[Ffmpeg-devel-irc] ffmpeg-devel.log.20130515

[burek]

[01:01] <cone-719> ffmpeg.git 03Clément BSsch 07master:7b43120c7c1e: lavfi: add dctdnoiz filter.
[01:06] <ubitux> Changelog for next ffmpeg is starting to get pretty nice :)
[01:29] <cone-719> ffmpeg.git 03Michael Niedermayer 07master:1a36c756d895: ffmpeg: free threads on error conditions.
[03:20] <cone-719> ffmpeg.git 03Clément BSsch 07master:bc1c90eabfa4: lavfi/dctdnoiz: fix 10l typo in pr_height computation.
[09:43] <cone-639> ffmpeg.git 03Matt Wolenetz 07master:252c0bfdc014: lavc: Fix msvc compilation of utils.c with -WX (warnings treated as errors)
[09:43] <cone-639> ffmpeg.git 03Michael Niedermayer 07master:6bf6b9c279ab: Merge commit '252c0bfdc014c1fb6ad4fe06242c7beca58a6b41'
[09:47] <cone-639> ffmpeg.git 03James Darnley 07master:b0ef0ae77608: yadif: restore speed of the C filtering code
[09:47] <cone-639> ffmpeg.git 03James Darnley 07master:d10499a32ab3: yadif: correct strides in filter_edges_16bit
[09:47] <cone-639> ffmpeg.git 03Claudio Freire 07master:7c71ada4cad3: aacenc: Fix a rounding bug in aacpsy channel bitrate computation
[09:47] <cone-639> ffmpeg.git 03Michael Niedermayer 07master:772c801c7db9: Merge commit '7c71ada4cad3c6aab5fa24155c379465c41cfd76'
[09:53] <cone-639> ffmpeg.git 03Claudio Freire 07master:f4d0a63b5b5c: aacenc: Fix target bitrate for twoloop quantiser search
[09:53] <cone-639> ffmpeg.git 03Claudio Freire 07master:8bbdd20a293e: aacenc: Fix erasure of surround channels
[09:53] <cone-639> ffmpeg.git 03Sebastian Sandberg 07master:1bd57a850a65: vc1dec: fieldtx is only valid for interlaced frame pictures
[09:53] <cone-639> ffmpeg.git 03Michael Niedermayer 07master:46430fd47c62: vc1dec: Don't attempt error concealment on field pictures
[09:53] <cone-639> ffmpeg.git 03Michael Niedermayer 07master:b602e47f12d7: Merge commit '46430fd47c6239ef8742d0a34f9412d5060fa798'
[10:01] <cone-639> ffmpeg.git 03Hendrik Leppkes 07master:fe06795d7837: vc1: implement frame interlaced b-frame header parsing
[10:01] <cone-639> ffmpeg.git 03Michael Niedermayer 07master:cb22692f05f9: Merge commit 'fe06795d783785a76c6cbcaa2cd3012e6deb5fa5'
[10:11] <cone-639> ffmpeg.git 03Martin Storsjö 07master:5d9d8461fbb8: vc1dec: Don't apply the loop filter on fields
[10:11] <cone-639> ffmpeg.git 03Michael Niedermayer 07master:23439cf99bcf: Merge commit '5d9d8461fbb87046a7ca220d3676a045bd7ab46e'
[10:20] <cone-639> ffmpeg.git 03Diego Biurrun 07master:a349458e6769: jpeg2000dec: Use correct printf length modifier for pointer differences
[10:20] <cone-639> ffmpeg.git 03Yoshihisa Uchida 07master:3dff1bf99c65: mmf: Write metadata into an Optional Data chunk
[10:21] <cone-639> ffmpeg.git 03Luca Barbato 07master:7cbc4cb44244: fate: update the mmf reference
[10:21] <cone-639> ffmpeg.git 03Michael Niedermayer 07master:85de84c3e039: Merge commit '7cbc4cb4424430586ce2d5c0dba81174c6d1db01'
[10:26] <cone-639> ffmpeg.git 03Kieran Kunhya 07master:ede75ebc9b9c: Clarify output of av_get_bits_per_pixel
[10:26] <cone-639> ffmpeg.git 03Anton Khirnov 07master:096696ef0dd3: avfiltergraph: simplify inserting conversion filters.
[10:26] <cone-639> ffmpeg.git 03Michael Niedermayer 07master:aa6454e6ca7b: Merge commit '096696ef0dd391d9430376d1444c1a3cde9171fd'
[10:38] <spanish> hi
[10:38] <cone-639> ffmpeg.git 03Anton Khirnov 07master:ffba2053edfc: lavfi: fix compatibility code for old vf_scale options syntax
[10:38] <cone-639> ffmpeg.git 03Michael Niedermayer 07master:5692ad00f488: Merge commit 'ffba2053edfc177d217bf4a95edf51cd0fc40753'
[10:39] <michaelni> hi
[10:40] <spanish> I'm using the following command: ` ffmpeg -i input.mp3 -acodec copy -vcodec copy out.swf' to have a mp3 file split on swf frames
[10:40] <spanish> but when I look at it with sfwdump, some frames are 441 bytes, others 442, is there anyway to pad or otherwise make them all the same size?
[10:42] <michaelni> CBR mp3 does for some bitrates/samplerates have frames that differ by +-1 byte
[10:42] <michaelni> thats how mp3 works
[10:43] <spanish> and there's no chance of padding them up to 442, say with some sort of nop?
[10:43] <cone-639> ffmpeg.git 03Anton Khirnov 07master:7c57a582a03f: jpeg2000dec: don't use deprecated PIX_FMT values
[10:43] <cone-639> ffmpeg.git 03Michael Niedermayer 07master:75835abe910f: Merge commit '7c57a582a03fb473091a88737ab92b9f2a5bb87a'
[10:44] <spanish> I would need this so I don't have to parse the swf when I want random access over the file
[10:44] <spanish> if in constrast, some are 441, other 442 randomly, I would either have to keep a mapping, or parse the file to see where each frame ends
[10:49] <michaelni> you can do all kinds of things, but i dont know if the resulting files are then still valid swf
[10:50] <michaelni> also you can just search for the next mp3 frame when seeking to some byte position
[10:52] <spanish> well, the thing is flash asks my cgi for contents, and it's gonna by much heavier having to do all that, than just computing offsets based on fiexed lengths
[10:53] <spanish> still I would obviously do it, if no other solution is avialable, gonna read up on mp3 and see
[11:17] <cone-639> ffmpeg.git 03Anton Khirnov 07master:e6c4ac7b5f03: pixdesc: rename PIX_FMT_* flags to AV_PIX_FMT_FLAG_*
[11:17] <cone-639> ffmpeg.git 03Michael Niedermayer 07master:ff4680922fc4: Merge remote-tracking branch 'qatar/master'
[11:37] <cone-639> ffmpeg.git 03Michael Niedermayer 07master:c7c71f95f8d3: replace remaining PIX_FMT_* flags with AV_PIX_FMT_FLAG_*
[12:03] <cone-639> ffmpeg.git 03Carl Eugen Hoyos 07master:86ea021b3a3b: Reduce probing score for flic animation files.
[12:03] <cone-639> ffmpeg.git 03Michael Niedermayer 07master:3ac3edf63baa: Merge remote-tracking branch 'cehoyos/master'
[12:15] <cone-639> ffmpeg.git 03Michael Niedermayer 07master:e003413fe969: murmur3: fix memleak
[12:21] <cone-639> ffmpeg.git 03Michael Niedermayer 07master:a9cadacdd982: dpxenc: dont shift into the sign bit.
[12:52] <cone-639> ffmpeg.git 03Paul B Mahol 07master:fa23f94eaca3: wv: make probing more robust
[12:59] <durandal_1707> michaelni: you forgot to sent to nut-devel?
[13:12] <michaelni> no
[13:23] <michaelni> durandal_1707, I did not notice the new mails in nut-devel
[13:49] <durandal_1707> all arch on fate have log1p?
[14:29] <rkrishna> Hi, need to info on how ffmpeg -re works, that is real time encoding
[14:29] <rkrishna> how do we limit the decoding/encoding frame rated in -re case
[14:39] <cone-639> ffmpeg.git 03Paul B Mahol 07master:c45b823bf80f: lavfi/histogram: logarithmic mode for levels
[14:50] <durandal_1707> why "git describe" shows nonsense since forever?
[15:02] <michaelni> i suspect we dont have a better tag on master for it to pick it, also see ./version.sh
[15:26] <ubitux> j-b: flac & vorbis assert has been replaced in ffmpeg already afaict
[15:26] <j-b> ubitux: I know.
[15:26] <ubitux> and afaik a bunch of others
[15:27] <j-b> I know that too.
[15:27] <j-b> That does not change the fact that I need it.
[16:43] <michaelni> j-b, so why dont you use ffmpeg if you need it ?
[20:21] <cone-783> ffmpeg.git 03Michael Niedermayer 07master:0313653928b4: vc1dec: Update destination pointers in context for fields
[20:21] <cone-783> ffmpeg.git 03Michael Niedermayer 07master:012b319d9147: vc1dec: Fix tff == 0 handling in init_block_index()
[20:21] <cone-783> ffmpeg.git 03Michael Niedermayer 07master:19673db34aa0: vc1dec: Factorize srcU/V offseting
[20:21] <cone-783> ffmpeg.git 03Michael Niedermayer 07master:2412ad171716: vc1dec: Factorize picture pointer selection code
[20:21] <cone-783> ffmpeg.git 03Michael Niedermayer 07master:201412ffec04: vc1dec: Factorize picture pointer selection in vc1_mc_4mv_chroma()
[20:21] <cone-783> ffmpeg.git 03Michael Niedermayer 07master:ccb148e4780e: vc1dec: Fix current ptr selection in vc1_mc_4mv_chroma()
[20:21] <cone-783> ffmpeg.git 03Martin Storsjö 07master:ec7d002e5559: vc1, mss2: Check for any negative return value from ff_vc1_parse_frame_header
[20:21] <cone-783> ffmpeg.git 03Michael Niedermayer 07master:4162fc62b30d: vc1dec: Do not allow field_mode to change after the first header
[20:21] <cone-783> ffmpeg.git 03Martin Storsjö 07master:3ca3709ad450: vc1dec: Remove some now unused variables
[20:21] <cone-783> ffmpeg.git 03Michael Niedermayer 07master:0379fc1487d4: vc1: Move INIT_LUT() further up to allow using it in more places
[20:21] <cone-783> ffmpeg.git 03Michael Niedermayer 07master:37f2ac36a909: vc1: Simplify code by using INIT_LUT()
[20:21] <cone-783> ffmpeg.git 03Michael Niedermayer 07master:6ce2c3106d78: vc1: Make INIT_LUT() self contained
[20:21] <cone-783> ffmpeg.git 03Alex Smith 07master:225a5f91cc83: configure: Use linker hardening flags on mingw
[20:21] <cone-783> ffmpeg.git 03Michael Niedermayer 07master:e3c24f663ea7: Merge remote-tracking branch 'qatar/master'
[20:28] <cone-783> ffmpeg.git 03Nicolas George 07master:ca2f6b693214: lavf/concat: reject empty scripts.
[20:28] <cone-783> ffmpeg.git 03Michael Niedermayer 07master:42af97dbc0fa: Merge remote-tracking branch 'cigaes/master'
[20:34] <j-b> michaelni: because I obide to my clients, you know :)
[20:34] <j-b> michaelni: and I'm not in position to tell them what libraries to use.
[21:09] <Daemon404> vlc developers are j-b's "clients" now?
[21:12] <j-b> Daemon404: nope
[21:12] <j-b> Daemon404: real clients :)
[21:13] <nevcairiel> clients should dictate functional requirements, not technical internals
[21:13] <nevcairiel> but then, you are french
[21:13] <j-b> es
[21:14] <j-b> and one functionnality is that it deploys on their infra
[21:14] <j-b> As I'm a fairly new business, I don't dictate much
[22:22] <microchip_> [mpeg2video @ 0xabb440] mpeg2 style quantization not supported by codec
[22:22] <microchip_> nice, MPEG2 doesn support MPEG quants LOL
[23:16] <ubitux> michaelni: between fspp & spp; which one should be ported?
[23:20] <michaelni> ubitux, i always used spp, i dont remember exactly what fspp does exactly but there sure was a reason why it was written as seperate filter and not as an optimization within spp
[23:23] <ubitux> fspp was pushed in jun 2005, without details about the reason it was split from spp
[23:24] <ubitux> ah, there is some stuff on the ml i missed..
[23:28] <ubitux> http://lists.mplayerhq.hu/pipermail/mplayer-dev-eng/2005-March/033866.html
[23:28] <ubitux> "If I patch existing, then it looses soft_threshold (softthresh_c)
[23:28] <ubitux> ability (I guess, forever). But you can play with threshold matrix
[23:28] <ubitux> coefficients instead."
[23:47] <ubitux> ok well
[23:47] <ubitux> i can't choose between the two
[23:47] <ubitux> so i will get busy with something else
[23:53] <iive> ubitux: i told you about fspp, f stands for fast, it is supposed to be faster because it does overlapped 1d tranformations, instead of 2d
[23:54] <durandal_1707> faster -> worse?
[23:54] <ubitux> iive: right, but that doesn't tell me which one i should port :p
[23:54] <durandal_1707> spp
[23:55] <ubitux> will that mean we will have to keep fspp until someone decide to port it?
[23:55] <iive> spp
[23:55] <ubitux> ok
[23:56] <ubitux> porting spp ’ dropping mp={fspp,pp7,spp,uspp}?
[23:56] <ubitux> % ls libavfilter/libmpcodecs/*.c|wc -l
[23:56] <ubitux> 20
[23:56] <ubitux> almost there. :)
[23:57] <iive> hum... i don't remember actually trying pp7. maybe you can find a way to merge it into spp?
[23:58] <ubitux> i'd better not dedicate too much time on this :)
[23:58] <durandal_1707> ubitux: there is 17 filters, not 20
[23:58] <ubitux> sure
[23:59] <durandal_1707> 3 of them are non correct or asserts
[23:59] <durandal_1707> *at least, not tryied others, (there is say 50% chance random filter you pick that its broken)
[00:00] --- Thu May 16 2013