[Ffmpeg-devel-irc] ffmpeg-devel.log.20120227

[burek]

[00:49] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r59affed23c 10ffmpeg/ (doc/eval.texi libavutil/eval.c tests/ref/fate/eval): 
[00:49] <CIA-17> ffmpeg: eval: add root() to solve f(x)=0
[00:49] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[03:31] <patrykmoura> Hey guys, im tryin to install ffmpeg in my CentOS, ive install all libraries with disabled-shared and prefix=/home/vamolaco/
[03:31] <patrykmoura> My problem is, ffmpeg keeps installing libs at /usr/lib... ive tried to change the --libdir=/home/vamolaco/lib, but ffmpeg again installed libasound.so.2 (link) in /usr/lib and the libasound.so in the same dir (/usr/lib)... i want to install everything under the prefix ive setted...
[03:31] <patrykmoura> Can you guys help me, pls?
[03:32] <patrykmoura> now im tryin with enable-static... idk if it ll work...
[03:32] <patrykmoura> =(
[03:42] <Compn> http://www.whatsmypass.com/the-top-500-worst-passwords-of-all-time
[04:47] <CIA-17> ffmpeg: 03Anton Khirnov 07master * r493a86e25b 10ffmpeg/tests/ (7 files in 2 dirs): 
[04:47] <CIA-17> ffmpeg: FATE: remove a bunch of useless -vsync 0
[04:47] <CIA-17> ffmpeg: No changes in the test results.
[04:47] <CIA-17> ffmpeg: 03Anton Khirnov 07master * r87d7a92b62 10ffmpeg/ (11 files in 2 dirs): rawdec: set timebase to 1/fps.
[04:47] <CIA-17> ffmpeg: 03Paul B Mahol 07master * rcf6914e27f 10ffmpeg/libavcodec/cdxl.c: 
[04:47] <CIA-17> ffmpeg: cdxl: remove early check for bpp
[04:47] <CIA-17> ffmpeg: Signed-off-by: Paul B Mahol <onemda at gmail.com>
[04:47] <CIA-17> ffmpeg: Signed-off-by: Justin Ruggles <justin.ruggles at gmail.com>
[04:47] <CIA-17> ffmpeg: 03Anton Khirnov 07master * rd43040e28b 10ffmpeg/avconv.c: 
[04:47] <CIA-17> ffmpeg: avconv: refactor vsync code.
[04:47] <CIA-17> ffmpeg: No functional changes, just make it more readable.
[04:47] <CIA-17> ffmpeg: 03Paul B Mahol 07master * r126daeb550 10ffmpeg/libavcodec/cdxl.c: 
[04:47] <CIA-17> ffmpeg: cdxl: set pix_fmt PAL8 only if palette is available
[04:47] <CIA-17> ffmpeg: Signed-off-by: Paul B Mahol <onemda at gmail.com>
[04:47] <CIA-17> ffmpeg: The code in parse_nal_units would always end up setting has_b_frames
[04:47] <CIA-17> ffmpeg: to "1", except when stream is explicitly marked as low delay.
[04:47] <CIA-17> ffmpeg: Since the parser itself would create 'extradata', simply reopening
[04:47] <CIA-17> ffmpeg: the parser would cause this.
[04:47] <CIA-17> ffmpeg: 03Anton Khirnov 07master * r13f6917ca9 10ffmpeg/avconv.c: avconv: set discard on input streams automatically.
[04:47] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r574dcb5baf 10ffmpeg/: (log message trimmed)
[04:47] <CIA-17> ffmpeg: Merge remote-tracking branch 'qatar/master'
[04:47] <CIA-17> ffmpeg: * qatar/master:
[04:47] <CIA-17> ffmpeg:  lavf: don't guess r_frame_rate from either stream or codec timebase.
[04:47] <CIA-17> (12 lines omitted)
[12:29] <cbsrobot> Compn: thew - mine is on place 501
[12:43] <Compn> its 123451 isnt it ?
[12:52] <iive> please don't put your passwords on publicly logged channel!
[12:53] <Compn> iive: http://www.whatsmypass.com/the-top-500-worst-passwords-of-all-time
[12:54] <iive> :)
[12:54] <cbsrobot> Compn: no it's: 501
[13:02] <merbanan> :/ my pincode is there
[13:04] <Compn> if its easy to type, its easy to use for lots of people :\
[13:04] <cbsrobot> merbanan: 666666 ?
[13:04] <merbanan> 1111
[13:04] <Compn> six sixes! :P
[13:05] <merbanan> I'll switch to 4444
[13:05] <merbanan> it is at least longer down the list
[13:09] <Compn> surprised people dont use their names represented by phone keypad
[13:10] <Compn> e.g. 2=abc 3=def 4=ghi 5=jkl
[13:10] <Compn> or however that is setup
[13:10] Action: Compn somehow forgets phone numbers already
[13:28] Action: mystica555_ has his name as his google voice last-4 digits ;)
[16:43] <gnafu> iive: I don't know what you're talking about.  All I saw was this: "<@Compn> its ****** isnt it ?"
[16:53] <Compn> do ho ho
[16:53] <Compn> hunter2 ! 
[16:53] <gnafu> Compn: "******* !"?
[16:53] <gnafu> ;D
[16:53] <gnafu> That's one of my favorites.
[17:23] <CIA-17> ffmpeg: 03Nicolas George 07master * re36504b5a9 10ffmpeg/tests/ (lavfi-regression.sh ref/lavfi/drawbox): fate: add test for the drawbox video filter.
[17:23] <CIA-17> ffmpeg: 03Nicolas George 07master * r4d27a5b2c2 10ffmpeg/tests/ (lavfi-regression.sh ref/lavfi/setdar ref/lavfi/setsar): fate: add tests for the setsar and setdar video filters.
[17:23] <CIA-17> ffmpeg: 03Nicolas George 07master * r47bd0bc461 10ffmpeg/tests/ (lavfi-regression.sh ref/lavfi/testsrc): 
[17:23] <CIA-17> ffmpeg: fate: add tests for testsrc.
[17:23] <CIA-17> ffmpeg: This test uses the lavfi device and
[17:23] <CIA-17> ffmpeg: is ready for more similar tests.
[17:23] <CIA-17> ffmpeg: 03Nicolas George 07master * r220de4fa20 10ffmpeg/tests/ (lavfi-regression.sh ref/lavfi/overlay): 
[17:23] <CIA-17> ffmpeg: fate: add test for the overlay video filter.
[17:23] <CIA-17> ffmpeg: The test also covers somewhat the split, scale,
[17:23] <CIA-17> ffmpeg: pad and fifo filters and the graph parser.
[17:23] <CIA-17> ffmpeg: 03Nicolas George 07master * r9380003087 10ffmpeg/tests/ (lavfi-regression.sh ref/lavfi/fade): fate: add test for the fade video filter.
[17:23] <CIA-17> ffmpeg: 03Nicolas George 07master * rf2a05174eb 10ffmpeg/tests/ (lavfi-regression.sh ref/lavfi/pad): fate: add test for the pad video filter.
[17:23] <CIA-17> ffmpeg: 03Nicolas George 07master * rb0de06aee6 10ffmpeg/tests/ (lavfi-regression.sh ref/lavfi/thumbnail): fate: add test for the thumbnail video filter.
[17:23] <CIA-17> ffmpeg: 03Nicolas George 07master * rd8cf9d8a6c 10ffmpeg/tests/ (lavfi-regression.sh ref/lavfi/transpose): fate: add test for the transpose video filter.
[17:23] <CIA-17> ffmpeg: 03Nicolas George 07master * r01f1468eff 10ffmpeg/tests/ (lavfi-regression.sh ref/lavfi/select): fate: add test for the select video filter.
[17:23] <CIA-17> ffmpeg: 03Nicolas George 07master * r48f3d43113 10ffmpeg/tests/ (lavfi-regression.sh ref/lavfi/unsharp): fate: add test for the unsharp video filter.
[17:23] <CIA-17> ffmpeg: 03Nicolas George 07master * ra4c22e3cac 10ffmpeg/tests/ (lavfi-regression.sh ref/lavfi/life): fate: add test for the life video source.
[17:47] <teratorn> hiya, if I decode a yuv420p frame from a video, and the frame.width == 320, but frame.linesize[0] == 352, something is definitely screwed up, isn't it?
[17:47] <ubitux> padding?
[17:47] <teratorn> o.O ?
[17:48] <teratorn> linesize includes padding between the planes!?
[17:48] <teratorn> err
[17:49] <nevcairiel> linesize is the stride or pitch, or whatever you call it
[17:49] <nevcairiel> it will most of the time include extra padding from the end of a line to the start of the next
[17:50] <nevcairiel> this is especially important if your image width is a rather odd size, but the decoder can only effectively work on mod16
[17:51] <ubitux> looks like a mod64 here, no?
[17:51] <iive> it is most likely no-emuedge
[17:52] <nevcairiel> 352 is only mod32, 320 is mod64. :p
[17:52] <nevcairiel> mod32 is quite common for the y plane of yuv420, because that means u/v are mod16
[17:53] <teratorn> hmmm.. OK then, so I should rely on width/heigh when copying plane data...errr this teams I have to copy plane data line-by-line if I must ignore padding on each line?
[17:54] <nevcairiel> if whatever you're copying to cannot cope with padding, then yes
[17:54] <teratorn> I'm copying the imagine to an SDL YUV overlay buffer
[17:54] <teratorn> its linesizes match he width/height without padding
[17:54] <teratorn> http://teratorn.org/code/sdlplayer/
[17:54] <teratorn> nevcairiel, ubitux: thx for help alot
[17:55] <teratorn> is there any trick or optimized way to get frames without padding?
[17:55] <nevcairiel> not really
[17:55] <teratorn> OK then - not a practical concern in my app
[17:55] <nevcairiel> usually copying line-by-line shouldnt be any slower really
[17:55] <nevcairiel> just a few lines of code more
[17:56] <teratorn> memcpy is a macro, and inlined so I guess not
[17:56] <teratorn> well maybe not a macro
[17:56] <teratorn> but inline
[17:57] <ubitux> coverage since nicolas' commit: lavd:1.5%’13.3% lavf:21.9%’38.5% lavu:71.9%’73.4%
[17:57] <ubitux> commits*
[17:58] <ubitux> (lavd increased because of lavfi device)
[18:35] <teratorn> would anyone like to help me fix my image copying problem? I don't know what is wrong with my brain today, but I really need to get this done at work - we would be happy to pay $50 USD even if it takes you 5 minutes
[18:35] <teratorn> busted code: http://teratorn.org/code/sdlplayer/
[18:36] <av500> money upfront?
[18:37] <teratorn> av500: I'm sorry, no
[18:37] <teratorn> av500: and I think I have someone helping me now, thanks
[18:37] <av500> what is there to fix?
[18:37] <av500> you copy the image line by line
[18:37] <av500> and you take stride into account
[18:38] <teratorn> it's corrupted in the playback...
[18:38] <av500> even ffplay should have code for that
[18:38] <av500> then you are copying it wrong
[18:38] <av500> ffplay uses SDL as well
[18:38] <av500> so see it there
[18:38] <teratorn> good tip, thanks
[18:39] <av500> it was free :)
[18:40] <gnafu> Such a nice guy.
[18:40] <gnafu> How uncharacteristic of you.
[18:40] <gnafu> ;D
[18:40] <av500> right
[18:40] Action: av500 invoices gnafu
[18:40] <gnafu> D:
[18:47] <teratorn> well, ffplay.c is just a little bit complicated - so offer still stands in case anyone wants to have a crack at fixing my C code :)
[18:51] <av500> overlay->pitches[0] = frame.linesize[0];
[18:51] <av500> why???
[18:51] <av500> you copy from frame to overlay
[18:51] <av500> each with its own stride
[18:51] <av500> aka linesize
[18:52] <av500> or (i = 0; i < rect.h / 2; i++)
[18:52] <av500> for (i = 0; i < rect.h / 2; i++)
[18:52] <av500> why /2?
[18:53] <av500> ah ok
[18:53] <av500> missed the other loop
[18:54] <av500> so drop  overlay->pitches[x] = frame.linesize[x];
[18:54] <teratorn> hm
[18:54] <teratorn> oh god
[18:54] <teratorn> that is still in there
[18:55] <teratorn> av500: paypal address?
[18:55] <av500> it works now?
[18:55] <teratorn> told you my brain wasn't working today
[18:55] <teratorn> yep
[18:55] <nevcairiel> shouldnt do any work on mondays
[18:56] <av500> s/mon/week/
[18:56] <teratorn> nevcairiel: I fully agree with that sentiment
[18:56] <nevcairiel> however, if you dont work on mondays, tuesday becomes the monday
[18:56] <nevcairiel> following that logic, we should all just troll around the web for a living
[18:57] <gnafu> nevcairiel: It works for him.
[22:23] <CIA-17> ffmpeg: 03Carl Eugen Hoyos 07master * r7b4e1466d4 10ffmpeg/ (ffmpeg.c libswresample/swresample.h): Exit ffmpeg if swr_convert() fails.
[22:38] Action: beastd wonders if Libav-devel moderation queue gets regularly reviewed
[23:10] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * rf1808e3048 10ffmpeg/libpostproc/postprocess.c: 
[23:10] <CIA-17> ffmpeg: libpostproc: add pp=help support to print the help text via av_log()
[23:10] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[00:00] --- Tue Feb 28 2012