[Ffmpeg-devel-irc] ffmpeg-devel.log.20130724

burek burek021 at gmail.com
Thu Jul 25 02:05:02 CEST 2013


[00:00] <durandal_1707> llogan: what?
[00:00] <saste_> llogan, nice? doesn't sound like the correct term
[00:01] <llogan> a worthless, but entertaining joke filter.
[00:01] <saste_> i'd like a naive voice synth
[00:02] <saste_> not sure if we can do timelined audio overlay at the moment
[00:02] <durandal_1707> what is that?
[00:03] <saste_> you overlay audio at the given interval
[00:03] <saste_> like overlay=...:enable=expr
[00:05] <llogan> does clement not code in the northern hemisphere summer?
[00:05] <saste_> clement, not active in this period, he will be back in august he said
[00:08] <durandal_1707> saste_: you have pan,amerge for this
[01:14] <cone-528> ffmpeg.git 03Michael Niedermayer 07master:274a50ab5738: avcodec/utils: Warn the user about the lack of a lock manager if insufficient locks are detected
[03:02] <kierank> that has got to be the greatest ticket known to man
[03:05] <drv> it has cool pictures
[03:05] <kierank> 5 years of no aac work and now this
[03:05] <kierank> encoder work
[04:49] <cone-815> ffmpeg.git 03Zhang Rui 07master:6bab3430a775: avformat/http: support reading compressed data
[10:23] <cone-292> ffmpeg.git 03Luca Barbato 07master:70ecc175c7b5: alsdec: Fix the clipping range
[10:23] <cone-292> ffmpeg.git 03Michael Niedermayer 07master:865f34823c57: Merge commit '70ecc175c7b513a153ac87d1c5d219556ca55070'
[10:39] <cone-292> ffmpeg.git 03Luca Barbato 07master:ca488ad48036: alsdec: Clean up error paths
[10:39] <cone-292> ffmpeg.git 03Michael Niedermayer 07master:a550926705fe: Merge commit 'ca488ad480360dfafcb5766f7bfbb567a0638979'
[11:38] <cone-292> ffmpeg.git 03Luca Barbato 07master:2f034f255c49: 4xm: Reject not a multiple of 16 dimension
[11:38] <cone-292> ffmpeg.git 03Michael Niedermayer 07master:82654172ced5: Merge commit '2f034f255c49050e894ab9b88087c09ebe249f3f'
[12:00] <cone-292> ffmpeg.git 03Luca Barbato 07master:bd7b4da0f462: 8bps: Bound-check the input buffer
[12:00] <cone-292> ffmpeg.git 03Michael Niedermayer 07master:da0f67b177ad: Merge commit 'bd7b4da0f4627bb6c4a7c2575da83fe6b261a21c'
[12:16] <cone-292> ffmpeg.git 03Luca Barbato 07master:6fd221e5f8b0: 8bps: Make the bound-checks consistent
[12:16] <cone-292> ffmpeg.git 03Michael Niedermayer 07master:7dec8512be12: Merge remote-tracking branch 'qatar/master'
[12:56] <cone-292> ffmpeg.git 03Michael Niedermayer 07master:800637709c82: alsdec: minor simplification in read_block()
[12:56] <cone-292> ffmpeg.git 03Anshul Maheshwari 07master:a553cb4683a2: documented that av_codec_close is required to call
[13:06] <cone-292> ffmpeg.git 03Michael Niedermayer 07master:101c7ea90673: avcodec/mpeg12dec: remove commented out assert
[14:10] <cone-292> ffmpeg.git 03Bernie Habermeier 07master:c5f3cc40e33a: matroskaenc: implement CueRelativePosition
[16:30] <cone-292> ffmpeg.git 03Timothy Gu 07master:11cb697501ed: doc/encoders: partially rewrite and reformat libx264 docs
[16:31] <Daemon404> o shit it finally went it
[16:31] <Daemon404> in*
[16:37] <saste_> Daemon404, why, was you waiting for it?
[16:37] <saste_> it took "only" a week, which is not so much for a 200+ lines documentation patch
[16:38] <Daemon404> i was just following the very large email thread
[16:57] <saste> ^^ epic ticket
[17:17] <Daemon404> yes epic
[17:57] <aballier> michaelni: (and others too) what is your opinion on an avframe based version of sws_scale/swr_convert that would allocate the output buffer as needed alike the avframe based avcodec_decode/encode_audio/video ?
[17:59] <aballier> if the interface is generic enough it might even be possible to plug it in avcodeccontext so that input frames are converted before being sent to the encoder and output frames are converted before being returned
[18:15] <michaelni> aballier, possible if someone wants to use it and someone volunteers to maintain it
[18:17] <aballier> i can try to bake up a patch
[18:17] <aballier> just wanted to be sure its not an idea that has been considered and discarded for some reason i dont know
[18:23] <michaelni> i dont remember that being suggested previously (at least not in that form)
[18:25] <wm4> michaelni: do I need to set any magical flags to make the libswscale chroma pos stuff for yuv->rgb take effect?
[18:26] <michaelni> wm4, no but its very easily possible theres a bug, for example some optimized X->Y converter could end up being used that doesnt know about chrma pos
[18:27] <nevcairiel> probably disable the inaccurate scalers, accurate_rnd and the full chroma flags
[18:27] <nevcairiel> or something
[18:27] <wm4> I was trying SWS_LANCZOS | SWS_FULL_CHR_H_INT | SWS_FULL_CHR_H_INP | SWS_ACCURATE_RND | SWS_BITEXACT
[18:29] <cone-292> ffmpeg.git 03Michael Niedermayer 07master:e31d20255d81: vf_scale: add output h/v_chr_pos parameterss
[18:30] <wm4> oh, it only does something if I set the dst_ parameters
[18:30] <wm4> but wouldn't refer dst_... for yuv->rgb position to RGB, which does not have chroma position stuff?
[18:31] <wm4> are they just swapped?
[18:31] <michaelni> i hope not
[18:42] <cone-292> ffmpeg.git 03Michael Niedermayer 07master:24a4367ae963: vf_scale: add input h/v_chr_pos parameters
[18:49] <hotwings> according to the author of the softhddevice output plugin for vdr, theres a problem with the vdpau decoder not ffw/rew'ing h264 (h264 paff to be exact, iirc). would this be the right place to talk to someone about that?
[18:50] <wm4> michaelni: so, is there something wrong with it?
[18:56] <michaelni> hotwings, if theres someone around who knows vdpau well, not sure there is atm (carl is listed as maintainer)
[18:56] <michaelni> reimar knows vdpau too
[18:56] <michaelni> and probably others
[18:57] <michaelni> hotwings, if its reproduceable with ffmpeg/ffplay then probably easiest to open a bug report
[18:57] <michaelni> wm4, dunno, how can i reproduce it ?
[19:07] <Compn> hotwings : upload a sample file that causes problems to the trac
[19:08] <Compn> er, and report the bug on the trac i mean
[19:19] <wm4> michaelni: I tried: ffmpeg -i chroma-pos-mpeg2.y4m -vf 'crop=32:32,scale=1000:1000:out_v_chr_pos=128:out_h_chr_pos=128:flags=accurate_rnd+full_chroma_int+full_chroma_inp+bilinear,format=rgba' avchroma.png -v debug
[19:19] <wm4> michaelni: changing the in_v_chr... etc. params didn't change anything
[19:19] <wm4> only out
[19:19] <wm4> input is yuv420
[22:53] <cone-292> ffmpeg.git 03Michael Niedermayer 07master:436616fd4251: swscale: exchange src & dst chroma pos
[00:00] --- Thu Jul 25 2013


More information about the Ffmpeg-devel-irc mailing list