[Ffmpeg-devel-irc] ffmpeg-devel.log.20130508

[burek]

[00:04] <llogan> Compn: no, i wasn't questioning the "n" (i looked that up once but forgot the details...). just the significance of "0.8".
[00:17] <xlinkz0> what are the reasons for getting Operation not permitted on av_seek_frame?
[00:19] <ubitux> return -1
[00:20] <xlinkz0> yeah but why wouldn't the seek work?
[00:21] <ubitux> that's all i can answer with the amount of info you provided
[00:21] <xlinkz0> http://codepad.org/g8PietjZ
[00:23] <xlinkz0> i don't think the whole code is relevant.. basically i'm trying to seek to video_stream->duration - 1500ms
[00:25] <ubitux> reproducible with all formats?
[00:26] <xlinkz0> Stream #0:0(und): Video: h264 (Baseline) (avc1 / 0x31637661), yuvj420p, 1280x720 [SAR 1:1 DAR 16:9], 3245 kb/s, 6.68 fps, 45.42 tbr, 90k tbn, 180k tbc
[00:26] <ubitux> that's a codec
[00:26] <xlinkz0> mp4
[00:26] <ubitux> so, only mp4?
[00:26] <xlinkz0> that's what i'm testing on
[00:27] <ubitux> then compare with something else
[00:27] <xlinkz0> any suggestions?
[00:27] <ubitux> % ./ffmpeg -v 0 -formats | wc -l
[00:27] <ubitux> 254
[00:28] <xlinkz0> but what do i have to do to make the seek work on mp4?
[00:28] <xlinkz0> i need it on mp4..
[00:28] <ubitux> i'm suggesting a way of getting a clue about the issue
[00:28] <ubitux> not a workaround
[00:29] <xlinkz0> ok so i should transcode the video to a different format and try again?
[00:38] <xlinkz0> ubitux: i added pos = pos - (pos % 90000); and now it works
[00:38] <xlinkz0> it seems it will only seek to specific seconds
[00:38] <xlinkz0> not a huge problem for me.. but i think it's a bug in ffmpeg
[01:18] <Compn> llogan : isnt 0.8 the 0.8 branch ?
[01:21] <llogan> i guess so. that's what is seems to show by default, as in "n0.8-21930-gbc0cbac"
[01:21] <llogan> *that's what "git describe" seems to show...
[01:22] <llogan> oh, i think i did something dumb
[01:22] <llogan> because 21930 isn't right either
[01:23] <Compn> no, i think what happeend is your git checkout is for the 0.8 branch
[01:23] <Compn> :P
[01:23] <Compn> you have to git voodoo master or something to switch it back to master
[01:26] <cone-605> ffmpeg.git 03Michael Niedermayer 07master:bca50e5cd522: avcodec/mpegvideo: Fix block height for lowres 3 interlaced blocks
[01:26] <cone-605> ffmpeg.git 03Michael Niedermayer 07master:c67bca2b5a94: avcodec/mpegvideo: Fix edge emu with lowres
[01:44] <cone-605> ffmpeg.git 03Stefano Sabatini 07master:c773adee3fda: lavu/mem: extend documentation for av_dynarray_add()
[01:44] <cone-605> ffmpeg.git 03Stefano Sabatini 07master:ee9794ed2052: lavu/mem: fix potential int overflow and crash in av_dynarray_add()
[01:44] <cone-605> ffmpeg.git 03Stefano Sabatini 07master:ae0ce99ca425: lavfi/unsharp: use verbal form in help message options
[03:00] <highgod> Hi, does anyone ever use ffmpeg rtsp? Is there any sample in ffmpeg? thanks.
[03:27] <Compn> highgod : sending rtsp or recieving rtsp ?
[03:27] <Compn> highgod : there are rtsp sample urls on http://wiki.multimedia.cx (search for rtsp)
[03:31] <highgod> both of them
[03:31] <highgod> Compn:thanks, I will read it
[03:39] <cone-605> ffmpeg.git 03Christophe Gisquet 07master:fc37cd43338b: x86: sbrdsp: force PIC addressing for Win64
[03:41] <Compn> http://wiki.multimedia.cx/index.php?title=RTSP#Samples_in_the_wild
[03:44] <highgod> Compn: so great, thanks, another question, do you know whether ffserver can by use as a transcode server?
[03:49] <highgod> Compn: hwo can I get the code?
[04:05] <Compn> you mean use it like an ffmpeg youtube backend ?
[04:06] <Compn> there are some projects that use ffmpeg in that way
[04:06] <Compn> i dont remember the names offhand
[04:09] <highgod> oh,hehe
[04:10] <highgod> yes
[04:10] <highgod> use the libs, not use command lines of ffmpeg.exe
[04:18] <Compn> let me google some projects up
[04:19] <Compn> i dont think many projects use ffserver to serve lots of users
[04:19] <Compn> there are more dedicated servers for rtsp and such
[04:20] <Compn> its hard to google. seems like every single web host now has ffmpeg hosting support :)
[04:25] <highgod> Thanks Compn
[09:53] <highgod> ffmpeg is so great
[09:56] <av500> of course
[10:30] <durandal_1707> the older versions of select filter used PICT_TYPE_I instead of I and etc...
[11:15] <Bor0> how would I go implementing buffering state in ffplay? I have something like buffering=1; av_read_frame(...); buffering=0; but when I display the variable it always shows buffering = 0 even when it's buffering
[11:29] <av500> Bor0: how much time is spent inside the av_read_frame() call?
[11:31] <Bor0> it looks like it's working now though, seems that I wrongly printed the value
[11:31] <Bor0> av500: I have no idea. it's a libff* internal function
[11:32] <av500> yes, but you can instrument around it
[11:32] <Bor0> how do you mean
[11:32] <Bor0> by timing it?
[11:33] <av500> indeed
[11:33] <av500> but it looks like you found your error
[11:33] <Bor0> seems I did, now it only remains to know if this is the right way to implement such a thing
[11:34] <Bor0> assuming that my stream is a http
[11:34] <av500> depends on what you want to achieve
[11:34] <Bor0> when I get the same picture displayed for a couple of seconds, I want the user to be notified that his player did not freeze but it is downloading from the slow http server
[11:35] <Bor0> "a couple of seconds", it could be instantly
[11:39] <av500> on slow connections you might in fact implement some buffering yourself
[11:39] <Bor0> how would I go for that?
[11:39] <av500> so read_frames() and store them for a while, then start playing
[11:46] <cone-474> ffmpeg.git 03Janne Grunau 07master:45a1694fde25: jpeg2000dec: cleanup properly after decoding errors
[11:46] <cone-474> ffmpeg.git 03Michael Niedermayer 07master:aecb3421f641: Merge commit '45a1694fde251ee56e23aec073d73172c00aa470'
[11:47] <durandal_1707> michaelni: merging again?
[11:54] <cone-474> ffmpeg.git 03Janne Grunau 07master:70f9661542a5: jpeg2000dec: initialize static data just once
[11:54] <cone-474> ffmpeg.git 03Michael Niedermayer 07master:1bad40ef2723: Merge commit '70f9661542a581dfe93b636b1c55b5558e4a4e3c'
[11:55] <durandal_1707> michaelni: say when you are done so i can push little bomb
[11:58] <Bor0> av500: could you show me some example implementation of that? I understand how it would be implemented but it'd be great if I can see an example
[11:59] <av500> sorry no
[11:59] <av500> also, its totally not about the channel topic :)
[12:00] <Bor0> okay, thanks about your help. yes I understand, but this channel seems to be far more active regarding development than #ffmpeg (where mostly ffmpeg command line stuff is discussed) :)
[12:01] <Bor0> I guess it's the devel keyword :p
[12:01] <av500> if you look at ffplay, there is already some buffering of frame in queues
[12:04] <Bor0> yeah, I am using ffplay. I see that buffering is already implemented, I just need to show an indicator to the user when the buffer is empty and it is waiting for a buffer fill
[12:04] <cone-474> ffmpeg.git 03Anton Khirnov 07master:a97d8cc16e0d: indeo3: use unaligned reads on reference blocks.
[12:04] <cone-474> ffmpeg.git 03Michael Niedermayer 07master:e077ba801644: Merge commit 'a97d8cc16e0da30c9ffefa1ede2a0adf3db5f3f8'
[12:06] <rkrishna> Hi, any plans to add support for AVCi50/100 to ffmpeg?
[12:06] <JEEB> is it one of those crazy spec-incompliant "AVC Intra" formats?
[12:07] <JEEB> because any sane'ish AVC will get decoded just fine in pretty much all cases
[12:07] <Compn> rkrishna : got a sample of a file that doesnt work in ffmpeg ?
[12:08] <durandal_1707> see c5142a95a51320e65d80d6bca0eb69fcff05508f
[12:08] <rkrishna> I am actually looking for Encoder support, if we can implement it in some form
[12:09] <JEEB> x264 is used for AVC encoding
[12:09] <JEEB> the problem is that there is no single specification for AVC Intra
[12:10] <JEEB> there's a "nice" mailing list thread about AVC Intra on x264-devel
[12:10] <rkrishna> ok, so I should be talking to x264 devel guys. 
[12:12] <rkrishna> my understanding is we can intra encode with x264. We probably need to handle the bit stream so that it is compatible with 3rd party softwares
[12:14] <cone-474> ffmpeg.git 03Anton Khirnov 07master:481575d25519: avconv: remove -deinterlace
[12:14] <cone-474> ffmpeg.git 03Michael Niedermayer 07master:42f5fc716e93: Merge commit '481575d255195c6c2f32e04686b94de984ae78d4'
[12:14] <Compn> rkrishna : did you check ffmbc project ? it is tailored for avc broadcasting support
[12:15] <JEEB> rkrishna, there was a thread in jan/feb 2012 @ http://mailman.videolan.org/pipermail/x264-devel/2012-January/009048.html , http://mailman.videolan.org/pipermail/x264-devel/2012-February/009078.html
[12:15] <JEEB> it seems like there just is not a single specification for AVC Intra :P
[12:15] <Paranoialmaniac> what is AVCi50/100?
[12:17] <Paranoialmaniac> afaik AVC-Intra itself is not an issue, A fatal issue i know is that AVC-Intra-in-MOV generated by FCP doesn't contain any SPS and PPS.
[12:17] <rkrishna> I checked ffmbc it does not say any thing about AVCi50/100
[12:17] <j-b> wow. so how do you decode those?
[12:18] <JEEB> j-b, I guess they are hardcoded in the decoder... sounds as awful as it is :D
[12:20] Action: j-b 's head explodes
[12:20] <j-b> JEEB: you got to be kidding me
[12:21] <av500> like mjpeg with fixed tables :)
[12:25] <cone-474> ffmpeg.git 03Anton Khirnov 07master:bce70f817059: configure: make avconv select (a)trim filters
[12:25] <cone-474> ffmpeg.git 03Michael Niedermayer 07master:bf0547504b73: Merge commit 'bce70f8170592599a4f2f5f02606c1246e88df57'
[12:40] <michaelni> mjpeg with fixed tables is not that bad, its just one or 2 kinds of tables, in h264 its the whole set of global headers that are missing
[12:40] <michaelni> they would set more than 1 table
[12:41] <michaelni> and its not one fixed value, IIRC avc intra uses different values depending on resolution and such
[12:42] <cone-474> ffmpeg.git 03Anton Khirnov 07master:9779f93066ee: configure: add dependencies to lavfi .pc file
[12:42] <cone-474> ffmpeg.git 03Michael Niedermayer 07master:28397de4df4b: Merge commit '9779f93066ee5ea6fba6b76f001d358cc2857891'
[12:43] <cone-474> ffmpeg.git 03Michael Niedermayer 07master:1aed0513a1c2: configure: use shorter variable name for "libavdevice_pc_deps"
[12:50] <durandal_1707> why was drawgrid rejected?
[12:53] <cone-474> ffmpeg.git 03Anton Khirnov 07master:cd4bfe3a50ee: FATE: use a less ambiguous end time for filter-trim-time test
[12:53] <cone-474> ffmpeg.git 03Luca Barbato 07master:e0c53c340818: nut: use meaningful error values
[12:53] <cone-474> ffmpeg.git 03Michael Niedermayer 07master:9818bbde5303: Merge commit 'e0c53c3408181d2e7d544eb65fc7355a79438dfe'
[13:37] <cone-474> ffmpeg.git 03Luca Barbato 07master:e55e8b2c362a: lavc: do not leak frames in avcodec_decode_audio4
[13:37] <cone-474> ffmpeg.git 03Michael Niedermayer 07master:fe6a36e7687b: Merge commit 'e55e8b2c362a60bad66929d83d8c83b0d7f5ffc9'
[13:41] <cone-474> ffmpeg.git 03Luca Barbato 07master:6f98508e2b86: oma: return meaningful errors
[13:41] <cone-474> ffmpeg.git 03Michael Niedermayer 07master:d13bc208da3b: Merge commit '6f98508e2b86dcb79614d8c9b8e97c5e83a21cb0'
[13:42] <michaelni> durandal_1707, it was rejected ?
[13:55] <cone-474> ffmpeg.git 03Luca Barbato 07master:db9aee6ccf18: oma: properly forward errors in oma_read_packet
[13:55] <cone-474> ffmpeg.git 03Michael Niedermayer 07master:f302ba4ddab7: Merge commit 'db9aee6ccf183508835acc325f5ad87d595eacc4'
[13:59] <cone-474> ffmpeg.git 03Diego Biurrun 07master:7edaf4edb5c3: x86: rnd_template: Eliminate pointless OP_AVG macro indirection
[13:59] <cone-474> ffmpeg.git 03Michael Niedermayer 07master:bf18810a215c: Merge commit '7edaf4edb5c3c04f34ad1242680cbc32d11f4087'
[14:05] <cone-474> ffmpeg.git 03Diego Biurrun 07master:92f8e06ecb43: x86: dsputil hpeldsp: Move shared template functions into separate object
[14:05] <cone-474> ffmpeg.git 03Michael Niedermayer 07master:2f9ef60c9765: Merge commit '92f8e06ecb431a427ea13d794e5a6bc927a034d2'
[14:12] <cone-474> ffmpeg.git 03Diego Biurrun 07master:63bac48f734f: x86: dsputil: Move rv40-specific functions where they belong
[14:12] <cone-474> ffmpeg.git 03Michael Niedermayer 07master:69d2eff5af67: Merge commit '63bac48f734fc69cca2ef2cfada92cd9a222734d'
[14:17] <cone-474> ffmpeg.git 03Diego Biurrun 07master:3334cbec0aa1: x86: dsputil: Remove unused MOVQ_BONE macro
[14:17] <cone-474> ffmpeg.git 03Diego Biurrun 07master:1b343cedd7cd: x86: dsputil: Remove unused ff_pb_3F constant
[14:18] <cone-474> ffmpeg.git 03Michael Niedermayer 07master:e3869dd17ee1: Merge commit '1b343cedd7cd68e7865aa5280d1568c7e5d79917'
[14:31] <cone-474> ffmpeg.git 03Diego Biurrun 07master:bd392785a75d: fate-run: Drop unnecessary regtest() indirection
[14:31] <cone-474> ffmpeg.git 03Michael Niedermayer 07master:63208c034bf6: Merge commit 'bd392785a75d7d51414a9788d50396a8d9f29d77'
[14:41] <cone-474> ffmpeg.git 03Martin Storsjö 07master:06122c253394: msvc: Move linker pragma from config.h to libavutil/internal.h
[14:41] <cone-474> ffmpeg.git 03Michael Niedermayer 07master:8d9537228f03: Merge remote-tracking branch 'qatar/master'
[14:41] <michaelni> durandal_1707, merge done, fire at will
[14:49] <cone-474> ffmpeg.git 03Paul B Mahol 07master:5c057433ccd3: libswscale: GBRAP input & output and GBRAP16 input support
[14:49] <cone-474> ffmpeg.git 03Paul B Mahol 07master:1f2baec7bd60: lavfi: add gbrap support to some filters
[14:49] <cone-474> ffmpeg.git 03Paul B Mahol 07master:94ff68ccb3d1: lavfi/lut: add support for YUVA422P and YUVA444P
[14:49] <cone-474> ffmpeg.git 03Paul B Mahol 07master:443181c0a9a8: libswscale: simplify isAnyRGB macro
[14:49] <cone-474> ffmpeg.git 03Paul B Mahol 07master:8e2972274e1b: lavfi/null: remove redudant lines
[15:29] <kurosu_> is there a macro to round a value A to the nearest multiple of N being >A ? (something like  (A+N-1) & -N if I'm not mistaken)
[15:30] <nevcairiel> there is FFALIGN, but it only works for N being a power of two
[15:30] <kurosu_> yeah, what I meant, sorry
[15:31] <nevcairiel> well then thats what you want :)
[16:07] <Hywan> Hi all. I would like to know if it is possible to record a Mac OS X screen without using VLC or another software. I read this post (from 2010) http://ffmpeg.org/pipermail/ffmpeg-devel/2010-May/095605.html that gives some clues. Are they any developments in this direction currently?
[16:15] <xlinkz0> Hywan: http://ffmpeg.org/trac/ffmpeg/wiki/How%20to%20grab%20the%20desktop%20(screen)%20with%20FFmpeg
[16:16] <Hywan> xlinkz0: Mac OS is not listed here.
[16:16] <xlinkz0> doesn't mac also use x11?
[16:17] <Hywan> xlinkz0: With XQuartz, yes, but you will be able to only grab X11 windows, not the other (native) windows
[16:18] <xlinkz0> idk
[16:18] <xlinkz0> doesn't apple provide its own software for doing so?
[16:19] <Hywan> Yes. Quicktime Player, but I can't use it as an input of ffmpeg (except if I do it in 2 steps: recording at first and then processing with ffmpeg). And Quicktime Player can not be used in CLI.
[16:22] <durandal_1707> ubitux: FF_RIGHT_SHIFT_ROUNDED_TOWARD_INFINITY()
[16:23] <durandal_1707> but FF_DO_NOT_USE() is less scary
[16:26] <ubitux> :/
[16:32] <Compn> Hywan : install qt_tools , command line interface for quicktime
[16:32] <Compn> its very useful :)
[16:33] <Compn> Hywan : http://omino.com/sw/qt_tools/
[16:33] <Hywan> Compn: Does it allow me to record a specific region of the screen? I didn't see this on the website.
[16:35] <Compn> actually i'm not sure if it will work for you
[16:35] <Compn> lets see if i can find something better
[16:37] <Compn> looks like a two step process so far , http://superuser.com/questions/164892/screen-capture-command-line
[16:38] <Hywan> Compn: I would like to precisely specify a region of my screen.
[16:38] <Compn> Hywan : this is what you probably want > https://github.com/csexton/captured
[16:40] <Compn> http://www.tuaw.com/2007/05/17/terminal-tip-6-easy-ways-to-capture-your-screen/
[16:40] <Hywan> Compn: "captured" only capture images.
[16:40] <Compn> i wonder if you can set screencapture command line program to pipe png files to ffmpeg for outputting as a movie or stream
[16:41] <Hywan> Compn: Wow, it would be very slow :-/, no?
[16:41] <Compn> mkfifo lotsofpng && capture | ffmpeg -f image2 -i lotsofpng out.mov
[16:41] <Compn> for encoding png ? shouldnt be that slow
[16:41] <Compn> depending on encoding codec
[16:41] <Compn> depending on what you are recording
[16:42] <Compn> are you screen recording video or just regular programs ?
[16:42] <Hywan> regular programs
[16:43] <Compn> should be fine, but i'm not sure the interval possible with those programs
[16:44] <Hywan> Compn: Very low :-(
[16:46] <ubitux> http://pastie.org/7818302  any suggestion before i add this to the wiki?
[17:05] <Hywan> ubitux: xwininfo --root would be better no?
[17:12] <ubitux> the idea is to record a window you clic on
[17:15] <durandal_1707> then add that explanation to wiki too
[17:51] <d```> is there a list or page of all the hardware supported by ffmpeg? I'm trying to find an ARM platform that's supported
[17:55] <funman> d```: if it runs linux it is supported
[17:56] <d```> funman: as far as I can tell, there's no support for marvell vmeta on ARM
[17:57] <d```> funman: http://lists.ffmpeg.org/pipermail/ffmpeg-devel-irc/2012-June/000708.html
[17:57] <d```> and yes, I realize that was from almost a year ago
[17:58] <funman> that's for some specific hardware decoder afaiu
[17:58] <funman> cpu decoding works fine
[17:59] <d```> that's what I'm looking for, what hardware is supported
[17:59] <d```> sorry if I was unclear
[18:00] <Compn> ubitux : i dont know what this is ... but its interesting. http://ebit.lbl.gov/~arun/ffmpeg/index.html
[18:00] Action: Compn searching for 'ffmpeg' on .gov sites
[18:00] <funman> d```: i'm fairly sure, 'none'
[18:01] <cptspiff> d```: no omx in ffmpeg afaik. which typically is the interface to these decoders (or properietary nda'd shit).
[18:02] <d```> funman cptspiff: thanks for the info, this makes me sad
[18:04] <Compn> d``` : you mean you want hardware acceleration support list ?
[18:04] <Compn> there should be one in ffmpeg, just grep hwaccel 
[18:04] <d```> Compn: that's exactly what I'm looking for
[18:04] <cptspiff> np. you can have a look at https://github.com/xbmc/xbmc/tree/master/xbmc/cores/omxplayer for a player using avformat for demux and omx for decode
[18:05] <Compn> d``` : unfortunately, all the arm apps i've seen using ffmpeg are closed source and no one felt like sharing code with us
[18:05] <Compn> theres a few ffmpeg-android projects which may have optimizations. but i havent done an exhaustive search to find out 
[18:05] <cptspiff> droid needs stagefright love..
[18:06] <Compn> d``` : its quicker just for us to say ffmpeg doesnt have any arm based hwaccels iirc
[18:07] <funman> Compn: vlc is open source ;)
[18:07] <funman> we use omx decoder on android
[18:07] <Compn> :)
[18:07] <Compn> funman : why u no share with us your code then ?
[18:07] <Compn> :P
[18:07] <funman> maintained by wbs
[18:07] <funman> Compn: just for evil :P
[18:08] <d```> Compn: thanks for all the info, I wish I knew all this a few days ago
[18:08] <cptspiff> funman: is omx avail on all hw?
[18:08] <Compn> http://limoa.sourceforge.net/docs/0.3/group__limoi__component__ffmpeg__omx.html
[18:08] <Compn> whats this limoa thing
[18:08] <Compn> it looks useful to us...
[18:09] <Compn> cptspiff : not a chance in hell
[18:09] <Compn> all these arm devices have different hw
[18:09] <Compn> and api
[18:11] <Compn> am curious what people use ffmpeg on a phone for anyway
[18:11] <Compn> streaming video ?
[18:11] <cptspiff> phones, not so interesting. pads, more interesting.
[18:11] <funman> cptspiff: omx is just an api
[18:11] <funman> which loads real decoders dynamically if present
[18:12] <cptspiff> yes, i know this. point is, we fell down on stagefright since if you use omx, there is no guarantee there are sw decoders around. with stagefright you have a minimum set of features that MAY be hw backed.
[18:12] <funman> and yes every device as its own decoder
[18:13] <Compn> but yeah, having xbmc on android is amazing...
[18:14] <JEEBsv> I still have my first-gen Xbox with XBMC on it :D
[18:15] <cptspiff> heh mine still lives as well. which is a damn miracle consider how i have treated it over the years
[18:37] <durandal_1707> http://imageshack.us/download/826/27283552.png
[19:08] <durandal_1707> one can use overlay with timeline?
[19:11] <ubitux> yes
[19:13] <durandal_1707> then i could add it to blend
[19:16] <durandal_1707> i get warnings, and wrong few frames, guess because relevant code is missing
[19:16] <ubitux> you need to use the passthrough callback
[19:18] <durandal_1707> ugh, that code in overlay is not nice
[21:38] <cone-463> ffmpeg.git 03Paul B Mahol 07master:deef4b1ac944: lavfi/drawutils: add 16 bit rgb and planar rgb to ff_fill_rgba_map()
[21:38] <cone-463> ffmpeg.git 03Paul B Mahol 07master:8245520b8800: lavfi/colorchannelmixer: remove now redudant code
[21:38] <cone-463> ffmpeg.git 03Paul B Mahol 07master:c4a5499d259b: lavfi/extractplanes: packed rgb support
[22:09] <ubitux> durandal11707: how did you do the dep thing?
[22:09] <ubitux> added #if CONFIG... around the specific extractplanes code?
[22:10] <durandal11707> ubitux: in configure
[22:10] <ubitux> what if you want only alphaextract?
[22:10] <durandal11707> also i moved #if CONFIG to right place (it was wrong)
[22:11] <durandal11707> that is little akward
[22:11] <ubitux> why?
[22:11] <durandal11707> well, removing deps from configure should make that possible
[22:12] <durandal11707> i could add ifdev around options if that would silence warnings
[22:13] <durandal11707> *ifdef
[22:13] <durandal11707> and unused class..
[23:15] <ubitux> durandal11707: could we use extract/merge planes filter to make some filter work only on specific planes?
[23:17] <durandal11707> not yes, as there is not mergeplanes filter
[23:17] <durandal11707> *yet
[23:18] <ubitux> i wonder what you are waiting for!
[23:18] <durandal11707> i want to make good chroma key filter
[23:19] <durandal11707> i already sent link to some experiment i had with colorkeymask and overlay
[23:20] <durandal11707> results is pretty good for such trivial implementation, but i can be improved (adding extra modes ...)
[23:22] <durandal11707> for merge planes filter you just want to be able to merge gray single planes?
[23:23] <durandal11707> well i can do it if only merging is going to be made (hence that name....)
[23:42] <ubitux> should i just name the macro FF_SHIFTUP()?
[23:56] <michaelni> thats possible but i think noone will understand what the macro does without reading the #define
[00:00] --- Thu May  9 2013