[Ffmpeg-devel-irc] ffmpeg.log.20111126

[burek]

[00:00] <pythonirc1011> preset didnt work for me
[00:00] <pythonirc1011> lemme show you my setting...
[00:00] <burek> can you please use pastebin.com, to show your command line and its output?
[00:01] <pythonirc1011> http://codepad.org/90RnTcyw
[00:01] <pythonirc1011> for some reason, presets dont work well on windows, nor does libfaac
[00:01] <burek> put d:\Videos\00019
[00:01] <burek> .MTS
[00:02] <burek> into quotes
[00:02] <pythonirc1011> it works, without quotes...I can change that for sure.
[00:02] <burek> it confuses somehow things, i donno
[00:02] <burek> all those errors seem to be because of that
[00:03] <pythonirc1011> I'm not getting any errors
[00:03] <pythonirc1011> sorry , please ignore that "output" on codepad
[00:03] <pythonirc1011> thats not the output of my execution
[00:03] <pythonirc1011> I'm running some batch conversions right now, which will take some time...so cant paste the output yet
[00:04] <pythonirc1011> burek: do you see any obvious errors in the command that i should look into/fix/tune?
[00:04] <burek> i would use -threads 0
[00:04] <burek> to let ffmpeg autodetect No. of threads
[00:05] <pythonirc1011> well with threads 8 it takes too much time for me...on my 8 core machine
[00:05] <pythonirc1011> ah ok
[00:05] <burek> its not quite 8
[00:05] <pythonirc1011> will change that...didnt know 0 meant that
[00:05] <burek> you have 1 decoding and 8 encoding threads
[00:05] <burek> something like that
[00:05] <burek> so you miss one core
[00:05] <burek> -threads 0 is the safe way to do it
[00:05] <pythonirc1011> k, will change that for sure then
[00:06] <pythonirc1011> anything else? 
[00:06] <burek> you see -cmp +chroma ... etc
[00:06] <burek> you should use that only if you are expert
[00:06] <burek> and know what you're doing
[00:06] <burek> for all the others, use -preset
[00:06] <burek> its much better tuned
[00:06] <burek> and -tune too
[00:07] <pythonirc1011> I found that on the web by someone who had done it...for converting between avchd and mpeg4
[00:07] <burek> well ok
[00:07] <pythonirc1011> I cant use presets on windows for some reason, so had to go with that
[00:07] <burek> well, give me the output, and I'll tell you why you can't
[00:07] <burek> and how to fix it
[00:07] <pythonirc1011> is there a "-tune" that i should add?
[00:07] <pythonirc1011> perfect...give me some time. I'll show you a pastebin
[00:07] <burek> ok
[00:09] <pythonirc1011> http://codepad.org/IAtIA2Yo
[00:09] <pythonirc1011> For some reason adding -threads 0 started giving me errors
[00:10] <burek> hm, why do you reencode h264 video to h264 video?
[00:10] <pythonirc1011> burek: here are the errors with thread 0 : http://codepad.org/dRFqSeej
[00:11] <pythonirc1011> burek: The MTS file does not play with windows media player...the output mpeg4 does -- have no clue whats the difference
[00:11] <burek> did you try ffmpeg -i input.ts -vcodec copy -acodec copy output.mp4
[00:11] <pythonirc1011> yes i did
[00:11] <burek> and?
[00:11] <pythonirc1011> windows media player does not understand the output
[00:12] <pythonirc1011> just a black screen with audio
[00:12] <burek> wmp is just a direct show player
[00:12] <burek> meaning
[00:12] <burek> it doesnt understand anything
[00:12] <burek> if you dont
[00:12] <burek> have codec installed
[00:12] <pythonirc1011> when i play an mts file -- no video -- only audio
[00:12] <burek> can you download vlc for windows
[00:12] <pythonirc1011> when i play the mpeg4 after encoding (with the long command) -- both audio / video play well
[00:12] <relaxed> why are you setting those options and not using a preset?
[00:12] <pythonirc1011> vlc plays MTS and output without problems
[00:12] <burek> well use VLC then
[00:13] <pythonirc1011> relaxed: because on windows compiles, preset doesnt work
[00:13] <pythonirc1011> burek: If WMP cant play a video, other video applications have trouble rendering it too
[00:13] <relaxed> '-preset veryslow' doesn't work?
[00:13] <burek> do you understand how wmp plays videos?
[00:13] <pythonirc1011> relaxed: nope
[00:14] <burek> it uses installed codec (encoder/decoder)
[00:14] <burek> if that installed version of codec is old/wrong/bad..
[00:14] <burek> wmp wont be able to play it
[00:14] <burek> the point is, your wmp might not play that video
[00:14] <burek> but mine will
[00:14] <pythonirc1011> burek: where can i get a codec for MTS files? I looked on the web...this is what i found as a solution...use ffmpeg to transcode
[00:14] <burek> so its not a good tool to measure the correctness of a video
[00:15] <burek> well, lets go a more precise way
[00:15] <pythonirc1011> burek: according to your theory, my MTS should play, since h264 codec clearly is on my system?
[00:15] <burek> what do you want to accomplish in global
[00:15] <burek> it is, but what version of it?
[00:15] <pythonirc1011> Ultimately, I want to be able to use it in adobe premier
[00:15] <pythonirc1011> and be able to change its size if need be
[00:16] <burek> you can resize (rescale) your video with ffmpeg too :)
[00:16] <pythonirc1011> indeed...hence I was looking into ffmpeg
[00:16] <burek> so, once again, whats the ulimate goal
[00:16] <pythonirc1011> I cooked up that command so that i can run it on any MTS that the camera captures
[00:16] <pythonirc1011> be able to work with the video in adobe premier...
[00:17] <burek> then just remux your video into flv container
[00:17] <burek> ffmpeg -i ... -vcodec copy -acodec copy out.flv
[00:17] <pythonirc1011> i want quality control as well -- my cam is 24mbps -- will flv retain quality?
[00:18] <burek> isn't it better to configure the camera's output quality on the camera?
[00:18] <pythonirc1011> burek: for some reason my machine doesnt render that output
[00:18] <pythonirc1011> flv or mp4 -- copy
[00:19] <burek> update your x264
[00:19] <pythonirc1011> how?
[00:19] <pythonirc1011> i talked to canon tech support
[00:19] <burek> i can bet adobe's site can tell you that :)
[00:19] <pythonirc1011> they said their videos dont play on windoez
[00:20] <burek> but you said, it plays in vlc?
[00:20] <burek> so the support might be wrong
[00:20] <pythonirc1011> absolutely -- vlc renders everything...
[00:20] <pythonirc1011> its not using win 7 codecs
[00:20] <burek> god bless for that :)
[00:20] <relaxed> pythonirc1011: you get get recent ffmpeg builds here that support -preset. http://ffmpeg.zeranoe.com/builds/
[00:21] <pythonirc1011> I agree, but the rendering is flickery...24Mbps is too much for even VLC
[00:21] <relaxed> s/get get/can get/
[00:21] <pythonirc1011> thats where i downloaded mine from
[00:21] <pythonirc1011> btw, most of those builds crash!
[00:21] <burek> well, you're on win 7..
[00:21] <burek> what did you expect :(
[00:21] <pythonirc1011> yes
[00:21] <relaxed> oh? I don't use them
[00:22] <pythonirc1011> burek: All I care about is to be able to get my videos in a format that win 7 understands without loosing quality
[00:22] <burek> now we are talking things :)
[00:22] <wolfman2000> alright, I'm doing something wrong. I figured all I needed to do was copy the static libavcodec.a and libavformat.a to my project, but that isn't doing it. I then wondered if I had to include the associated header files, but now I'm given what seem to be silly errors such as 'avcodec_decode_video' is not a member of 'avcodec'. What is the proper way to use the static libraries in a project again?
[00:22] <pythonirc1011> and i wont take a russian codec pack on my machine :)
[00:22] <burek> so your final goal isnt to be able to load video into adobe premiere, but to be able to play it in win 7?
[00:22] <pythonirc1011> both actually... :)
[00:23] <pythonirc1011> burek: whats really going on -- so MTS is also h264 and ffmpeg output is also h264 -- its just a version difference because of which the MTS doesnt render?
[00:23] <burek> wolfman2000, i think your problem exceeds ffmpeg
[00:23] <burek> you should google more for static linking
[00:24] <wolfman2000> burek: I know about --enable-static. What else do you think I'm missing?
[00:24] <burek> pythonirc1011, im not sure how to compare two h264s
[00:24] <burek> you can maybe use mediainfo
[00:24] <burek> or some other tool
[00:24] <burek> but most probably its a different profile or something
[00:25] <pythonirc1011> burek: When a camera manufacturer claims that their videos are 24Mbps -- should i set -b 24000k?
[00:25] <pythonirc1011> or is there something like a VBR for best quality?
[00:25] <burek> wolfman2000, did you try google first?
[00:25] <wolfman2000> yes. so far, not helping
[00:25] <burek> pythonirc1011, yes.. -crf
[00:26] <pythonirc1011> burek: whats wrong with the command that i used...it converts my h264 that win7 doesnt understand to h264 that it does understand?
[00:26] <ghostbar_> burek, got it, I'm trying right now
[00:26] <burek> pythonirc1011 i told you already, it probably changes profile of h264 video
[00:26] <burek> and your h264 can read that profile, but not your camera's
[00:26] <burek> you should in any case update your h264 win codec
[00:27] <pythonirc1011> burek: if you know where i can find a good h264 codec for win 7, please let me know.. :)
[00:27] <pythonirc1011> Also: why does -threads 0 give these errors: http://codepad.org/dRFqSeej
[00:28] <burek> well, im not sure, but i would try google for those errors and figure out why :)
[00:29] <burek> wolfman2000, i just tried: http://www.google.com/search?q=ffmpeg+static+linking+example
[00:29] <pythonirc1011> burek: i would be ok with WMP not liking MTS videos, but what i dont like is premier not liking them...the preview pane is pain to work with with mts
[00:29] <burek> and I've found several links which are promissing
[00:29] <burek> http://soledadpenades.com/2009/11/24/linking-with-ffmpegs-libav/
[00:30] <burek> well
[00:30] <burek> maybe you have problems with how is it called in windows..
[00:30] <burek> not splitter..
[00:30] <burek> hmh
[00:30] <wolfman2000> I'm on Mac, now Windows
[00:31] <burek> filter i guess
[00:31] <burek> oh sorry, i was answering pythonirc1011 :)
[00:31] <burek> anyway, demuxer, which gets the video stream out of the container (flv, mp4, avi)
[00:32] <pythonirc1011> burek: vlc sucks with the 24Mbps videos on my machine....some frames render after a second!
[00:32] <pythonirc1011> another reason i've to move these videos to another format
[00:32] <burek> pythonirc1011, tools - messages (log level = 2)
[00:32] <pythonirc1011> or get a codec that works
[00:32] <burek> and see what does it complain about
[00:33] <burek> well, ok, then loose -threads completely
[00:33] <burek> and try like that
[00:33] <burek> of course, use -preset and -tune
[00:33] <burek> otherwise, don't complaint for bad quality
[00:33] <pythonirc1011> burek: http://paste.pocoo.org/show/512988/
[00:34] <pythonirc1011> thats from vlc
[00:34] <burek> your cpu can't manage to play it that fast
[00:34] <burek> i would try to tune the camera
[00:35] <pythonirc1011> burek: If an i7-2600K 4GHz CPU cant play a video...Its not the CPU...its VLC :)
[00:35] <burek> its 24mbps..
[00:35] <pythonirc1011> yes...and i want to keep it 24Mbps 1080p
[00:36] <pythonirc1011> sucks : adobe discontinued its media player! :(
[00:37] <burek> fun never ends :)
[00:37] <pythonirc1011> canon and microsoft fighting...consumer gets killed...
[00:38] <pythonirc1011> burek: does -crf take an argument?
[00:39] <pythonirc1011> whats the float? what value should i use?
[00:41] <burek> try -crf 2-
[00:41] <burek> try -crf 20
[00:41] <pythonirc1011> trying 25 :)
[00:42] <pythonirc1011> looks like its rendering 3Mbits
[00:43] <pythonirc1011> burek: is there a way to tell vlc to use hw acceleration?
[00:43] <pythonirc1011> got it
[00:43] <pythonirc1011> use gpu acceleration
[00:43] <pythonirc1011> trying it now
[00:44] <burek> try asking in #videolan
[00:44] <pythonirc1011> plain flat green video with gpu acceleration on :)
[00:45] <pythonirc1011> burek: what does the number "-crf 25" stand for?
[00:47] <ghostbar_> burek: it didn't worked. It grows the number to the infinite, not just until 9 :-(
[00:55] <pythonirc1011> burek: what does CRF really mean? At 15 it seems it encodes at 20Mbps+ levels. 
[00:55] <burek> pythonirc1011, try x264 --help and -fullhelp :)
[00:56] <burek> ghostbar_ :(
[01:23] <darkstarbyte> Are the -async and -vsync options important, and if so how should I use them?
[01:37] <burek> http://ffmpeg.org/ffmpeg.html
[01:46] <darkstarbyte> thanks
[01:46] <darkstarbyte> Would anyone know where I would go to get help with mkisofs?
[01:55] <sacarasc> :|
[01:55] <ubitux> :|
[01:55] Action: burek sets mode: +beer sacarasc
[01:55] Action: burek sets mode: +beer ubitux
[01:56] <burek> feel any better? :)
[01:56] <ubitux> i don't drink alcohol
[01:56] <ubitux> gimme apple juice
[01:56] <burek> thats what we all say :)))
[01:57] <ubitux> i really don't like alcohol actually :p
[01:58] <teratorn> any example code around of transcoding with a frame-rate change?
[02:01] <makario> Is it possible to take a certain video's audio track and mute specific moments in audio?
[02:03] <burek> teratorn what exactly do you want to accomplish
[02:03] <burek> makario, i think not
[02:03] <burek> you can extract certain parts though
[02:03] <makario> burek, darn. do you know of any other way to do that? basically i've been given the task of creating a quick 'swear word' muter
[02:03] <teratorn> burek: I'm transcoding, and I want to go from source frame-rate to a fixed, lower frame-rate
[02:04] <burek> skipping unwanted ones
[02:04] <burek> makario, using voice recognition?
[02:04] <ubitux> makario: extract the audio track, edit the audio with sth like audacity, and remux the stream
[02:05] <makario> burek, ubitux: I need to do it programatically using the subtitles track
[02:05] <burek> teratorn: ffmpeg -i input -r 25 ...
[02:05] <ubitux> then you might want to create an audio filter
[02:05] <ubitux> but this is not a simple task
[02:05] <makario> ubitux: audio filter?
[02:05] <teratorn> burek: yeah, it's just somewhat of a task tracing around in ffmpeg.c trying to figure out how it does its magic some times :)
[02:05] <ubitux> makario: do you know how to write in C?
[02:06] <makario> ubitux: yes, kind of. wrote a gba game, but have never done anything for a desktop (or using ffmpeg)
[02:07] <makario> ubitux: point me in the right direction and i'll bang my head against it until i figure it out
[02:07] <ubitux> that will require some time (let's say a few days) to make it i guess
[02:07] <ubitux> look at the various libavfilter/af_*.c files
[02:07] <makario> ubitux: okay, cool.
[02:07] <ubitux> libavfilter/af_pan.c is the most recent one
[02:08] <ubitux> it's a filter which allows you to "levelize" the channels
[02:08] <makario> are all of these in the source?
[02:08] <ubitux> yes
[02:08] <ubitux> http://git.videolan.org/?p=ffmpeg.git;a=blob;f=libavfilter/af_pan.c;hb=HEAD
[02:09] <teratorn> ubitux: do you know what technique is used to lower the frame-rate? does it just drop frames here and there?
[02:10] <ubitux> no i don't know
[02:10] <makario> ubitux, yeesh
[02:11] <makario> it'll be a few weeks, i think
[02:11] <teratorn> ubitux: k, thanks, I'll just try to decipher ffmpeg.c later
[02:11] <ubitux> makario: git show 1fbf7165d59907a0632f8b72664a31f97f218656 for the complete way of adding the filter
[02:11] <burek> teratorn
[02:11] <makario> thanks
[02:11] <burek> you can look at the -vsync too
[02:12] <teratorn> erm, I mean to address those question to burek :/
[02:12] <teratorn> burek: OK
[02:13] <burek> oh you mean how does it resample the video
[02:14] <burek> well isn't it obvious :) if the input rate is lower, it duplicates frames
[02:14] <burek> if higher, it drops
[02:15] <ubitux> makario: a simple way of doing it is to make a filter which takes ts ranges; something like silent=3-4.5,10.3-15.7,...
[02:15] <ubitux> then when the pts is in this range, you just put 0
[02:15] <ubitux> otherwise you copy the data
[02:16] <ubitux> that should do the trick
[02:16] <ubitux> if you suceed in it, you can send it to ffmpeg-devel at ffmpeg.org for review
[02:16] <makario> ubitux, yeah, i think that's the plan. 
[02:16] <ubitux> so it can be added
[02:16] <makario> all right, cool
[02:16] <ubitux> good luck :)
[02:17] <makario> very much appreciated!
[02:17] <ubitux> you can ask for help on #ffmpeg-devel for that matter btw
[02:19] <makario> will probably do
[02:19] <makario> now to read up!
[02:26] <ronag> why does a filter graph with yadif choose to convert yuv422p10 to yuv420p instead of yuv420p?
[02:26] <ronag> *instead of yuv422p 
[02:45] <burek> can you please use pastebin.com, to show your command line and its output?
[02:50] <Nagy> not commandline, code: http://pastebin.com/AHjMM3Gm
[02:53] <Nagy> this the log http://pastebin.com/FkYUabF1
[02:53] <Nagy> for some reason it does fmt:yuv422p10le -> w:720 h:486 fmt:yuv420p
[02:54] <Nagy> instead of what I would want: fmt:yuv422p10le -> w:720 h:486 fmt:yuv422p
[02:54] <Nagy> I'm not sure how libavfilter chooses formats, but that would be the logical conversion, wouldnt it?
[04:15] <jimlkl> Pasteeater....I figured out the problem I was having with certain videos being put on my iPod which had the sides cut off of them.
[04:17] <jimlkl> Well, if he's not here I'll tell you all- the iPod video setting of "fit to screen" had to be selected as OFF.
[04:31] <pasteeater> e
[04:32] <pasteeater> ...i know this isn't my terminal, damnit.
[05:05] <o3u> hi, how can i stream multiple files in a row, and do processing on each, at the same time? Right now i'm able to stream a single file, when i try to do consecutive calls to ffmpeg it doesnt seem to work, is there a way to do this?
[05:14] <johnnydude> hey guys, i'm trying to encode a dvd .VOB to webm, resize the video so its small and get the video to be about 5mb... 
[05:16] <johnnydude> for some reason the file keeps coming out massive
[05:16] <sacarasc> Set -v:b X
[05:16] <sacarasc> Where X is the right bitrate to make it 5MB.
[05:17] <sacarasc> Depends on how long the video is, though.
[05:17] <johnnydude> alright cool let me try... 
[05:18] <sacarasc> I think it's -v:b. These new changes confuse me sometimes.
[05:18] <johnnydude> i used some other tool to make a mp4, and it came out around 5mb with perfect quality... video is only 5min
[05:18] <johnnydude> hmm well, i was trying like -b and other stuff, but it didnt seem to apply 
[05:19] <johnnydude> or if it did, not how i was expecting, im pretty new to this
[05:20] <johnnydude> ffmpeg.exe -i D:\VIDEO_TS\VTS_01_1.VOB -f webm -s 480x384 -v:b 120 output3.webm
[05:20] <johnnydude> and it crashes when i start it
[05:20] <sacarasc> What does it say? Use pastebin.com for a large paste.
[05:21] <johnnydude> well im on windows atm, so i just get a windows error exception thing
[05:22] <johnnydude> w:720 h:576 pixfmt:yuv420p tb:1/1000000 sar:16/15 sws_param: [scale @ 032DE900] w:720 h:576 fmt:yuv420p -> w:480 h:384 fmt:yuv420p flags:0x4 [libvpx @ 032DDF20] err{or,}_recognition separate: 1; 1 [libvpx @ 032DDF20] err{or,}_recognition combined: 1; 65537 [libvpx @ 032DDF20] v0.9.7-p1
[05:23] <johnnydude> my god, i thought this was going to be easy
[05:23] <johnnydude> spent like 12 hours just trying to convert a video loll
[05:25] <johnnydude> ooh is it supposed to be "-vb" isntead of "-v:b" ?
[05:26] <sacarasc> I think it depends on version.
[05:27] <sacarasc> And it should be -b:v anyway. :D
[05:27] Action: sacarasc blames the cold.
[05:27] <johnnydude> im using some static git win32 version
[05:29] <johnnydude> alright, i do -b:v 160
[05:29] <johnnydude> but then it encodes at around 120?
[05:30] <johnnydude> do you know where the man page for latest ffmpeg is? since that switch isnt in what im reading at all..
[05:32] <sacarasc> http://ffmpeg.org/ffmpeg.html
[05:33] <johnnydude> damn running out of time.. gonna get fired cos im too retarded to encode a video lol
[05:33] <johnnydude> yeah thanks man :)
[05:36] <johnnydude> yeah i dont think -b:v -v:b or -bv are doing anything
[05:37] <johnnydude> ooh i see, i have to put a k on the end
[05:40] <o3u> any help on the streaming question? anyway to do it straight from ffmpeg?
[05:40] <o3u> imma try using a pipe, but it'd be nicer if i could do it with ffmpeg straight up
[06:42] <pasteeater> o3u: stream to what? "when i try to do consecutive calls to ffmpeg it doesnt seem to work" can you show your command(s) and output(s) on a pastebin site?
[06:56] <o3u> stream it to ffserver 
[06:57] <o3u> pasteeater: i cant seem to get pipes working with that either
[08:38] <o3u> concatenation sort of worked by using a pipe, then using ffmpeg to read that pipe to the server, and several other instances consecutively writing to the pipe, however i get a 'rc buffer underrun 'now, 
[08:38] <o3u> underflow rather
[08:42] <o3u> here are my terminals: http://pastebin.com/Qh8WbwQj 
[10:16] <EOF-sensei> how would one mux 6 channels of flac from separate files into an mkv?
[10:17] <EOF-sensei> (as well as encode a sequence of png files to dirac)
[10:58] <burek> EOF-sensei, use -map
[10:58] <burek> and -f image2
[10:58] <burek> more info here: http://ffmpeg.org/ffmpeg.html
[11:02] <burek> o3u: TCP connection to localhost:8090 failed: Connection refused
[11:02] <burek> read your output more carefully..
[11:04] <EOF-sensei> hmm
[11:04] <EOF-sensei> god
[11:04] <EOF-sensei> schroedinger really needs optimization work
[11:04] <EOF-sensei> and people need to stop wimpering and figure out a way to make wavelets compress well
[11:05] <EOF-sensei> </minirant>
[11:06] <burek> you can always help by sending such a patch :))
[11:07] <burek> or donating a box of beers or such ^^
[11:12] <EOF-sensei> indeed
[11:13] <EOF-sensei> every signal around me telling me I should put on my developer cap and stretch my C/++ muscles again
[11:13] <ubitux> and apple juice/
[11:13] <ubitux> oo/
[11:13] <EOF-sensei> I should probably work at finding workable motivators
[11:14] <EOF-sensei> I should probably start preemptive patenting
[11:15] <EOF-sensei> to avoid problems proving prior art to keep patents free
[11:20] <shevy> ffmpeg -ss 10 -i input.mp3 -t 30 new.mp3
[11:20] <shevy> am I reading this right: it says, start at position at 10 seconds, and copy the next 30 seconds into the file new.mp3
[11:20] <shevy> Because somehow, that does not seem to be the case... the resulting mp3 is 40 seconds long
[11:21] Action: shevy scratches his head.
[11:21] <shevy> hmm it seems also not quite accurate :(
[11:22] <EOF-sensei> shevy: re-encoding mp3s is a lossy process
[11:22] <EOF-sensei> I suggest using an mp3 clipper
[11:22] <EOF-sensei> as it will not further degrade the audio quality
[11:22] <shevy> ah, the above command will re-encode?
[11:23] <EOF-sensei> I would imagine
[11:24] <EOF-sensei> use mp3split
[11:24] <EOF-sensei> you won't have issues
[11:27] <EOF-sensei> although I suppose mp3split isn't in all distro repositories
[11:28] <EOF-sensei> (including my own)
[11:28] <EOF-sensei> I wonder why
[11:28] <EOF-sensei> patent issues I guess
[11:41] <EOF-sensei> is theholyduck everywhere?
[11:50] <burek> EOF-sensei no need for other tool, ffmpeg can also do it
[11:50] <burek> shevy, you can add -acodec copy
[11:51] <burek> and put -ss after -i (to seek after decoding frames)
[11:51] <burek> like this: ffmpeg -i input.mp3 -ss 10 -acodec copy -t 30 new.mp3
[12:01] <shevy> hmm
[12:01] <shevy> thx!
[12:03] <EOF-sensei> I wasn't sure if -acodec copy would split without artifacts at the split points
[12:03] <EOF-sensei> good to know there's an option for that
[12:03] <EOF-sensei> although
[12:03] <EOF-sensei>  /ffmpeg getting less unix-y/
[12:05] <shevy> I think I know what's with that file
[12:05] <shevy> it seems to not have a proper header
[12:06] <shevy> that explains why ffmpeg was reporting wrong duration, I think :)
[12:56] <TryDent1> what ffmpeg version fixes the 2 byte bug
[12:58] <burek> TryDent1 what 2 byte bug
[12:58] <TryDent1> ffmpeg -i music.file output.wav
[12:58] <TryDent1> it creates 2 bytes bigger than it should
[12:59] <JEEB> what format is the music file? and is it in a container or a raw stream?
[12:59] <TryDent1> input or output
[12:59] <JEEB> input
[12:59] <TryDent1> flac
[12:59] <JEEB> ok, not sure about flac
[13:00] <TryDent1> if i use flac program it creates exact same file as original.wav
[13:01] <ubitux> TryDent1: we already explained to you
[13:01] <JEEB> aac and mp3/ac3 and similar lossy formats at least have priming samples which would get decoded by ffmpeg because there's no standard on how to specify how many priming samples were used by the encoder
[13:01] <JEEB> (unless it's muxed into a container)
[13:01] <TryDent1> jeeb i am only comparing with lossless formats
[13:02] <TryDent1> ubitux  yes i was told there was a bug
[13:02] <ubitux> no there isn't
[13:02] <ubitux> it was told to you to report it to ffmpeg & flac if you think it is
[13:02] <ubitux> the original flac was certainly made by flac
[13:02] <TryDent1> how is it not a bug: it's creating 2 bytes bigger
[13:02] <ubitux> so it's no surprise it is the same
[13:03] <ubitux> if you remux the flac with ffmpeg, convert it to wav, then convert it back to flac, you will obtain the same
[13:03] <TryDent1> what if  zip was doing that; creating 2 bytes bigger
[13:03] <ubitux> TryDent1: the audio data is the same
[13:03] <ubitux> the headers might be different
[13:03] <ubitux> just like you can have different zip with the same content
[13:04] <TryDent1> ubitux  i  encoded using  ffmpeg -acodec flac   and decoded using flac app;  and it created exact file as original.wav
[13:05] <ubitux> i can create you 2 zip with the exact same content, one with 2 more bytes than the other
[13:05] <ubitux> and this is *not* a bug
[13:05] <TryDent1> ubutux  show me: prove it
[13:05] <ubitux> are you really trying to make me waste my time?
[13:06] <JEEB> I feel like he's not really understanding what you're trying to say :V
[13:06] <TryDent1> because i don't believe zip/rar/etc would have that kind of serious bug
[13:06] <ubitux> JEEB: i think he's just trolling so i'm going to do something else :p
[13:06] <TryDent1> ubitux  i am serious
[13:06] <TryDent1> ubitux i want to know about 2 bytes bigger in zip scenario
[13:06] <JEEB> ubitux, sometimes it's a good idea to take a pause :3
[13:07] <TryDent1> ubitux i am just perfectionist; i want everything to be exact
[13:08] <JEEB> TryDent1, if you do ffmpeg -i input.flac -acodec copy output.flac and  then read that, does the resulting output.flac give the same output as input.flac?
[13:09] <TryDent1> jeeb okay; let's start with original.wav
[13:09] <TryDent1> my goal is  original.wav to match  output.wav
[13:10] <ubitux> TryDent1: http://ubitux.fr/pub/shots/_flacdiff.png
[13:10] <ubitux> see?
[13:10] <ubitux> only the header is affected
[13:10] <ubitux> basically what encoder was used
[13:10] <TryDent1> that picture means nothing to me
[13:10] <JEEB> ah
[13:10] <JEEB> yeah
[13:10] <JEEB> you should only compare the audio data in the wav :)
[13:10] <ubitux> TryDent1: it's a diff between original flac and remuxed ffmpeg file
[13:11] <JEEB> uh, wait
[13:11] <ubitux> (with a wav step between)
[13:11] <ubitux> again, the data doesn't change
[13:11] <ubitux> only the header (meta information and such) change
[13:11] <TryDent1> why you comparing flac files?
[13:11] <ubitux> that is *not* a bug
[13:11] <TryDent1> why not compare  original.wav and output.wav
[13:11] <ubitux> you were talking about flac.
[13:12] <TryDent1> i never said  ffmpeg had  flac encoding bug;  i said    ffmpeg has  decoding bug
[13:12] <ubitux> share you sample
[13:12] <ubitux> and the exact procedure
[13:13] <TryDent1> okay
[13:15] <TryDent1> ubitux i think there is big misunderstanding  i was complaining that md5sum or original.wav and output.wav didn't match;  i wasn't complaining about flac files
[13:16] <JEEB> if you bindiff the files, is the difference in the beginning or end?
[13:16] <TryDent1> let me try
[13:16] <ubitux> TryDent1: give the exact steps to reproduce, and the sample
[13:16] <JEEB> since I don't think that ffmpeg gives a completely different result
[13:17] <ubitux> yes
[13:17] <ubitux> that's exactly the same issue as flac
[13:17] <ubitux> wav has a header too
[13:18] <TryDent1> where do i get bindiff
[13:19] Action: JEEB doesn't know of any good (visual) binary differs
[13:20] <TryDent1> then why are you recommending me to use it
[13:21] <JEEB> hell, you can open the file in a hex editor of your choice
[13:21] <JEEB> and check the beginning and the end of the file :P
[13:22] <burek> if i understand correctly, you encoded 1.wav with flac tool, making 1.flac and now you are decoding 1.flac with ffmpeg to 2.wav, and 1.wav differs from 2.wav?
[13:22] <JEEB> yes
[13:22] <JEEB> that seems to be the case
[13:22] <TryDent1> burek correct   but i used ffmpeg to encode
[13:22] <burek> and flac to decode?
[13:22] <JEEB> and it also differs from 3.wav that comes out of 1.flac done with flac's reference decoder it seems
[13:23] <TryDent1> burek if i use flac to decode  it matches; if i use  ffmpeg to decode it does not match
[13:23] <burek> oh i see, so if you do it all with ffmpeg, then 1.wav differs from 2.wav
[13:23] <TryDent1> correct
[13:23] <burek> can you upload somewhere those 2 wavs?
[13:23] <TryDent1> sure
[13:24] <TryDent1> but let me use small samples
[13:24] <burek> ok
[13:24] <JEEB> it's fine
[13:24] <JEEB> as long as it replicates
[13:24] <TryDent1> burek  but if you have a wav file you can do it yourself
[13:25] <TryDent1> assuming yuou have both ffmpeg and flac app
[13:25] <burek> i dont :(
[13:25] <burek> i dont use flac generally
[13:25] <burek> only aac+ :)
[13:26] <TryDent1> what is aac+?
[13:26] <burek> what is aac+ ???
[13:26] <burek> dear god, forgive him
[13:26] <burek> :)
[13:26] <TryDent1> i know aac  but not aac+
[13:26] <burek> its the future :)
[13:27] <burek> lets just say that at 32/48 kbs you get "near cd quality"
[13:27] <burek> its HE-AAC v2
[13:27] <burek> http://en.wikipedia.org/wiki/High-Efficiency_Advanced_Audio_Coding
[13:27] <burek> with sbr and pps
[13:28] <TryDent1> HE-AAC is better than LC?
[13:28] <burek> you should've heard of it, since youtube for example encodes all its audio with that
[13:28] <burek> way better
[13:28] <TryDent1> i see
[13:28] <burek> take a look at the link i gave you
[13:29] <burek> and see the image on the right
[13:29] <TryDent1> i had a  aac-LTP  once and  some many players had problem with it
[13:29] <burek> where is AAC LC and where is HE-AAC
[13:29] <TryDent1> does he-aac v2  have compatibility problem like  he-ltp
[13:29] <burek> i donno
[13:30] <TryDent1> i mean aac-ltp
[13:30] <burek> but i know that, when something is so good, the others will converge to it
[13:30] <burek> meaning, the manufacturers of audio players will tend to incorporate aac+ into devices
[13:30] <TryDent1> what aac encoder support he-aac v2?
[13:31] <burek> Scientific testing by the European Broadcasting Union has indicated that HE-AAC at 48 kbit/s was ranked as "Excellent" quality using the MUSHRA scale.[8] MP3 in the same testing received a score less than half that of HE-AAC and was ranked "Poor" using the MUSHRA scale. Data from this testing also indicated that some individuals confused 48 kbit/s encoded material with an uncompressed original.
[13:31] <burek> nuff said :)
[13:32] <burek> libaacplus (for ffmpeg)
[13:32] <burek> its already in the ffmpeg git
[13:33] <TryDent1> what file extension does aac+ use
[13:33] <TryDent1> Unknown encoder 'libaacplus'
[13:34] <burek> you need to ./configure --enable-libaacplus
[13:34] <TryDent1> huh
[13:34] <JEEB> you have to use the library, configure it and IIRC the encoder did use example implementation's code so I'm not sure if the resulting binary will stay GPL
[13:34] <burek> it will not
[13:34] <JEEB> (the last point only matters to people who actually give the binaries out tho)
[13:34] <burek> exactly
[13:35] <TryDent1> does itunes support aac-plus
[13:35] <burek> surely it does
[13:36] <JEEB> doesn't QT's aac encoder support he-aac anyways?
[13:36] <TryDent1> what about nero or faac?
[13:36] <JEEB> so you could as well use it
[13:36] <burek> JEEB which one is that?
[13:36] <JEEB> it has an API, several command line apps use it
[13:36] <burek> faac is AAC LC only
[13:36] <JEEB> I don't think you can use it within ffmpeg :)
[13:36] <burek> nero is good too
[13:37] <JEEB> nero has HE-AAC, a bit worse than QTAAC tho IIRC
[13:37] <burek> JEEB ok :) i thought you are talking about libvo_aacenc, which also doesnt support he-aac
[13:38] <burek> TryDent1, http://tipok.org.ua/node/17 go to Download section, get the file, unpack, configure, make, install
[13:38] <TryDent1> i don't remember itunes has settings of changing lc or he-aac
[13:38] <burek> and then get ffmpeg from git, configure (with --enable-libaacplus), make, install and thats it
[13:39] <burek> well, why would player had such an option?
[13:39] <JEEB> you don't need a setting in a decoder
[13:40] <JEEB> because the bitstream tells the decoder which type of AAC it is
[13:41] <TryDent1> there is no option of LC or he-aac  in itunes
[13:41] <burek> :)
[13:41] <burek> read above
[13:41] <JEEB> and there shouldn't be
[13:42] <JEEB> or wait
[13:42] <JEEB> you mean the encoder?
[13:42] <JEEB> not sure if itunes has the settings
[13:42] <JEEB> it's meant to be 'dumb'
[13:43] <JEEB> http://sites.google.com/site/qaacpage/news/qaacrelease107
[13:43] <JEEB> this uses the QuickTime API
[13:44] <JEEB> (cabinet = downloads)
[13:44] <TryDent1> wow
[13:44] <TryDent1> why is it in google site not in apple site
[13:44] <JEEB> because it's not an apple app?
[13:44] <JEEB> sites.google is free hosting
[13:44] <JEEB> the QuickTime API is free for everyone to use :P
[13:45] <TryDent1> then why can't  qaac be used with ffmpeg
[13:45] <JEEB> because no-one has made a patch?
[13:45] <JEEB> d'oh
[13:47] <burek> wait what is the diff between libaacplus and that link above (if there was a patch)
[13:48] <JEEB> different encoder used
[13:49] <JEEB> also, depending on the understanding of how linking to Apple's API is considered GPL-wise (MPC-HC and friends link to it without caring), it might actually be possible to give out binaries that can link to it.
[13:49] <burek> so it could be considered GPL?
[13:50] <JEEB> LGPL or GPL, the Apple SDK is...
[13:50] <JEEB> I think it's BSD or something
[13:50] <JEEB> of course it only works on systems that have Apple's libs
[13:50] <JEEB> which limits it to Win/Mac
[13:50] <JEEB> unless you use wine, that is
[13:52] <JEEB> yeah, MPC-HC has the whole SDK in their repo it seems https://github.com/jeeb/mpc-hc/tree/master/include/qt
[14:01] <TryDent1> is there big difference between he-aacv1 and he-aacv2
[15:05] <mehmetali> Hi, how i can enable svq3 pixel format.
[15:08] <BlackBishop> can I stream by any chance the output of /dev/video0 ( my webcam ) to the internet so it can be viewed by vlc/mplayer or other players ? :)
[15:10] <BlackBishop> as in .. is there any easy way ? ( one line thing )
[15:13] <BlackBishop> I'm currently using : mplayer -tv device=/dev/video0:driver=v4l2:input=1:width=1920:height=1080 tv://1 -zoom -aspect 4:3
[15:13] <BlackBishop> but I want it to be able to stream
[15:38] <BlackBishop> mhm, trying to use ffserver, bind(port 3343): Address family not supported by protocol
[15:43] <BlackBishop> ffmpeg 0.7.7 if it matters :|
[16:58] <burek> BlackBishop, it has got something to do with ipv6
[16:58] <burek> it's a bug :(
[17:04] <BlackBishop> ow, fsck, yeah .. I have an ipv6 address .. :/
[17:05] <BlackBishop> I guess it might be fixed in newer versions but newer ones break my xbmc :/
[17:06] <burek> what version of ffmpeg are you using
[17:07] <burek> is it the latest git?
[17:08] <BlackBishop> nope, 0.7.7 right now ...
[17:08] <BlackBishop> last time I used ~amd64 ( in gentoo's repo ) broke xbmc .. 
[17:08] <BlackBishop> so .. I decided to wait 
[17:10] <burek> well
[17:10] <burek> you can do more than wait
[17:11] <burek> you can try the latest git
[17:11] <burek> since, multimedia tools, like ffmpeg, get updated every day
[17:11] <burek> bug fixed every hour
[17:11] <burek> so, it might be wise to use git, rather than "stable" releases :)
[17:11] <burek> also, it compiles in like 10-15 mins
[17:12] <BlackBishop> yeah, compile time isn't a problem .. BUT .. I have xbmc .. and I want that to work more than the ffserver thing .. 
[17:14] <burek> can you try vlc-nox
[17:17] <BlackBishop> what for ?
[17:17] <BlackBishop> streaming ?
[17:35] <burek> BlackBishop yes
[17:35] <burek> or use udp if possible
[17:35] <burek> ffmpeg -i ... -f mpegts udp://target.subnet
[17:40] <BlackBishop> neah, I'll wait 'till this gets sorted out ..
[18:40] <khali> can anyone suggest a simple frontend to ffmpeg for Windows?
[18:50] <burek> why frontend?
[18:50] <burek> why not cli
[18:51] <sacarasc> Which is a front end, really.
[18:51] <khali> burek: I'm happy with the command line, but I suspect my sister-in-law won't be
[18:51] <burek> just give her the link to the documentation :)
[18:52] <khali> sacarasc: it certainly is; but I asked for a _simple_ frontend ;)
[18:52] <sacarasc> It is simple.
[19:06] <khali> sacarasc: thanks for your help
[19:24] <pasteeater> khali: winff
[19:28] <khali> pasteeater: will suggest that, thanks
[22:40] <monstaRtruck> guys how do i sync up sound with my video
[22:41] <monstaRtruck> no recording setting will make it sync
[22:41] <monstaRtruck> i jus hav to do it afterwards
[23:20] <burek> what is the synonim for vlc's ffmpeg-hurry-up in ffmpeg?
[00:00] --- Sun Nov 27 2011