to make merges "easier"?
[17:04] <michaelni> durandal_1707, sure
[17:05] <durandal_1707> michaelni: even if that makes merges harder?
[17:06] <michaelni> if someone volunteers to do the extra work
[17:07] <durandal_1707> i can't volunter 24/7/12/356
[17:07] <wm4> durandal_1707: lazy!
[17:10] <durandal_1707> wm4: no, i have real life
[19:47] <durandal_1707> michaelni: so what's best scaler quality wise?
[19:52] <michaelni> i dont really know, bicubic & lanczos are popular.
[20:13] <cbsrobot> durandal_1707: I read once that sinc is best for scaling down, mitchell for scaling up
[20:13] <ubitux> durandal_1707: that's a large encoder you ahve there :)
[20:21] <durandal_1707> ubitux: hmm. perhaps i should add more macros
[20:21] <ubitux> :)
[21:32] <Daemon404> " Thank you for your participation in the Coverity Scan program. We are pleased to announce that we have enhanced our service to address the feedback of our members. These enhancements enable you to: "
[21:32] <Daemon404> wtf?
[21:32] <Daemon404> i dont recall havign a coverity acct
[21:58] <cone-963> ffmpeg.git 03Michael Niedermayer 07master:03cd3bec6e66: avformat/iff: Byte seek is unsupported
[22:00] <Compn> Daemon404 : you never looked at ffmpeg in coverity ?
[22:00] <Daemon404> nope.
[22:00] <Compn> its possible that you requested it, and michael sent in your name, but they never got around to making you an account ?
[22:02] <Daemon404> Compn, i explicitly said i didnt
[22:02] <Daemon404> iirc.
[22:03] <Compn> ok then
[22:06] <wm4> Daemon404: so what exactly is that "removed" copyright in paul's patch about?
[22:08] <Daemon404> what do you mean
[22:09] <Daemon404> he moved code from wavpack.c, copyright by kostya, to a new file, with no copyright
[22:09] <wm4> ah
[22:12] <Compn> is no good to do that 
[22:12] <Compn> gotta watch it 
[22:13] <Compn> seen a lot of people make that mistake lately
[22:13] <wm4> might get sued by Libav otherwise
[22:15] <Compn> wm4 : its in our rules that every file has to have a copyright , diego did a lot of work to get that 
[22:15] <Compn> standardized
[22:17] <Compn> BBB : videolan wants h265 and vp9 decoders in ffmpeg :)
[22:22] <JEEB> smarter's HEVC decoder is pretty ready, just needs some finishing up and the OpenHEVC dudes to merge in their useful changes
[22:22] Action: michaelni also wants H265 & vp9 in ffmpeg
[22:22] <michaelni> decoders and encoder
[22:22] <michaelni> S
[22:22] <JEEB> the OpenHEVC folk are the worst part I guess
[22:22] <Daemon404> supposedly in aug BBB, myself, + others are workign on vp9
[22:22] <JEEB> since they mostly just hack stuff up on their C++ stuff
[22:23] <JEEB> and don't even care if their hacked up libavcodec works with upstream
[22:23] <JEEB> elenril ended up merging their stuff the last time
[22:23] <JEEB> and now they have written even more
[22:26] <smarter> I know I keep saying that, but openhevc stuff should get cleaned up/merged Soon//"  :)
[22:26] <smarter> they just got tiles working which is nice
[22:27] <JEEB> it's not your problem, really :s
[22:27] <JEEB> it's theirs
[22:27] <JEEB> they have no respect for the upstream
[22:27] <JEEB> if they had, the commits could be merge as-is
[22:27] <smarter> they do, but it's not their priority
[22:27] <JEEB> yes, getting money from Ateme is :-P
[22:28] <smarter> mraulet keeps me up to date won what they're doing
[22:28] <smarter> ateme is sponsoring them? I didn't know that
[22:28] <JEEB> well all that PR they're getting by letting Ateme use them as a platform, I would guess they'd be getting money :P
[22:28] <Daemon404> 'as a platform' ?
[22:30] <JEEB> ugh, I'm kind of tired after a bicycle ride, but what I meant is that all the public PR messages about their project pretty much has Ateme's name on them, so I guess Ateme said that they'd get some money to get something working for various dates
[22:30] <JEEB> not that I really know, it just feels just like x264's opencl
[22:30] <JEEB> and "x265"
[22:30] <JEEB> esp. the first
[22:31] <JEEB> which was a case of "everyone works until the OpenCL is 'working' under AMD's money and then it's done. Except it isn't."
[22:31] <JEEB> and yes, I have a rather negative opinion on those guys, just by looking at how they lack any kind of courtesy
[22:31] <JEEB> (regarding upstream)
[22:32] <smarter> well mraulet bought me beer once, that's courteous :p
[22:32] <JEEB> well, sure -- your decoder is now being marketed as the first open source solution, with Ateme's name in the press releases :P
[22:33] <JEEB> I'm sorry but this just smells like "let's just grab that code and hack it up for our use" :D
[22:33] <JEEB> maybe I'm wrong
[22:33] <JEEB> I hope I am, actually
[22:34] <smarter> the PR stuff comes from the GPAC folks I think, I don't know them/their motivations
[22:34] <Daemon404> ugh gpac
[22:34] <JEEB> GPAC and OpenHEVC seem to be rather entangled :P
[22:35] <Daemon404> i have nothing good to say about gpac
[22:37] <JEEB> I think most people who have actually touched the code have that opinion
[22:37] <Skyler_> JEEB: I'm willing to cut them a little bit of slack on the opencl mainly because they just got saddled with a mostly-impossible task :/ and it ran months and months beyond the intended date and stuff
[22:38] <Skyler_> at least now we will maybe have fewer people asking "where is my gpu x264 support"
[22:38] <Skyler_> "it's here, but it's not very good"
[22:38] <JEEB> yeah... at least muggs used his food time for it and kind of finished it
[22:39] <JEEB> I just used it as an example of a project where "<smarter> they do, but it's not their priority" kind of things lead to "budget ended? well, that's it"
[22:39] <JEEB> kind of stuff
[22:39] <Skyler_> Yeah
[22:39] <Skyler_> I hope that doesn't happenthis time
[22:39] <Skyler_> I'm mainly willing to go with it because, gosh, at least they're offering to do something and nobody else is, but...
[22:39] Action: smarter is optimistic/naïve :)
[22:40] <JEEB> also this was mostly regarding the hacking OpenHEVC/GPAC is doing with smarter's decoder
[22:40] <JEEB> not regarding "x265"
[22:40] <JEEB> I'd just hate it if elenril or someone had to start merging stuff again from OpenHEVC
[22:41] <kierank> 21:27:57 <JEEB> yes, getting money from Ateme is :-P --> that's not true
[22:41] <kierank> they get money from the french government
[22:41] <smarter> they really should call it "x265" with the quotes :p
[22:41] <Daemon404> x264++
[22:41] <JEEB> yes, they're a university-related project so naturally they get money from the French government
[22:42] <kwizart> Hello, Can I default to --enable-thumb in arm arches such as armv7-a and armv6 hard float ?
[22:42] <JEEB> anyways, I'm just being cynical as always
[22:42] <smarter> they're also working with the industry on http://www.4ever-project.com/
[22:42] <kierank> ateme don't give them money
[22:50] <JEEB> well, it was just a crazy idea in my head, so you shouldn't really put too much weight in that :) In general it just felt like they were dancing in order to get X ready, and then get whatever was promised from a third party Y. And screw the upstream :D But as I said, I am just being a cynical, negative person in this case.
[23:51] <durandal_1707> michaelni: did you subscribed me to coverity spam?
[23:57] <durandal_1707> is someone going to port lossywav as filter?
[00:00] --- Fri Jun 21 2013


More information about the Ffmpeg-devel-irc mailing list