[Ffmpeg-devel-irc] ffmpeg.log.20120626

[burek]

[00:49] <Freakshow> < saste: DelphiWorld: rtmp native support in ffmpeg is not very good at the present moment> in what way is rtmp support in ffmpeg 'not very good'?
[00:50] <saste> Freakshow: in the sense that not all protocols are supported, and that people had some issue in the past, so they're usually better served by the librtmp implementation, which is more tested and generally more used
[00:51] <saste> i don't know the current status of affairs though
[00:51] <Freakshow> and holy shit... burek / ubitux I just read through this evenings discussion re: trac vs. mediawiki, I'm sorry... but I seriously lol'd
[00:51] <saste> maybe it's improved in the meaningwhile...
[00:51] <Freakshow> true that... but generally speaking, isn't librtmp more for handling encrypted output rather than standard output?
[00:52] Action: llogan waits for IRC log
[00:52] <Freakshow> rtmpt, etc.?
[00:52] <Freakshow> yes llogan: grep for HerbertPumpkin
[00:52] <Freakshow> :D
[00:53] <llogan> nice nick. must be full o' gold.
[00:54] <Freakshow> my favorite:
[00:54] <Freakshow> [07:11am] ubitux: then just report typo issues by opening an issue in the trac
[00:54] <Freakshow> [07:11am] ubitux: and complete the doc in the wiki
[00:54] <Freakshow> [07:12am] HerbertPumpkin: what's a trac?
[00:54] <Freakshow> [07:12am] ubitux: it's a facebook group
[00:54] <Freakshow> sorry for that spam....
[00:59] <Freakshow> seriously though... I'm curious if anyone has had a chance to review:
[00:59] <Freakshow> https://ffmpeg.org/trac/ffmpeg/ticket/1446
[01:00] <Freakshow> how would I go about trying to get some eyes on this issue?
[01:00] <llogan> is that your ticket?
[01:02] <burek> well, to be honest, if nobody is interested to make the docs better without raping the free time of developers, then we did waste our time today on that discussion about wiki and docs
[01:05] <heimlich> hi guys.. if i have a quicktime movie with timecode... is there a way to extract specific frames from the quicktime movie starting with timecode xxxx and ending with timecode yyyy...when I try to use the -ss option, I always get invalid duration specification for ss
[01:09] <Freakshow> llogan: yes, that's mine
[01:10] <Freakshow> burek: to be fair... I wasn't saying it was a waste of time. I think that there were some valid points brought up in that conversation
[01:12] <burek> heimlich, can you please use pastebin.com, to show your command line and its output?
[01:12] <burek> Freakshow, well yes, but we seriously offtopiced this channel :D
[01:12] <heimlich> sure.. one sec
[01:13] Action: Freakshow nod
[01:14] <Freakshow> no one else was talking anyway
[01:14] <heimlich> burek: http://pastebin.com/94mSGWqQ
[01:16] <heimlich> weird.. didnt seem to paste all
[01:16] <burek> heimlich, try -ss 02:29:04.200 instead of -ss 02:29:04:20
[01:18] <heimlich> burek: same result: http://pastebin.com/2brseNf2
[01:18] <burek> heimlich, note the dot "."
[01:18] <burek> instead of a column ":"
[01:18] <heimlich> ungh.. sorry.. one sec
[01:19] <heimlich> yeah.. that worked..
[01:22] <heimlich> now end timecode
[01:28] <burek> * llogan waits for IRC log -> see live logs here (if you need): http://ffmpeg.gusari.org/irclogs/
[01:29] <heimlich> burek: http://pastebin.com/6LTKUGva
[01:29] <heimlich> btw.. thanks
[01:29] <burek> heimlich, can you please use pastebin.com, to show your command line and its output?
[01:29] <burek> uncut, entire output
[01:29] <Freakshow> llogan: do you have any thoughts on that ticket? I'm just trying to get some movement on it if possible
[01:30] <heimlich> here yo go: http://pastebin.com/EwQH6YKi
[01:31] <burek> heimlich, your input file's duration is: Duration: 00:00:01.20
[01:31] <burek> and you are seeking to -ss 02:29:04.200 (2 hours, 29 minutes, 4 seconds and 200 mseconds)
[01:31] <burek> :)
[01:32] <heimlich> well.. the starting timecode of the file is 02:29:04:02     my initial goal, was to extract frames from a long quicktime, using certain timecode segments
[01:33] <burek> I see
[01:35] <burek> well, in stream #0.1 there is "timecode" metadata
[01:35] <burek> you might get it using ffprobe or ffmpeg (parse the output of ffmpeg -i inputfile)
[01:35] <burek> and subtract from the creation time maybe
[01:36] <burek> (which is also in metadata)
[01:36] <burek> I didn't say it right.. what I meant is to subtract from the first timecode (of the first segment) you have
[01:37] <burek> so that you get the actual time into the video
[01:38] <burek> -ss can only handle time in seconds or in "hh:mm:ss[.xxx]" form
[01:39] <Freakshow> brb
[01:39] <llogan> Freakshow: sorry. i've never used rtsp output.
[02:12] <alyawn> using libavformat, how can I specify a specific output codec rahter than taking the AVFormatContext default?
[02:13] <alyawn> think mpegts format with an h264 stream, rather than the default mpeg2video
[02:48] <Nickname123> Can anyone help me get ffmpeg compiling on my windows mingw setup?  Ffmpeg configure gives me the following error: "ERROR: libx264 not found"  See http://pastebin.com/qTwYas8t for directory layout and command line
[02:49] <Nickname123> oh i think i made a copy and paste error
[02:53] <Nickname123> http://pastebin.com/xaMLLDUa
[02:57] <Nickname123> I tried to fix it for the / paths that msys uses... still no luck: http://pastebin.com/sEwjDwfA
[03:01] <grepper> I think you need to compile and install x264 first, at least that is how it is done on linux afaik
[03:02] <grepper> then if need be give the path the the installed libraries and headers
[03:02] <Nickname123> I did but the msys / mingw search paths are throwing me off
[03:02] <Nickname123> so I tried to explicitly list them
[03:02] <Nickname123> oooh
[03:02] <Nickname123> I think i got it
[03:02] <grepper> I see.  Sorry, I don't use windows so dunno
[03:02] <Nickname123> started over on my configure line
[03:03] <Nickname123> Found http://stackoverflow.com/questions/8812827/build-ffmpeg-with-x264-for-android
[03:03] <grepper> did you do a make distclean or at least a make clean first ?
[03:03] <Nickname123> and it made me think to use relative paths for some reason lol
[03:03] <grepper> ah
[03:03] <grepper> okay, as long as you solved it
[03:04] <Nickname123> Thanks
[03:04] <Nickname123> I think the -I and -L options were choking on C:/
[03:04] <Nickname123> but a relative path woked
[03:04] <Nickname123> *worked
[03:04] <llogan> Nickname123: there is this: https://ffmpeg.org/trac/ffmpeg/wiki/MingwCompilationGuide
[03:05] <llogan> but it's not as comprehensive as i'd like
[03:05] <Nickname123> I read through that first
[03:05] <Nickname123> It helped me find all the extra libs. Ty
[03:06] <llogan> i would prefer that it had actual step-by-step commands for users to copy and paste.
[03:07] <llogan> (un)fortunately i don't use mingw
[03:08] <Nickname123> This is my first time compiling on Windows instead of compiling on Linux for Windows
[03:19] <alyawn> figured mine out... I needed to re-assign the codec, not just the codec_id after calling avformat_new_stream
[03:27] <Nickname123> it built =)
[06:01] <joecool> /usr/lib/gcc/x86_64-pc-linux-gnu/4.7.1/../../../../x86_64-pc-linux-gnu/bin/ld: libavcodec/x86/fft_mmx.o: relocation R_X86_64_PC32 against undefined symbol `memcpy@@GLIBC_2.14' can not be used when making a shared object; recompile with -fPIC
[06:01] <joecool> /usr/lib/gcc/x86_64-pc-linux-gnu/4.7.1/../../../../x86_64-pc-linux-gnu/bin/ld: final link failed: Bad value
[06:02] <joecool> i run a no-multilib system
[06:02] <joecool> why is this happening?
[07:32] <ubitux> Freakshow :)
[07:53] <alyawn> am I supposed to re-init my AVPacket before I call avcodec_encode_video2()?
[10:46] <sgfgdf> hello, guys! i'm trying to convert a video for an iPad. here is my attempt: my preset file -- http://cl.ly/2G1v0y3n0M292G2g053V , log -- http://cl.ly/0B1o1P300F0z372i1g2z . anyone can help, please?
[10:46] <Vardan> hi all
[10:46] <Vardan> people how can I get my ffmpeg library version number via code?
[10:51] <Mavrik> Vardan: there's a "version.h" header defined with versions for each library
[10:51] <Mavrik> also macros that give you better output
[10:52] <Mavrik> sgfgdf: your preset sets invalid audio settings
[10:54] <sgfgdf> Mavrik, which one is incorrect? ar, ab, ac?
[10:55] <Mavrik> considering this line:     Stream #0:1(eng): Audio: aac, 48000 Hz, stereo, s16, 1200 kb/s
[10:55] <Mavrik> it's ar
[10:56] <Mavrik> since ffmpeg thinks you're setting audio bitrate to 1200kbps :)
[10:58] <sgfgdf> Mavrik, so i should change b= or ar= ? it's strange because i found the command from here -- http://www.ioncannon.net/meta/1040/how-to-create-ipad-formatted-videos-using-handbrake-or-ffmpeg/ and it appears to work probably for the author. is it my mistake or it is wrong from the source i get it?
[10:59] <Mavrik> b is video bitrate
[10:59] <Mavrik> nothing wrong with it
[11:02] <ssuclinux> kg2
[11:07] <sgfgdf> Mavrik, i put b above ab, because it appears to override ab. now it is working. thank you!
[11:10] <burek> sgfgdf, -ab controls audio bitrate and -ar controls the sampling frequency (44.1KHz)
[11:10] <burek> or 48KHz in your case
[11:10] <burek> so try -ab 96k
[11:12] <burek> btw, your -ab gets overriden most probably because you use -b to set video bitrate
[11:13] <burek> but the syntax changed a little bit in latest ffmpeg
[11:13] <burek> -b:v is used for video bit rate and -b:a for audio bitrate
[11:13] <burek> setting just -b affects both video and audio
[11:14] <sgfgdf> burek, can i use b:v=1200k and b:a=128k in the ffpreset file rather than command arguments?
[11:14] <burek> yes
[11:15] <sgfgdf> burek, okay so better use them to eliminate confusions. thank you!
[11:16] <burek> :beer: :)
[11:19] Action: sgfgdf sends http://blogsensebybarb.files.wordpress.com/2012/04/beer.jpg to burek. hope i don;t have so much future questions :)
[11:20] <burek> :excited:
[11:23] <sgfgdf> burek, before drinking them you should count them and think again :)
[11:25] <burek> one beer at a time :}
[11:33] <varaderoguy> morning all
[11:33] <varaderoguy> ffmpeg compiling fun....
[11:33] <varaderoguy> need some help and advice
[11:35] <varaderoguy> So - I'm running the standard Centos 5.7 RPM version of FFMPEG 0.6.5
[11:35] <varaderoguy> I'm trying to build the night build with the same arguments.....
[11:35] <varaderoguy> This is not as simple as it sounds.....
[11:37] <varaderoguy> The good news is that just using the standard ./configure; make and make install on Centos, things work just fine
[11:37] <varaderoguy> Its just that libraries are not being included
[11:37] <JEEB> better not to do it with exactly the same arguments, distros tend to just EnableEverything while you usually have a more limited set of things that you need (given the fact that 99% of the decoders are inside libavcodec)
[11:37] <JEEB> but look at it that you basically have everything LGPL enabled by default
[11:37] <JEEB> and then you enable stuff that you need (libx264., libfaac, for example)
[11:38] <JEEB> and yes, the libraries on an old centos are probably old
[11:38] <JEEB> libfaac is one of those libraries that was pretty much left to rot so that might be up-to-date
[11:38] <varaderoguy> JEEB: lovely - well that good news....I will need libx264 though (for transcoding to flv file format)
[11:38] <varaderoguy> ...and the ac3 audio stack
[11:38] <rainmaker1> Hi, is it possible to output the same stream to multiple destinations?
[11:38] <JEEB> ac3 decoder and encoder are included in libavcodec
[11:39] <JEEB> ok, you then most probably need a current yasm and current x264
[11:39] <varaderoguy> Varaderoguy HUGS JEEB
[11:39] <varaderoguy> (with applogies to everybody) - ffmpeg -version give me.....
[11:39] <varaderoguy> libavutil      51. 62.100 / 51. 62.100 libavcodec     54. 29.100 / 54. 29.100 libavformat    54. 11.100 / 54. 11.100 libavdevice    54.  0.100 / 54.  0.100 libavfilter     3.  0.100 /  3.  0.100 libswscale      2.  1.100 /  2.  1.100 libswresample   0. 15.100 /  0. 15.100
[11:40] <JEEB> go to your home dir, do 'mkdir builds' and move into that folder
[11:40] <JEEB> wget http://www.tortall.net/projects/yasm/releases/yasm-1.2.0.tar.gz && tar xvf yasm-1.2.0.tar.gz && cd yasm-1.2.0
[11:41] <JEEB> do you want to install to some specific place or /usr/local, which is often the default?
[11:41] <burek> rainmaker1, yes: ffmpeg -i input (output options) output1 (output options) output2 (output options) output3 ...
[11:42] Action: JEEB pokes varaderoguy 
[11:42] <varaderoguy> jeeb: sorry - I was just fetching yasm
[11:42] <JEEB> k
[11:43] <JEEB> then you basically ./configure and set --prefix=/where/you/want/to/install
[11:43] <rainmaker1> burek: tnx :)
[11:43] <varaderoguy> jeeb: /usr/local will be fine
[11:43] <JEEB> k
[11:43] <burek> :beer: :)
[11:43] <JEEB> then you configure, make and make install -- and then you can check if yasm --version gives you 1.2.0
[11:44] <JEEB> I guess you have git already installed so the url for x264 would be git://git.videolan.org/x264.git
[11:45] <varaderoguy> got git installed....just doing a make install on yasm
[11:45] <JEEB> you clone it and configure it with --enable-static
[11:45] <varaderoguy> lovely - yasm is 1.2.0
[11:45] <JEEB> and set prefix accordingly (but I think /usr/local is the default)
[11:46] <varaderoguy> hummm.....got a slight problem in so much that I can do an https, but not git protocol
[11:46] <JEEB> http://git.videolan.org/git/x264.git is on http://git.videolan.org/gitweb.cgi?p=x264.git;a=summary
[11:46] <varaderoguy> jeeb: **
[11:48] <JEEB> basically --enable-static will build a static library. They're easier to deal with in this context so I recommend them unless you know your ways around ldconfig and friends
[11:48] <JEEB> (x264 with a default config will only build the command line encoder tool)
[11:48] <JEEB> also you should naturally check that asm is enabled in the configuration screen that comes up after you run the configure
[11:48] <varaderoguy> okay....just having fun and games with git....
[11:49] <JEEB> I'm actually waiting for your compiler to fail at some point because I remember many centos 5 users stumbling somewhere along the line...
[11:49] <varaderoguy> jeeb: stupid Q
[11:51] <varaderoguy> jeeb: just doing make on x264
[11:52] Action: JEEB is left a bit irritated by a non-asked question
[11:52] <varaderoguy> jeeb: humm...that passed off without issue
[11:52] <varaderoguy> jeeb: okay - what didn't I ask?
[11:53] <JEEB> <varaderoguy> jeeb: stupid Q <-
[11:54] <varaderoguy> jeeb: sorry - I have a nasty habit of not finishing sentences....and then figuring the problem out for myself....my fault
[11:55] <varaderoguy> okay - so I'm now ready to rock.....
[11:55] <varaderoguy> It is a just a case of reconfiguring ffmpeg and letting it pick up the libraries?
[11:56] <JEEB> ok, so I guess you got libx264 installed? and it had asm enabled?
[11:57] <JEEB> then ./configure --enable-gpl --enable-libx264 and see if it spots the newer library in /usr/local or the older one from the package management
[11:57] <varaderoguy> jeeb: I can confirm that asm is Enabled in x264
[11:57] <JEEB> also, if you hadn't gotten git before, how exactly did you get the source for ffmpeg?
[11:58] <varaderoguy> no - it was my fault and the fact that I'm trying to type too quickly.....
[11:58] <varaderoguy> I really apprechiate your help JEEB
[12:02] <varaderoguy> jeeb: ./configure looks good; now just trying the make; make install
[12:02] <JEEB> you could check the contents of config.log for which it actually picked up to be sure :)
[12:08] <varaderoguy> ls
[12:10] <varaderoguy> jeeb: That is fantastic....many thanks Jeeb
[12:34] <varaderoguy> Chaps: another question: does the nightly builds come with a man file for ffmpeg?
[12:37] <varaderoguy> ls
[12:51] <varaderoguy> okay: the -metadata function is still broken in the latest build :-(
[12:52] <varaderoguy> has anybody managed to get ffmpeg to write metadata to the file header for an FLV file?
[12:52] <kcm1700_> does anyone know what's YUV440 format? I was searching for the information but I don't see any.
[12:54] <varaderoguy> kcm1700: its a broadcast video format
[12:55] <varaderoguy> kcm1700: Its to do with the number of bits of information in terms of lumance, and chromance
[12:56] <kcm1700_> varaderoguy: thank you. Is there any web page(or search keyword) to learn the format?
[12:56] <varaderoguy> kcm1700: try this for starters: http://en.wikipedia.org/wiki/YUV
[12:57] <kcm1700_> ah..
[12:57] <kcm1700_> Luminance/chrominance systems in general <- this section has information, I missed it.
[12:57] <kcm1700_> thank you
[13:01] <kcm1700_> Ah, is YUV440 reordered format of YUV422?
[13:03] <zap0> probably.
[13:04] <varaderoguy> kcm_1700 - in essence, although you will find that there is no V sampling information, which for the purposes of colour correction can be an issue....but in terms of most data packets these days, in don't really need the V packets....its saves space
[13:07] <kcm1700_> thanks for the helps, zap_0, varaderoguy!
[13:12] <zap0> glad my huge contribution was recognized ;)
[14:18] <AlRazi> I am trying to convert mp3 -> aac+, the resulting file has 7000+ hours duration on itunes, i tried running qt-faststart on it but to no avail
[14:19] <AlRazi> I tried using ffmpeg to produce a wav file, and then convert the wav file to m4a using faac, it works fine, but faac doesn't support HE-AAC
[14:47] <AlRazi> I am trying to convert mp3 -> aac+, the resulting file has 7000+ hours duration on itunes, i tried running qt-faststart on it but to no avail, I tried using ffmpeg to produce a wav file, and then convert the wav file to m4a using faac, it works fine, but faac doesn't support HE-AAC
[14:56] <rainmaker1> Hi, is there any tool to help me find the atom size?
[15:02] <rainmaker1> found :)
[15:17] <AlRazi> I am trying to convert mp3 -> aac+, the resulting file has 7000+ hours duration on itunes, i tried running qt-faststart on it but to no avail, I tried using ffmpeg to produce a wav file, and then convert the wav file to m4a using faac, it works fine, but faac doesn't support HE-AAC
[15:35] <Spideru> Hi. I would to connect ffplay to an rtp stream created with ffmpeg. On linux ffplay -f rtp rtp://<ip>:<port> works, on window Xp (with different prebuilds) nope. What i am missing? Thank you
[16:49] <jesk> adding/modifying metadata in mp4 is a pita
[16:49] <jesk> always rewrite from the scratch
[16:50] <jesk> how can I modify a mp4 so that is has enough free atom space, something around 5MB would make sense
[16:53] <Na_Klar> Can I affect the PNG pre-filter with ffmpeg? I want to decode a movie to a png image sequence but want to use NONE pre-filter as specified in the PNG defination.
[16:58] <Diogo> hi one question please..
[16:58] <Diogo> i need to comunicate with ffser using ffmpeg
[16:58] <Diogo> i'm using this command: /servers/ffmpeg/bin/ffmpeg -i HD.mp4 http://localhost:8090/test.asf
[16:59] <Diogo> but it is to fast...
[16:59] <Diogo> i'm doing something wrong?
[17:28] <xero-exez> First things first: "Thanks again for FFMPEG!"
[17:32] <xero-exez> Would someone be so kind to check my syntax. It transcodes all input files except ProRes.
[17:32] <xero-exez> Error: Error setting profile baseline
[17:32] <xero-exez> http://pastebin.com/Nw65Ck4u
[17:36] <sacarasc> You have -profile and -vprofile...
[17:36] <sacarasc> Also, if you had read...
[17:36] <sacarasc> x264 [error]: baseline profile doesn't support 4:2:2
[17:38] <xero-exez> @sacarasc Thank you for your response! I tried with one of -profile and -vprofile there is no differance
[17:38] <sacarasc> But you're trying to use 4:2:2 with a profile that doesn't support it.
[17:40] <xero-exez> @sacarasc Allright I'm reading about YUV an downsampeling now....If i understand right the color input scheme cannot be downscaled?
[17:43] <gnarface> hey guys i'm trying to transcode some video (a decoded VOB ripped from a commercial dvd i own) and though it successfully produces video with little complaint, the apparent playback of resulting video is unsteady
[17:43] <gnarface> it speeds up and slows down
[17:43] <gnarface> on an even interval
[17:43] <gnarface> like it plays real fast for a half second then slows down to less than 1fps, then plays a bunch real fast, etc... nice even cycle
[17:44] <gnarface> same behavior with single and multi-pass encoding
[17:45] <gnarface> same behavior with multiple audio codecs
[17:45] <gnarface> http://pastebin.com/hyeheLaf
[17:45] <xero-exez> @sacarasc Thanks for putting me on the right track.... I added -pix_fmt yuv420p
[17:45] <gnarface> here's an example run; i just tried adding -vsync 2
[17:46] <delicado> hi guys what function is available in ffmpeg library to convert the YUV format to RGB? i tried looking in avutils and avpicture for functions that start with "convert" but i did not find any.
[17:46] <xero-exez> @delicado -pix_fmt yuv420p      :)
[17:49] <xero-exez> @gnarface try to lower your freames-per-keyframe -g 25
[17:49] <delicado> xero-exez: but im using the library. :(
[17:49] <gnarface> xero-exez: same behavior if i don't specify any at all :(
[17:50] <Mavrik> delicado: that's the task of libswscale
[17:50] <Mavrik> delicado: look for sws_scale and other sws_* parameters
[17:51] <Mavrik> er, functions, not parameters, sorry :)
[17:51] <gnarface> xero-exez: trying it anyway
[17:51] <gnarface> xero-exez: think this could be relevant? "[mpeg @ 0x20d4ec0] max_analyze_duration reached"
[17:51] <delicado> thanks Mavrik.
[17:52] <xero-exez> @gnarface Don't listen to me I'm just a n00b that had a question befor yours...and trying to do something back
[17:52] <cen|3> could someone explain to me how to convert an asf to gif using ffmpeg/avconv??? I keep getting two errors - one stating that the gif only handles rgb24 pix form (use -pix_fmt rgb24 (but errors when I try to do this)) and another error stating that it could not write header for output file...
[17:54] <Mavrik> cen|3: make sure you're using ffmpeg and not avconv and then paste the whole output with command line.
[17:55] <cen|3> what I type is: avconv -i file.asf file.gif
[17:55] <cen|3> k - hold on...
[17:58] <cen|3> pastebin.com/CveXpEW0
[17:58] <cen|3> http://pastebin.com/CveXpEW0
[17:59] <cen|3> I'd like to make the gif resolution in the end much smaller as well (like 640xXXX)
[18:06] <gnarface> does anyone happen to know if this implies a bug in ffmpeg?  "Seems stream 0 codec frame rate differs from container frame rate: 59.94 (60000/1001) -> 59.94 (60000/1001)"
[18:08] <cen|3> any idea as to what my problem is Mavrik?
[18:10] <Mavrik> cen|3: hmm
[18:11] <Mavrik> what do you get when you pass the pix_fmt parameter?
[18:12] <cen|3> Option pixel_format not found.
[18:13] <Mavrik> yes
[18:13] <Mavrik> now read again the previous error message and check for typos.
[18:18] <gnarface> wuldn't gif be rgb8 ?
[18:19] <gnarface> nevermind i guess
[18:19] <ePirat> hello
[18:20] <ePirat> can ffmpeg segment a given file into segments of x segments length in specified format? cause I just saw the -segment_* options and wondered how to use them&
[18:23] <jesk> anyone knowing how to create free?
[18:23] <jesk> i mean a free atom for MP4
[18:25] <cen|3> sorry bout' that - I had to take a call
[18:25] <cen|3> gnarface - I tried that also and it gave the same error
[18:27] <cen|3> and if I try to use pix_fmt gif it says: Failed to set value 'gif' for option 'pix_fmt'
[18:28] <cen|3> what is the difference between musing and demuxing support? gif format only has muxing support... would that matter?
[18:28] <cen|3> musing=muxing
[18:29] <sacarasc> cen|3: Muxing is putting the file together, demuxing is taking it apart.
[18:29] <gnarface> cen|3: demuxing is the reverse of muxing... since the gif format only supports one video stream and no audio/subtitle/camera angle streams there's nothing from which to demux...
[18:30] <cen|3> ah...  ;) thx
[18:36] <meekohi_> Are there any clever ways of reducing filesize of a video if you know it *loops*? I'm encoding to webm.
[18:36] <meekohi_> I'm trying to explore the settings available, but it's a little overwhelming to guess what will and won't work.
[18:53] <cen|3> got it - apparently gif support sucks for ffmpeg - so everyone extracts the images with ffmpeg and using another program like imagemagick to create the gif - thanks anyways....
[18:53] <cen|3> thanks anyways guys
[18:54] <Mavrik> doh
[18:54] <Mavrik> you were trying to create an animated gif
[18:54] <Mavrik> yeah, that doesn't work
[18:54] <Mavrik> sorry, thought you're just trying to get an image :)
[18:55] <cen|3> yup - sorry - guess I should have said that...  ;)
[18:55] <cen|3> I assumed it was obvious considering I was pulling from a video file... haha - But this process works great....
[18:55] <cen|3> I really appreciate your efforts though
[18:58] <Mavrik> cen|3: haven't used those in a while, completely forgot you can do that :P
[18:58] <delicado> hi is the video format in each call to avcodec_decode_video2 always YUV420P?
[19:00] <Mavrik> delicado: nope
[19:00] <Mavrik> it's what the video is encoded in
[19:00] <cen|3> Mavrik - Yeah I haven't had much need of them for a long time myself, but I'm trying to post an image on a forum and it won't accept video formats, but it does accept an animate gif (never used ffmpeg before though)... So...  :)
[19:01] <delicado> oh not YUV420P, when i read an AVFrame::format i always get 0. so its YUV.
[19:05] <Mavrik> delicado: most videos are encoded to YUV420P, however you shouldn't rely on that
[19:05] <Mavrik> H.264 can take YUV422P and other formats
[19:05] <Mavrik> even RGB
[19:14] <gnarface> question: is this bad?   Duration: 00:00:04.98, start: 0.205433, bitrate: -2147483 kb/s
[19:15] <Mavrik> gnarface: or a bug
[19:15] <Mavrik> :)
[19:15] <gnarface> Mavrik: well, is that normal output for a (decrypted) VOB ripped from a commercially-encoded, region 1 DVD?
[19:16] <gnarface> actually i'm more interested in what the possible complications are that this could cause
[19:17] <gnarface> namely whether it could cause the complication i'm having (where in the apparent video framerates wildly vary in the resulting transcoded video)
[19:21] <Freakshow> Spideru: how are you outputing the rtp stream locally with ffmpeg? I'm just curious
[19:24] <jesk> damn
[19:24] <jesk> no way to add a free atom?
[19:25] <jesk> anyone?
[19:29] <varaderoguy> night all
[19:43] <Rockj> How would I go about reporting an issue where ffmpeg exits with return code 0 on read error?
[19:44] <Rockj> nor does it look like error messages are printed to STDERR, I guess its not supposed to be like this?
[19:54] <AlRazi> I am trying to convert mp3 -> aac+, the resulting file has 7000+ hours duration on itunes, i tried running qt-faststart on it but to no avail, I tried using ffmpeg to produce a wav file, and then convert the wav file to m4a using faac, it works fine, but faac doesn't support HE-AAC
[20:08] <delicado> how can i get a 'PixelFormat' from an AVFrame? so i can use it as parameter 3 for 'sws_getContext'. because it does not accept the AVFrame::format that i give to it. the compiler says i need an PixelFormat?
[20:10] <Mavrik> format of the frame, -1 if unknown or unset Values correspond to enum PixelFormat for video frames, enum AVSampleFormat for audio)
[20:10] <Mavrik> read the dcs
[20:10] <Mavrik> it's int because it can be used for audio and video
[20:11] <Mavrik> you'll have to cast it
[20:11] <delicado> Mavrik: i tried putting 0
[20:11] <Mavrik> why 0?
[20:11] <Mavrik> what would that accomplish?
[20:11] <delicado> it says i need PixelFormat.
[20:11] <delicado> yeah
[20:12] <delicado> i mean it does not accept int
[20:12] <Mavrik> yes.
[20:12] <delicado> it needs a PixelFormat? how do i get it?
[20:12] <Mavrik> --->> "int Values correspond to enum PixelFormat for video frames" <<---
[20:12] <Mavrik> do you even know C?
[20:13] <Mavrik> cast the int to pixelformat
[20:13] <delicado> i dont know C in depth. does '(PixelFormat)0' work?
[20:13] <delicado> ah okay. ill try
[20:14] <sente> i have an flv and a bunch of mp4s, how can I find the specific format/codec info of the flv, and reencode all the mp4s to be of that same type?
[20:14] <sente> (this is so they can be streamed with amazon's streaming cloudfront abilities)
[20:14] <delicado> it works. thanks
[20:15] <Mavrik> delicado: why 0?
[20:15] <Mavrik> if you want to convert from INPUT format to whichever OUTPUT format you want
[20:15] <Mavrik> you're supposed to pass the INPUT format and the OUTPUT format to sws_scale
[20:16] <Mavrik> sente: that's not a trivial problem
[20:16] <delicado> Mavrik: no it was just a sample. because AVFrame::format is int. sorry im bad in english.
[20:16] <Mavrik> delicado: ah, yea
[20:16] <Mavrik> so basically
[20:16] <Mavrik> (PixelFormat)AVFrame::format
[20:16] <delicado> yes.
[20:17] <Mavrik> sente: that's not an easy problem - run "ffprobe -i <file.flv>" to see it's data
[20:17] <Mavrik> but it won't show exact encoding settings
[20:18] <sente> i just need to make sure the new videos are "as streamable" as the flv is
[20:18] <sente> if that makes sense
[20:18] <sente> (i dont know much at all about video/codec stuff)
[20:19] <Mavrik> hrm
[20:19] <Mavrik> I guess you'll have to learn ;)
[20:19] <ePirat> when trying to compile a program I've written using the libav libraries I get "Undefined symbols for architecture x86_64" (more: http://pastebin.com/yQFDL90j)& Would be great if someone could help me. thanks in advance&
[20:19] <Mavrik> sente: what are you streaming them with? what's your target? who are you streaming them to?
[20:20] <Mavrik> ePirat: that sounds like you either have 32-bit libs while compiling 64-bit target, or you forgot "-l" linking parameters, or your library path is broken
[20:20] <Mavrik> (you're working on Linux?)
[20:20] <sente> Mavrik: using flowplayer
[20:20] <sente> ffmpeg -i MP4/1_Code.mp4 -sameq MP4/1_Code.flv
[20:20] <Mavrik> ICK. Don't use sameq.
[20:20] <Mavrik> :D
[20:20] <sente> [flv @ 0x1ff27a0] flv does not support that sample rate, choose from (44100, 22050, 11025).
[20:20] <Mavrik> sente: the question is
[20:20] <Mavrik> why aren't you streaming mp4 files to flowplayer?
[20:21] <Mavrik> it's the preferred way in new flash
[20:21] <ePirat> Mavrik, oh right forgot -l *blush* (I am using Mac OS) thanks
[20:22] <Mavrik> hrm
[20:22] <Mavrik> can't help you with MacOS sadly
[20:22] <sente> Mavrik: hrm, well the mp4's won't stream from amazon
[20:22] <sente> but an flv will
[20:22] <Mavrik> OS X has some wierd cocktail of 32/64bit arch with wierd compilers, don't have any at hand sadly :\
[20:22] <Mavrik> sente: I see
[20:22] <Mavrik> sente: "sameq" doesn't do what you think it does, plus it reencodes whole video
[20:22] <sente> Mavrik: it says the clip isn't found, when I try and stream the mp4
[20:23] <Mavrik> sente: I suggest you try just remuxing them to flv first
[20:23] <sente> Mavrik: i see
[20:23] <Mavrik> with ffmpeg -i <file.mp4> -codec copy <file.flv>
[20:25] <ePirat> Mavrik, can i find a sample file using the av libraries in order to see if it's a problem with my code or another?
[20:26] <Mavrik> ePirat: that looks like a linker problem
[20:26] <Mavrik> ePirat: if you're compiling with gcc
[20:26] <Mavrik> ePirat: I suggest you add "-L" path to the libav 64-bit libraries
[20:32] <zap0> anyone do video editing for final cut pro ?
[20:33] <ePirat> Mavrik, ok now i get a lot of implicit declaration errors http://pastebin.com/VPN8Cxd0 huh&
[20:41] <sente> Mavrik: thanks
[20:41] <sente> ill give that a shot
[20:41] <sente> Mavrik: didn't work
[20:41] <Mavrik> ePirat: that seems like your include files aren't being included correctly
[20:42] <Mavrik> ePirat: use -I on compile stage
[20:42] <Mavrik> to pass directory for include files
[20:42] <sente> http://i.imgur.com/g2lJw.png
[20:43] <Mavrik> sente: ah, your audio has wrong samplerate
[20:43] <Mavrik> you'll have to reencode it
[20:44] <sente> hmm, no idea how to do that
[20:44] <Mavrik> ffmpeg -i <file.mp4> -codec:v copy -codec:a libmp3lame -b:a 128k -ar 44100 <file.flv>
[20:44] Action: sente debates calling the customer and telling them to just fix it on their side
[20:45] <sente> nice, that's atleast working
[20:45] <sente> thank you
[20:51] <dericed> i have a collection of .Y.U.V files and am having trouble reading them. I use "ffmpeg -i frame_%06d.Y" and get "frame_%06d.Y: No such file or directory". Am I referring to it wrong?
[20:53] <zap0> dericed, zero based or 1 based?
[20:55] <dericed> zap0: ?
[20:55] <zap0> ?
[20:55] <dericed> :)
[20:55] <beandog> try quoting the filename?
[20:56] <dericed> same
[20:56] <beandog> bummer
[20:56] <beandog> dunno
[21:04] <dericed> found problem, my filenames started at the wrong number. fixed now
[21:13] <llogan> dericed: there is also the -start_number option. allows you to choose the image to begin the sequence.
[21:22] <dericed> llogan: I say that, glad it got added. I just wrote a loop to renumber the files because I have some gaps in the numbers as well.
[21:55] <relaxed> dericed: also, if they're jpgs you can cat them to ffmpeg as input
[22:14] <alyawn> does anyone have a good link to an example of properly setting PTS when encoding using libavcodec?
[22:20] <alyawn> I was under the impression that having time_base set correctly (1/60) and setting PTS to the frame number was sufficient
[22:29] <burek> alyawn, could you check the source code of the setpts filter?
[22:30] <alyawn> burek, ok
[22:32] <alyawn> burek, it seems that all setpts does is set the PTS to the evaluated value given to the filter
[22:34] <burek> I see
[22:35] <burek> well, can you check then how does a specific encoder does it
[22:35] <burek> which video encoder are you currently trying to write
[22:35] <burek> I mean, which video codec do you use for your code
[22:36] <alyawn> I'm attempting to use H264 wrapped in a mpegts
[22:38] <burek> hm then you should check how libx264 is doing it
[22:38] <burek> http://git.videolan.org/?p=ffmpeg.git;a=blob;f=libavcodec/libx264.c;h=d56dfe76fd6a42766604bc7444f046f74a481fa1;hb=37b5959d9689f5310640c7a0beaa7784c58bfa6f
[22:41] <alyawn> thanks, burek
[22:44] <delicado> hi im leaking memory in this code http://codepad.org/oTIoxAKo. i have SDL functions in there before i isolated the code that leaks. why is it still leaking? other than av_free_packet(). what function is missing so i dont have the memory leak? sorry for the bad english.
[22:45] <sente> Mavrik: the ffmpeg -i <file.mp4> -codec:v copy -codec:a libmp3lame -b:a 128k -ar 44100 <file.flv> example created flv's which had a vcodec error
[22:45] <sente> any idea?
[22:45] <Mavrik> what's a "vcodec error"?
[22:45] <sente> when i play it in VLC it says there's an error
[22:46] <Mavrik> *shrug*
[22:46] <sacarasc> What is the error?
[22:46] <Mavrik> flv is an old deprecated container which has problems with modern video codecs
[22:48] <delicado> anyone? am i missing something?
[22:48] <JEEB> it's not really deprecated and I'm not sure if it has problems with modern video codecs (compared to say, avi -- which has no support for b-frames at all)
[22:48] <JEEB> it might just not have anything else to go for it but the fact that it's simple
[22:48] <Mavrik> delicado: and where do you free the decoded frame?
[22:49] <Mavrik> JEEB: there's some compatibility problems with H.264 which is the cause for Adobe to push mp4 for flash in newer versions
[22:49] <JEEB> wut
[22:49] <JEEB> H.264 goes just fine into FLV as far as I know
[22:49] <Mavrik> lemme find the article
[22:50] <JEEB> it better be a technical one
[22:50] <JEEB> mp4 for apple is just natural because their mov is the base for mp4, and naturally not being limited by Adobe's specs is a good thing
[22:50] <JEEB> it's also much more robust generally
[22:50] <Mavrik> hrrm:  Use of the H.264 and AAC compression formats in the FLV file format has some limitations and authors of Flash Player strongly encourage everyone to embrace the new standard F4V file format.[7]
[22:50] <JEEB> ...
[22:50] <Mavrik> from here: http://en.wikipedia.org/wiki/Flv#cite_note-kaourantin-6
[22:51] <Mavrik> however the linked article doesn't say anything concrete
[22:51] <JEEB> yes
[22:51] <ePirat> Can anyone take a look at this: http://pastebin.com/5pDPDPSp and maybe give me an hint what could be wrong? Really annoying sitting here for 3 hours trying to get it work&
[22:51] <JEEB> also that's from 2007 and adobe is still using flv just fine for its streaming solutions (rtmp and friends)
[22:52] <JEEB> and as far as I know technically the format fits H.264/AAC just fine
[22:52] <Mavrik> JEEB: perhaps
[22:52] <rm-rf> i'm running ffmpeg using daemontools to capture video from an IP cam, and every 180 seconds it cuts the stream and starts a new one so that the filesize doesn't get ginormous. in doing this, i'm finding that i gets a bunch of little files (200k-2MB) scattered among the actual 3 minute files, but can't figure out why
[22:52] <rm-rf> any thoughts?
[22:53] <sente> VLC does not support the audio or video format "undf". Unfortunately there is no way for you to fix this.
[22:53] <JEEB> Mavrik, if you ever get any technical insight on the actual problems related to H.264/AAC audio in FLV other than "the timescale can't be !1000"
[22:53] <sente> the audio is there
[22:53] <juanmabc> ePirat: you would actually need to link "-lavcodec" or whatever
[22:53] <JEEB> Mavrik, do feel to tell
[22:54] <JEEB> anyone else is free too, of course
[22:54] <Mavrik> JEEB: will do
[22:55] <delicado> Mavrik: do i have to free the frame? what function can i use? is it avpicture_free?
[22:55] <Mavrik> delicado: of course you have to
[22:56] <Mavrik> at least at the end
[22:56] <burek> rm-rf, how do you mean "using daemontools"
[22:58] <delicado> yeah i added it. but it is still leaking.
[22:59] <burek> sente, can you please use pastebin.com, to show your command line and its output?
[23:00] <JEEB> Mavrik, looking at the FLV spec I can't really find any technical limitation
[23:00] <Mavrik> JEEB: hmm, deleting that statement from Wiki would be a good move then
[23:01] <JEEB> the move by Adobe from FLV to F4V seems just purely "we can build upon an openly specified format and we get to be able to stuff a whole lot of more stuff into this"
[23:01] <rm-rf> burek: daemontools is a process that monitors other processes, runs certain processes based on criteria, etc.
[23:01] <JEEB> also, man -- the flv spec is short
[23:02] <JEEB> I bet people who liked AVI will like this
[23:02] <alyawn> ok... I'm an idiot... I was reseting PTS to zero after each frame encode. thanks again, burek
[23:03] <burek> alyawn, :beer: :)
[23:03] <rm-rf> burek: one process i have, starts up ffmpeg to grab an rtsp stream and write it to a file. another process that accompanies it, monitors the runtime of the ffmpeg process, and kills it when it gets to 180 seconds. daemontools then sees that the process has "died" and starts a new one
[23:03] <ePirat> juanmabc, ok seems to work with shared libs but static seems to fail& hm
[23:03] <rm-rf> burek: does that makes sense?
[23:03] <JEEB> (one of the main reasons why certain tool authors never moved away from AVI-based stuff is because AVI seriously was simple, and mp4 was a... clusterfuck, and matroska had problems of its own and/or wasn't known)
[23:03] <burek> rm-rf, why do you that?
[23:03] <burek> what is your overall goal?
[23:03] <sente> burek: http://pastie.org/pastes/4156527/text
[23:03] <rm-rf> to limit file sizes
[23:04] <rm-rf> and the runtime of each file
[23:04] <burek> why don't you use -t for example
[23:04] <burek> to limit the time of the output
[23:04] <rm-rf> would that end the process, or just cut the file off and start a new one?
[23:04] <burek> end the process
[23:05] <burek> sente, mpeg4 video and mp3 audio inside flv..?
[23:05] <burek> why?
[23:05] <sente> burek: i have no idea what i'm doing
[23:06] <burek> :))) at least you are honest :) :beer: :)
[23:06] <sente> I just need to convert the mp4 to flv so i can stream it with amazon's CDN
[23:06] <sente> :)
[23:06] <burek> I see
[23:06] <sente> I know i should be able to stream mp4, but i can't get it to work
[23:06] <JEEB> burek, he is doing video copy and at least that right
[23:06] <sente> i can stream the flv just fine though
[23:06] <JEEB> but the problem of course is that flv doesn't support that format
[23:06] <JEEB> (MPEG-4 Part 2)
[23:06] <burek> sente, do you see video when you stream it?
[23:06] <rm-rf> burek: that might work well. is it '-t' measured in seconds?
[23:07] <sente> burek: no, just black
[23:07] <JEEB> wait... ffmpeg can actually copy mpeg-4 part 2 into flv...?
[23:07] <burek> rm-rf yes, or you can set it in hour:min:Sec.msec format
[23:07] Action: JEEB goes check
[23:07] <JEEB> sente, yeah... I guess you're able to do it
[23:07] <JEEB> but you should check the flv specification, page... 72
[23:07] <JEEB> mpeg-4 part 2 is not a supported video format for FLV
[23:08] <burek> sente, is it viable for you that you re-encode your video too?
[23:08] <sente> define viable?
[23:08] <JEEB> http://pastebin.com/pFt55P3v
[23:08] <JEEB> what the specs say
[23:08] <sente> i dont care much about quality, i just need these things as FLV files, and preferably not 100x larger than the original mp4 files
[23:09] <JEEB> use libx264 and crf
[23:09] <JEEB> find the highest crf value that still looks good for you and the slowest preset that is fast enough for you
[23:09] <burek> i see
[23:09] <burek> try this then instead
[23:09] <JEEB> http://mewiki.project357.com/wiki/X264_Settings#preset
[23:09] <JEEB> and start from crf around 23
[23:09] <burek> ffmpeg -i 4_Energy.mp4 -codec:v libx264 -codec:a libmp3lame -b:a 128k -ar 44100 4_Energy2.flv
[23:09] <JEEB> also why are you re-encoding the audio?
[23:10] <JEEB> it seems like the input already is mp3
[23:10] <sente> JEEB: Mavrik suggested I had to
[23:10] <JEEB> sente, nah
[23:10] <JEEB> it should work
[23:11] <sente> JEEB: okay so what's the command line I need, without the extraneous stuff?
[23:11] <sente> I appreciate this
[23:11] <JEEB> in the basics it's like what burek just now said
[23:12] <sente> JEEB: okay trying that line
[23:12] <sente> thanks
[23:13] <JEEB> just switch -codec:a to copy
[23:13] <JEEB> and remove the -b:a and -ar
[23:15] <JEEB> ffmpeg's libx264 interface, if it uses the same defaults as libx264 should default to crf 23 and preset medium
[23:15] <JEEB> so if it looks good, you can try a higher -crf
[23:15] <JEEB> and if it looks bad, you can try lowering the -crf a bit
[23:16] <JEEB> if it feels like it's going slow, you can tweak the -preset (it controls the internal speed vs more compression settings, while the crf is "how much do I compress this" [qp etc. wise, which basically makes it the 'quality' lever])
[23:16] <JEEB> and if it's fast enough you can try a slower preset
[23:18] <rm-rf> i have two streams on this machine currently, one is an asf (foscam) and the other is mp4 (intellinet). is it possible to use the same command that i'm using to grab the mp4 stream with ffmpeg to grab the asf stream? cmd='ffmpeg -i <url> -vcodec copy -ar 44100 -t 300 /path/to/file'
[23:19] <burek> hmh.. basically to grab the stream, it is enough just to use ffmpeg -i url
[23:19] <burek> but what did you want to do with the rest of the command?
[23:20] <rm-rf> i'm using the rest of the command to output the stream into a file. it fails if i don't have the '-ar' option, and i'm not sure if i completely need the '-vcodec' option or not
[23:22] <rm-rf> burek: btw, thank you for the '-t' suggestion. that works so much better than killing the process and restarting it :)
[23:26] <burek> rm-rf, well, if you want to put them all into a specific format/container (like flv or mp4) then you can't just blindly use -vcodec copy
[23:27] <burek> because those formats might not support just any vcodec that you encounter
[23:27] <burek> -ar just changes audio rate (sampling rate, usually 44.1 KHz)
[23:27] <rm-rf> burek: i have a process that runs and converts them to mp4
[23:28] <burek> and it was maybe suggested to you for a specific case (for example mp3 in flv, which doesn't support 48 KHz mp3 audio, or something similar)
[23:28] <rm-rf> i had to use the -ar, because i tried forcing the samplerate to 128k, then calling '-ar 128000', and it sounded like crap
[23:28] <burek> well -ar 128k is just wrong
[23:28] <rm-rf> but i'm all ears
[23:28] <burek> -ab 128k might sound normal
[23:28] <rm-rf> this is all new to me
[23:28] <burek> but -ar no..
[23:28] <rm-rf> ok, noted
[23:29] <burek> sample rate tells you at which rate you take audio samples (actually your audio card does that)
[23:29] <burek> human ears can hear up to 21 KHz audio
[23:29] <burek> so the double the rate is fair enough for catching all the frequencies human ear can hear
[23:29] <rm-rf> i'm willing to look at anything suggested to me since i don't know much about the mplayer/vlc/ffmpeg world
[23:30] <burek> 48 KHz is just being fancy :)
[23:30] <burek> and -ab or the audio "bitrate"
[23:30] <burek> is how much of the bandwidth would you use
[23:30] <burek> for each second of your audio
[23:30] <burek> i.e. 128kbps
[23:31] <rm-rf> ah, not actually a samplerate
[23:31] <burek> so, tell me first
[23:31] <burek> what is your goal
[23:31] <burek> what is your logical requirement
[23:31] <burek> without much technical details
[23:31] <rm-rf> basically surveillance video, in 5 minute increments, searchable on a php based webpage
[23:32] <rm-rf> this is for a client, so i can't divulge too much more, and also, i'm not a paranoid freak
[23:35] <burek> ok
[23:35] <burek> so, you get your video from the camera
[23:35] <burek> is it an ip cam
[23:40] <rm-rf> burek: yes, one is a foscam IP camera, and the other is an intellinet IP camera
[23:41] <rm-rf> the foscam stream is ASF formatted, the intellinet stream is MP4 formatted
[23:47] <burek> so, ffmpeg -i http://url_to_cam -vcodec libx264 -acodec aac -strict experimental -ab 128k -ar 44100 -ac 2 out.flv
[23:47] <burek> or you can use .mp4 instead of .flv
[23:47] <burek> and then for your second cam you can just use -vcodec copy
[23:48] <burek> to save your cpu/time without encoding the video
[23:50] <alyawn> is this the correct way to specify x264opts using libavcodec? av_opt_set(c->priv_data, "x264opts", "no-ssim", 0)
[23:51] <alyawn> guess I can look in ffmeg.c, nvmd
[23:57] <rm-rf> burek: ah, so you are thinking that one cam will always encode video on the file, then the other cam would capture the raw stream and encode after the fact?
[23:58] <rm-rf> burek: i have to run, but thank you for your help
[00:00] --- Wed Jun 27 2012