[Ffmpeg-devel-irc] ffmpeg.log.20111219

burek burek021 at gmail.com
Tue Dec 20 02:05:03 CET 2011


[00:02] <norbert_> hi all, I'm using Ubuntu 11.10 and I'm trying to screen record, but the colors are (gamma is) to bright when I playback; I tried both xvidcap and ffmpeg for recording, and vlc, totem and mplayer for playback... any ideas how I can fix this?
[00:04] <burek> ShadesEdge, ./configure --help
[00:04] <burek> also read README and INSTALL if present
[00:05] <burek> norbert_, did you try ffmpeg with X11gran
[00:05] <burek> grab
[00:05] <ShadesEdge> burek: ./configure --help won,t help me
[00:05] <norbert_> yes, with -f x11grab
[00:06] <norbert_> ffmpeg -f alsa -ac 2 -i pulse -f x11grab -r 25 -s 100x100 -i :0.0+100,100 -acodec pcm_s16le -vcodec libx264 -preset ultrafast -crf 0 output.mkv
[00:06] <norbert_> is what I'm running
[00:06] <burek> ShadesEdge
[00:07] <burek> norbert_
[00:07] <burek> can you please use pastebin.com, to show your command line and its output?
[00:07] <ShadesEdge> sudo apt-get install yasm, good luck with that ¬¬
[00:07] <norbert_> burek: will add output to pastebin; command line I already mentioned above
[00:09] <norbert_> burek: http://pastebin.com/WDigJv2A
[00:10] <burek> crf 0 ?
[00:12] <norbert_> probably the integer for one of "VPX_CBR", "VPX_CQ", "VP8E_SET_CQ_LEVEL"
[00:12] <norbert_> I don't know why it's there, the command line options were suggested by someone else
[00:12] <norbert_> should I remove it
[00:12] <norbert_> ?
[00:13] <burek> try -crf 20
[00:13] <burek> its an option for libx264
[00:13] <burek> you used kinda contradicting options
[00:14] <burek> anyway try with -crf 20
[00:14] <norbert_> I tried 20, colors are still bright
[00:14] <norbert_> I will upload screenshots of what I'm recording and what the result is
[00:14] <burek> ok
[00:15] <norbert_> oh wait
[00:15] <norbert_> I did mplayer -vo png output.mkv and the png's look ok
[00:16] <norbert_> so I tried mplayer -vo x11 output.mkv and then it plays ok
[00:16] <norbert_> so ffmpeg is not the problem
[00:16] <norbert_> sorry about that
[00:16] <burek> np :)
[00:16] <norbert_> anyways, thanks for helping
[00:16] <burek> :beer: :)
[00:17] <norbert_> hehe
[02:43] <freezway> how can I try to fix broken audio files (aac, .m4p)
[02:43] <freezway> i cant seem to play them
[02:48] <freezway> ffmpeg tells me [aac @ 0xed3900] channel element 0.7 is not allocated when i do ffmpeg -i filename
[02:49] <freezway> is it a corruption or drm issue?
[04:58] <Shimmy> Can anyone help me please: https://ffmpeg.org/trac/ffmpeg/ticket/747
[06:28] <freezway> Shimmy, i can try, probably cant tho
[06:28] <freezway> nope, cant help
[08:55] <davecotter> Q: what character encoding should the file paths be in?  i hope utf8?
[08:56] <davecotter> anybody there?
[08:57] <Wizord> davecotter, it depends on your locale
[08:57] <Wizord>  _  
[08:57] <davecotter> okay, so it's in the current, multibyte locale?
[08:57] <davecotter> mm'k.  so, like for us that would be WindowsLatin1, right?
[08:58] <Wizord> I don't know
[08:58] <Wizord> and by locale I mean the locale, not localization
[08:58] <davecotter> ANSI codepage 1252
[08:58] <davecotter> yeah,sure. k
[08:58] <Wizord> what system are you using?
[08:59] <davecotter> any win: windows 7, xp, vista& and could be any language.  
[08:59] <Wizord> /sigh
[08:59] <davecotter> user name might be in korean or something
[08:59] <davecotter> want to make sure the paths will work in any case
[08:59] <Wizord> then, most likely, it's utf16-le
[09:00] <davecotter> i could not find any definitive doc on it
[09:00] <Wizord> winapi docs on openfileW funtions
[09:00] <davecotter> mmm'k
[09:00] <davecotter> is there like a "-w" command for ffmpeg.exe then?
[09:01] <davecotter> for wide chars?  that'd be swell
[09:01] <davecotter> then i wouldn't have to worry about locale / encoding
[09:02] <Wizord> /sigh
[09:02] <davecotter> sorry, maybe i wasn't clear: this is for the file paths passed to ffmpeg via &..  DOH
[09:02] <davecotter> sorry i just answered my own question
[09:02] <davecotter> was going to say "via ShellExecuteEx", duh, which takes wide chars.
[09:03] <davecotter> <-- dork
[09:05] <davecotter> but thanks :)
[11:15] <fling> ffmpeg -i ../World.mp4 -f image2 -c:v ppm fling%d.ppm
[11:15] <fling> Unrecognized option 'c:v'
[11:15] <fling> what am i doing wronjg?
[11:15] <fling> followed that > http://ffmpeg.org/faq.html
[11:15] <sacarasc> Which version of ffmpeg are you using?
[11:18] <fling> sacarasc: 0.7.7
[11:18] <sacarasc> That's too old of a version to use the -c:v thing, I think. You can use -vcodec ppm, though.
[11:19] <fling> thanks
[11:19] <fling> what version should i use? 0.9?
[11:21] <sacarasc> It's generally advised to use the latest git version.
[11:24] <fling> sacarasc: thanks, will emerge it :p
[11:29] <fling> sacarasc: how to add padding zeros? fling%d.ppm
[11:33] <Tjoppen> fling: number between % and d
[11:33] <Tjoppen> and zero
[11:33] <Tjoppen> like: %09d
[11:33] <fling> thx
[11:34] <fling> if i need 000-999 > %03d ?
[11:34] <Tjoppen> yes
[11:35] <Tjoppen> number start at 1 I think, but elsewise correct
[11:35] <Tjoppen> *numbers. so 001-999
[11:37] <iive> zero is a number too, so 000-999
[11:39] <Tjoppen> yes, but ffmpeg counts from 1
[11:46] <kralikba> Hi
[11:47] <kralikba> I'm trying to compile ffmpeg on arm but I'm getting errors of an undefined reference to log2f and exp2f
[11:50] <kralikba> ld is complaining
[11:58] <fling> sacarasc: Tjoppen: thanks, ffmpeg rocks!
[12:55] <fb0x> hi, trying to compile latest ffmpeg 0.9 and i get "ERROR: libx264 version must be >= 0.118." i'm compiled that version of x264 but i still get error
[12:55] <fb0x> what is wrong?
[14:31] <cbreak-work> so, I finally managed to get ffmpeg to compile and link for iOS/iOSSimulator
[14:31] <cbreak-work> the major part was instrumenting ./configure with the correct parameters, and then lipoing the library into a fat binary
[14:32] <cbreak-work> but I also had to remove some asm so that apple's gcc can compile the code
[14:34] <cbreak-work> how would I submit that patch for consideration? I am sure that commenting out code is not the correct solution, but my knowledge of arm asm is non-existent, so I can't do anything better
[14:38] <Alasdairrr> cbreak-work: i'd guess, mail the mailing list with a patch attached or a link to a github/bitbucket commit
[14:59] <someone-noone> hello! Is ffmpeg "friendly" with OMAP processors? Hardware acceleration ?
[16:39] <rizlah> Hello. How powerful would a computer need to be to compress a DVD quality video in a reasonable time-frame?
[16:40] <rizlah> I would need to recompress some avis of DVDs but not sure what flags to use and don't want to wait four or five hours per video
[16:44] <sprayd> hello michaelni 
[16:44] <sacarasc> 500 hours? :o That's crazy...
[16:47] <sprayd> michaelni, i've been trying to contact them (peoples and committees) but I'm not getting any reply... even an acknowledgment that it may not be possible to fulfill my requests and thus redirect me or advise me something else..
[16:48] <sprayd> and now again some minutes are left until the deadline of the task michaelni 
[16:48] <brbrbr> rizlah please define you "reasonable time-frame"(or in FPS) and "DVD quality" :) "in ganeral", studios[even home-made/based]use render-frams to do its both reasonably quick and accurate
[16:48] <brbrbr> if you insista on single-case solution, try pick Opteron-based workstation with 1-2 16x core CPU's
[16:48] <sprayd> i fear i've not done enough michaelni 
[16:49] <brbrbr> and stick with x264 codec[versions above 2120 perform quite nice :]
[16:50] <rizlah> Well, DVD quality would generally be a one hour video (AVI container) of +/-500MB. I'm wondering if I can compress that down to maybe 250 or so without losing much picture / audio quality
[16:51] <rizlah> Reasonable time frame being 30 mins per video
[17:02] <michaelni> sprayd, can you upload the mails that you did send?
[17:03] <sprayd> okay
[17:04] <sprayd> they are in pdfs okay?
[17:04] <michaelni> any format you like
[17:07] <brbrbr> heavily-tuned "high" AVC profile can do that, esp in 10-bit mode, but hardly i can call "DVD quality", in 250mb case :/
[17:10] <rizlah> hmm
[17:10] <rizlah> Will look into that profile.
[17:11] <sprayd> so michaelni i have uploaded them
[17:13] <sprayd> and as a not michaelni  the dates on the one of the letter may perplex you but it is because my computer's clock's always displays wrong time on computer startup so...
[17:27] <brbrbr> rizlah check doom10.org forum for specific details about x264 tuning, where Dark Shikari moved with contributors after conflict or in #x264 channel
[17:28] <brbrbr> genrally, "built-in" presets was ok. both in Handbrake and [if you on Windows]Avidemux 2.55/2.6 , MedicaCoder
[17:29] <michaelni> sprayd, if you want to wait longer for a reply, i can extend the deadline ?
[17:29] <joecool> hi, I have an mp3 that was recorded with a 16khz mic and was saved as a 44.1khz file, it plays back too quickly, is there anyway to resample it to 16khz and get the correct speed?
[17:31] <sprayd> see michaelni waiting means i can't do other tasks, so that is a problem for me. What about I see this task done when I get the mails even though I am not engaged with the tasks? that would be okay for me and maybe you...
[17:43] <vega13> sup
[17:44] <vega13> what does av_packet_split_side_data do
[17:45] <vega13> i try to feed a packet in avcodec_decode_video2 then it segmentation fault @ avpacket_split_side_data
[17:52] <vega13> what could cause this
[17:52] <vega13> i'm sending the packet over network
[17:53] <vega13> then on the other side I try to decode with avcodec_decode
[17:58] <swkide> Greetings! I am trying to figure out how to use the new segmenter with .m3u8 output. Is there any documentation how to use it? I am kind of stuck
[18:02] <michaelni> sprayd, i can unassign it and if you get a reply and if noone else successfully finished it iam happy to assign it to you again and mark it as done
[18:12] <sprayd> okay agreed michaelni 
[18:13] <sprayd> please unassign it michaelni 
[18:14] <sprayd> but can i get bonus points for the existing work?
[18:14] <sprayd> or no michaelni ?
[18:21] <michaelni> sprayd, i have no idea what bonus points are ...
[18:22] <michaelni> sprayd, unassigned
[18:30] <sprayd> okay michaelni 
[18:32] <sprayd> michaelni, i saw there were some bonus points for some tasks but it's okay
[18:45] <japa-fi> Fedora 14 user with FFmpeg version 0.6.3-rpmfusion. trying to encode to .mp4 (libx264) by specfying quality (-qscale 25). Yet, when I start ffmpeg using a preset which does not contain bitrate parameter, ffmpeg reports that bitrate has been set to 200kbs.
[18:46] <japa-fi> that will result to a file with quality somewhere around 40 instead of what I specified ( 25 )
[18:46] <japa-fi> how can I "unset" the bitrate
[18:47] <sacarasc> Upgrade to a newer versions and it should be good...
[18:49] <japa-fi> Darn, need to either find another repo or compile myself
[18:51] <japa-fi> Wow.. use the google.. atrpms.net ffmpeg-0.8.7-50.fc14.x86_64.rpm  
[18:51] <japa-fi> Seems that my current version is bit off...
[18:56] <japa-fi> Arh, I never use yum -y, except now.
[18:57] <japa-fi> And It removed crapload of other stuff
[19:09] <twid> I have my costume program to convert file from one format to another format, But when i convert file and play then after it speed is seems like 4X, How can i make it slow? 
[19:11] <twid> I mean which structure specify the playing speed? Thx..
[19:16] <japa-fi> Even the new ffmpeg seems to ignore my settings: No matter what I set as quality, result is always the same: http://fpaste.org/V6S6/
[21:05] <MDijkstra> ie, m/win close
[21:05] <MDijkstra> oops
[21:20] <dericed> hi all, i updated my ffmpeg-git today and some of my mxf handling scripts are working very differently, leaving the output file truncated. I'm running the exact same command as last week on the same file.
[21:22] <dericed> ffmpeg from dec. 14th gives this output: http://pastebin.com/b656EB9F and provides a file of the right duration
[21:23] <dericed> ffmpeg from today gives this output http://pastebin.com/dE4cvhyA and outputs a file about 10% of what is expected (both outputs run the same command on the same file, just a git update is the difference)
[21:39] <cbsrobot> Tjoppen: ^
[21:40] <cbsrobot> dericed: maybe Tjoppen can help you out - wait for him to return ...
[21:40] <dericed> cbsrobot: thanks, i've been noticing the overlap of my regression and his patch
[21:44] <Tjoppen> dericed: upload a sample
[21:44] <Tjoppen> not sure what's wrong - 10x fewer frames?
[21:52] <dericed> i tried on a few files of different lengths, and the output was about 10%
[21:52] <dericed> Tjoppen: you want an 11GB sample? or some amount of the head of the file?
[21:56] <Tjoppen> lay it on me
[21:56] <Tjoppen> I deal with file ~ that size at work quite often
[21:57] <Tjoppen> *files. as long as the host is decent
[22:11] <dericed> Tjoppen: ok i'll look for a file clear to upload that has the same qualities
[22:11] <dericed> Tjoppen: These files are straight off XDCam disc btw
[22:15] <cbsrobot> Tjoppen: btw the latest patches didn't resolve my issue
[22:17] <cbsrobot> if you have time,please check: http://ijghv1.dl4free.com/
[22:17] <cbsrobot> it says: [mxf @ 0x10300c000] failed to find absolute offset of 1ef94 in BodySID 1 - partial file?
[22:18] <cbsrobot> and in September I was able to play this files in ffplay - now I can't
[22:18] <Tjoppen> this is relevant to my interests as we have to support xdcam at work
[22:20] <cbsrobot> and if I try to decode the file I can only decode one frame
[22:20] <cbsrobot> the file above is a test file for dcp
[22:21] <cbsrobot> it should be a timecode running for about 2min 
[22:22] <cbsrobot> avconv version N-33762-g284aa07 decodes it without problems
[22:24] <cbsrobot> Tjoppen: should I fillout a bugreport ?
[23:15] <Tjoppen> cbsrobot: sure
[00:00] --- Tue Dec 20 2011


More information about the Ffmpeg-devel-irc mailing list