[Ffmpeg-devel-irc] ffmpeg.log.20111117

burek burek021 at gmail.com
Fri Nov 18 02:05:02 CET 2011


[01:09] <Brad_> I'm using ffmpeg on FreeBSD. Every time I try to configure it stops with ERROR: libaacplus >= 2.0.0 not found
[01:11] <sacarasc> Do you have libaacplus installed?
[01:14] <Brad_> Yes
[01:15] <Brad_> Includes and so's are there...
[01:15] <sacarasc> And it's version 2.0.0 or greater?
[01:15] <Brad_> 2.0.2_3
[01:17] <sacarasc> Check config.log or whatever it is to see why it doesn't like it.
[01:19] <Brad_> LOL
[01:20] <Brad_> It's failing on ffw3
[01:20] <Brad_> Fftw3
[01:20] <Brad_> Because I compiled fftw3 with cloog
[03:23] <gunvalkyrie> Hi all, I can't seem to get the -ab option to set the bitrate in OSX. I did the build myself, and the command runs, but I'm trying to for a 192k bitrate and it goes straight to 128 :(. Here's my command line & output: http://pastebin.com/aCYyP7cS . Did my build screw up or am I just doing something dumb?
[03:27] <bradsoto> hello I'm back with more problems
[03:27] <bradsoto> libaacplus: bad aac setting: br:64, AACch:65536, AACsr:22050
[03:27] <bradsoto> there's two audio channels :(
[03:27] <bradsoto> not 65535
[04:40] <gunvalkyrie> I just check my command line options using my windows build of ffmpeg, and they work. Do I need to do something special when I build ffmpeg or lame so that I can change the bit rate in OS X?
[05:03] <gunvalkyrie> Hi all, it's not the build, I tried a simpler command line and it worked. Still not sure what the trouble is though. I'll play with the command line options though.
[05:32] <gunvalkyrie> silly me, it's the qscale option. It doesn't work with MP3s on my windows build for some reason, but it does on my OSX build, probably a bug that was fixed between builds.
[05:57] <sasishiju> Hi all. 
[06:06] <sasishiju> I have an x264 decoding sequence in my windows application on the x264 encoded data. Initially I perform av_parser_parse2 and when the parse is complete, I call the avcodec_decode_video2 method. I scale this decoded output as per the format/size requirement. 
[06:07] <sasishiju> But I have an issue here. My video output seems to have pixellation in a "non-negligible" manner. I have not been able to track where the issue might be in. Can any of you evaluate the flow ?
[06:07] <sasishiju> I have given the main code snippet at http://pastebin.com/Uk9DhRwC   Only a few lines. Could you please see if I miss anything obvious?
[06:14] <sasishiju> This works!! That is the most difficult part... But it doesnt give the expected quality. Whenever there is big movement in the captured video ( such as waving hands towards and away from the camera), the decoded frames are pixellated by a large extent...  Any suggestion?
[06:41] <sasishiju> hi. anybody around?
[08:57] <ocs1> hi, has ffmpeg sctp support? if so, how can I stream over sctp myfile.avi ?
[08:58] <ocs1> hi burek
[10:02] <sasishiju> hi. any body out there? 
[10:03] <sasishiju>  My video output seems to have pixellation in a "non-negligible" manner. I have not been able to track where the issue might be in. I have given the main code snippet at http://pastebin.com/Uk9DhRwC . Could any of you please see if I miss anything obvious?
[10:05] <sasishiju> FYI, the pixellation is maximum, when the captured video has lot of movements...  Any settings required to achieve non pixellated decoding?
[10:34] <faLUCE> hi, has ffmpeg sctp support? if so, how can I stream over sctp myfile.avi ?
[11:03] <foo-nix> I'm trying to follow the tutorials in the ffmpeg site, to have a better understanding of avcodec (frames, packets, hinting, etc). However, it won't compile (link to be more specific):
[11:03] <foo-nix> http://codepad.org/nvq8jDse
[11:04] <foo-nix> I noticed many of the undefined references are declared in avcodec.h, so I assume (by heuristics) -lavcodec should give the linker what it needs.
[11:05] <foo-nix> but it obviously doesn't. These functions are also deprecated, does that mean they are in the header files, but not in the object-files?
[12:12] <kriegerod> how should i push a frame into filtergraph, if it starts with nullsrc? for "buffer" source, it would be av_vsrc_buffer_add_frame()
[12:47] <foo-nix> why can't my linker find av_register_all() ?  I have compiled ffmpeg from source (0.8.6), an am linking against that by basically giving g++ all the .a file:  http://pastebin.com/krryUeyF
[12:47] <foo-nix> So if it's not in these obect files, then where is it?
[12:52] <foo-nix> aaaaah! sorry, it's a classic c++/c-object files confusion :(
[12:52] <foo-nix> sorry
[13:55] <sasishiju> Will avcodec_decode_video2 return in a single call?
[13:58] <sasishiju> Assume I have parsed an input buffer completely and created an AVPacket from the parsed output, will I get decoded output in a single call to avcodec_decode_video2?
[13:59] <sasishiju> I have been hanging around with similar question for a few days now in this channel. Hope some learned ffmpeg-ians will come to my rescue :-(
[14:08] <foo-nix> sasishiju, In version 0.8.6 the function completely disappeared, If I am correct, and one needs to use AVCodec::decode(...) (I assume)
[14:09] <foo-nix> What is it with that? There is not documentation on what it expects as void *data, int *data_size, etc.
[14:10] <sasishiju> foo-nix: thank you. Let me get the latest ffmpeg then. you mean to say it is deprecated or not in use at all ? 
[14:10] <foo-nix> sasishiju, I am not an expert, It is what I am figuring out right now :)
[14:11] <foo-nix> sasishiju, This is a function-pointer in a struct, which feels like methods in OO (c++), but I guess it is not.
[14:11] <foo-nix> sasishiju, I assume this way libavcodec tries to implement polymorfism, i.e. different codec's, they way you call decode is the same.
[14:11] <sasishiju> foo-nix: are you getting any replies in the channel? I didn't get any single reply other than yours so far for the past 3 days :-(
[14:12] <foo-nix> sasishiju, I have asked some questions, and no replies indeed, the general thing to do is just keep digging :)
[14:12] <foo-nix> sasishiju, and I must admit that I have a tendency to answer the question myself a few minutes after asking the question, so it is good practice to ask yourself lots of questions.
[14:12] <sasishiju> foo-nix: fine, I am stuck with some fine tuning in the current implementation. For me the decode works and video is rendered. But some how, the output is pixellated. 
[14:13] <sasishiju> When there is large movement at the video capture, the decoded output is pixellated very badly
[14:13] <foo-nix> sasishiju, but yes, I am having a hard time decoding the documentation of libavcodec as well. Especially since they seem to change the api every once in a while for no apparant reason.
[14:14] <foo-nix> sasishiju, when using libraries which are pre-1.0, this is what you are doomed to :P
[14:16] <sasishiju> By the way, I downloaded 0.8.5 git
[14:16] <sasishiju> you know which is the most recent? 
[14:29] <khali> burek: hi
[14:46] <mantas322> Hello!
[14:46] <mantas322> first time here I have a question
[14:46] <mantas322> anyone here?
[14:48] <mantas322> I'm looking for particular (optimal) arguements that will convert a variety of different videos into mp4
[14:51] <mantas322> ffmpeg -i source_video.avi input -acodec aac -ab 128kb -vcodec mpeg4 -b 1200kb -mbd 2 -flags +4mv+trell -aic 2 -cmp 2 -subcmp 2 -s 320x180 -title X final_video.mp4
[14:51] <mantas322> I found this ^
[14:51] <mantas322> so I was thinking to strip out the the -s 320x180
[14:52] <mantas322> any thoughts?
[14:54] <khali> mantas322: "optimal" doesn't mean anything
[14:55] <khali> mantas322: it all depends on what your target and objective is
[14:55] <mantas322> how about functional then
[14:55] <khali> functional isn't any better
[14:55] <mantas322> just to convert to mp4 i guess
[14:55] <mantas322> so it'll work on mobile browsers
[14:55] <khali> do you want fast compression ? small files ? good quality ?
[14:55] <mantas322> same quality
[14:55] <mantas322> all I care about is mobile browser functionality
[14:56] <khali> so you care about file size, it has to be small
[14:56] <mantas322> no
[14:56] <mantas322> in fact I'd like to preserve as much of the original quality as possible
[14:56] <khali> don't reencode the file then :p
[14:57] <mantas322> which arguement is that?
[14:57] <khali> my point is that you don't know what you want
[14:57] <mantas322> sry im a nob
[14:57] <mantas322> :<
[14:58] <khali> or are unable to express it
[14:58] <mantas322> I want a variety of videos like avi mpeg wmv   etc.... to be turned into mp4
[14:58] <khali> "mobile browser" implies low bitrate to me, but apparently it doesn't to you
[14:58] <mantas322> same size quality aspect ratio
[14:58] <khali> oh
[14:58] <mantas322> just as long as it works on the mobile browser
[14:59] <khali> so basically you want to change the _container_ and leave everything else unchanged?
[14:59] <mantas322> I guess so
[14:59] <mantas322> any advice
[14:59] <khali> then you should first check the original codecs, and only reencode if you really have to
[14:59] <mantas322> ?
[14:59] <khali> else use -vcodec copy and/or -acodec copy
[15:00] <khali> it is the best way to preserve quality - and it is very fast obviously
[15:01] <mantas322> hmmmm
[15:01] <khali> mantas322: if you have to reencode and want to preserve as much as possible of the original settings, then you really want to drop -s, indeed
[15:02] <khali> but IMHO setting -b 1200k arbitrarily regardless of the input is going to do the wrong thing... if you preserve the quality, output bitrate should be somewhat related to the input bitrate
[15:03] <khali> I'm also curious what kind of display your "mobile browser" offers... I've never used bitrates as high as 1200k even for my desktop machine
[15:04] <mantas322> so remove the -b 1200kb aswell?
[15:04] <mantas322> hmm
[15:04] <mantas322> im going to try this out
[15:05] <mantas322> Thanks for your help
[15:08] <khali> mantas322: no, removing -b would make it default to something ridiculously small such as 200k
[15:09] <mantas322> Oh.
[15:09] <khali> mantas322: my advice would rather be: check the input bitrate and use the same for output
[15:09] <khali> mantas322: or apply some adjustment depending on the input codec
[15:09] <mantas322> wouldnt it do that be default if I omit that arguement?
[15:09] <khali> mantas322: no
[15:10] <mantas322> can u help me understand what bitrate means?
[15:10] <mantas322> is it like the frames per second ?
[15:10] <khali> not really
[15:10] <mantas322> what would change if a video with a lower bit rate gets converted to a higher bitrate
[15:10] <mantas322> and vice versa
[15:10] <khali> it's expressed in bit/second
[15:10] <khali> (or byte/second, or kilobit per second, etc... all the same)
[15:11] <khali> to make it simple : (video bitrate + audio bitrate + container overhead) * duration = filesize
[15:11] <mantas322> hmmmm
[15:11] <khali> higher bitrate = higher quality and larger files
[15:11] <mantas322> -maxrate 1500k -minrate 1500k
[15:12] <mantas322> ?
[15:12] <mantas322> thats somethign different right
[15:12] <khali> however, this "= higher quality" only works up to a certain point... it takes a lot of bitrate to gain the last few % of quality, and beyond the quality of the input, you can't go
[15:13] <khali> -maxrate and -minrate are used when you have a decoding constraint of buffer size
[15:13] <khali> I usually don't set them
[15:13] <khali> but you have have to, depending on your decoding device
[15:14] <mantas322> welp im going to test some using     -b 1200kn
[15:14] <mantas322> kb*
[15:15] <mantas322>  -flags +4mv+trell -aic 2 -cmp 2 -subcmp 2
[15:15] <mantas322> not sure what any of these do
[15:15] <mantas322> but im sure they help[ >:
[15:15] <mantas322> :>
[15:16] <rexbron> hey could anyone tell me what are the supported pix formats for libx264? I keep getting yuv420p10le when I would like yuvj420p
[15:21] <alyawn> rexbron: for 8bit:  yuv420p, yuvj420p, yuv422p, yuv444p 
[15:23] <rexbron> alyawn: Incompatible pixel format 'yuvj420p' for codec 'libx264', auto-selecting format 'yuv420p10le'
[15:24] <alyawn> sounds like you are using 10bit libx264
[15:24] <rexbron> alyawn: is it an either or situation?
[15:24] <alyawn> yes
[15:24] <rexbron> bugger
[15:24] <rexbron> well that explains it :P
[15:25] <JEEB> yeah, libx264 can't yet be built so that it can encode both 10bit and 8bit H.264
[15:29] <rexbron> that would be nice
[15:30] <rexbron> problem solved thanks alyawn!
[15:32] <kriegerod> how should i push a frame into filtergraph, if it starts with nullsrc? for "buffer" source, it would be av_vsrc_buffer_add_frame()
[15:34] <mantas322> welp that was a huge failure
[15:34] <mantas322> stupid videos >:O
[15:37] <TimNich> OK if I ask ffmbc ffmpeg to code an h264 file for me, it runs the thread at around 80% cpu loading (out of 1600%, multiple cores) with sub threads all niced up to 10. However it only runs at around 14fps. I can run multiple instances on different files to up the throughput, but how do I make it run faster for the one instance?
[15:41] <watermark> I'm writing a script to check if ffmpeg supports decoding of a certain codec.  Is there is way I get get ffmpeg to output the supported codecs, (-formats does't seem to be detailed enough.)  As an example, I know the video codec is wmv1, can I get ffmpeg to output whether this codec is supported for decoding or not?
[15:43] <watermark> nevermind, I'm stupid
[15:47] <khali> mantas322: these flags are for higher quality encoding compared to the defaults - I use them as well
[15:48] <mantas322> welp im certainly creating new mp4 alright
[15:48] <khali> mantas322: note though that they are not recommended for real-time encoding unless you have a lot of processing power
[15:48] <mantas322> however they dont work :<
[15:48] <mantas322> yeah
[15:48] <mantas322> I do
[15:48] <mantas322> this is like a one time script i plan on running on thousands of videos
[15:48] <faLUCE> hi, has ffmpeg sctp support? if so, how can I stream over sctp myfile.avi ?
[15:49] <khali> ok, then it's OK to go high quality
[15:49] <faLUCE> JEEB: :-)
[15:49] <khali> mantas322: what "does not work" with your mp4 file?
[15:49] <mantas322> when i try to open it with windows media player
[15:49] <mantas322> it says windows media player encountered a problem whilse playing the file
[15:55] <khali> mantas322: please put the whole command line + the output of ffprobe for the generated file on some pastebin site
[15:55] <mantas322> lemme dick around with this some more
[16:02] <mantas322> YAY I got it to work!
[16:02] <mantas322> :D
[16:03] <mantas322> my arguments are bitrate 1200, maxwidth 1000, maxheight 1000, and -f mp4
[16:03] <mantas322> im going to try to add more arguements now
[16:07] <khali> I didn't know about  maxwidth and maxheight, it's not even in the manual
[16:13] <mantas322> hah
[16:13] <mantas322> those where just -s params
[16:13] <mantas322> or script program or whatever
[16:13] <mantas322> but called differently in the xml file of my application
[16:13] <mantas322> setting them to 0 fixed the issue
[16:17] <mantas322> the black doesnt look as black
[16:18] <mantas322> everything else is great
[16:18] <mantas322> hmmmmmmmmmmm
[16:28] <caseyburk> Could someone on here explain to me why the conversion of AVI to FLV using ffmpeg results in botched colors? (i.e. - blues turn into green for some reason). Would anyone have any suggestions?
[16:29] <khali> caseyburk: try converting to a different format and see if the problem persists...this will tell you if the problem is with decoding the avi or encoding the flv
[16:30] <caseyburk> khali: Would you happen to have any suggestions as for a different format or would any do the trick?
[16:31] <khali> caseyburk: any should do... try mp4 for example
[16:31] <khali> caseyburk: are you reencoding the video, or doing pass-through (-vcodec copy)?
[16:32] <caseyburk> khali: It looks like I'm re-encoding it. Being completely honest here, the original intention was to take AVIs and simply convert them into FLVs for my website, so if I'm doing something unnecessary, by all means, feel free to let me know. I just thought using this program was about as easy as it could get.
[16:33] <khali> caseyburk: yes, ffmpeg would definitely be the right tool for the job
[16:33] <caseyburk> khali: Okay, so I tool the AVI and re-encoded it to AVI (ha) and the color issue is still present. I'll now try the MP4.
[16:33] <caseyburk> khali: *took
[16:34] <khali> caseyburk: in fact, an even easier check if to play the original video with ffplay... if it looks bad then it's definitely a decoding issue
[16:34] <caseyburk> khali: I've been looking at the input vs. output notifiers and it looks like the input = rgb whereas the output = yuv. Might this be an issue?
[16:35] <caseyburk> khali: I'll try to play it back--and yes, it was the same result with MP4.
[16:37] <caseyburk> khali: Okay, so do I need a separate player to use ffplay or am I not calling it correctly from ffmpeg... (Obviously, I'm new to this--sorry.)
[16:38] <caseyburk> khali: I was trying to use ffplay as follows: "ffplay video.avi" from the command prompt in the directory where tyhe ffmpeg is.
[16:42] <khali> caseyburk: if ffmpeg works, then ffplay should work as well
[16:43] <khali> caseyburk: unless it was not built for some reason
[16:43] <khali> caseyburk: are you using a self-built ffmpeg, or a package for your distribution?
[16:44] <caseyburk> khali: I must be using a pre-built because when I try to call ffplay or whatever, it acts like it can't find it.
[16:48] <caseyburk> khali: Okay, I got a hold of a build that had the ffplay in it and it plays the video perfectly... So I'm going to try and re-encode the video using this new build I just downloaded in hopes that it might've had something to do with the previous build...
[16:51] <khali> caseyburk: rgb to yuv could indeed be related to color distortion... it _should_ not, but as far as I know rgb sources are relatively rare so a bug in rgb decoding could go unnoticed for some time
[16:55] <caseyburk> khali: The MP4 turned out okay. Ha.
[16:55] <caseyburk> khali: I'll try the FLV now...
[16:58] <caseyburk> khali: It really must have been the build because the FLV turned out fine, too.
[16:58] <caseyburk> :)
[16:59] <khali> caseyburk: great; problem solved then?
[16:59] <caseyburk> khali: Definitely solved. Thanks for your help!
[16:59] <khali> caseyburk: you're welcome
[17:00] <khali> usually I come here to _ask_ for help, I'm glad I could return some of it today :)
[17:00] <caseyburk> Ha, I know the feeling man... I'm outta here for now. See ya.
[17:01] <necromancer> hey guys, i'm having a bit of a problem with FFmpeg's handling of an FLV input and encoding to WAV
[17:02] <necromancer> it's tripping up this script i wrote...FFmpeg converts FLV->WAV, then pipes that data to SoX which mixes the WAV with another WAV, then it pipes that back to FFmpeg to encode into MP3 (though that last step may not be as important)
[17:02] <necromancer> here's the logs: http://stackoverflow.com/questions/8160358/mixing-a-flv-audio-stream-with-a-wav-background-track-and-converting-to-mp3-wit
[17:03] <necromancer> it gets tripped up on the first ffmpeg pipe from sox
[17:06] <khali> necromancer: you should isolate the steps to find out where the problem lies
[17:07] <necromancer> khali: i think i know where it lies. when i do it manually (without piping) i get a number of warnings, but if i output to a WAV it works
[17:12] <khali> necromancer: is the audio in your input file actually encoded with speex?
[17:16] <faLUCE> hi, has ffmpeg sctp support? if so, how can I stream over sctp myfile.avi ?
[17:20] <khali> necromancer: what are you trying to achieve with -t wav?
[17:22] <khali> necromancer: I think you really meant -f wav, not -t wav
[17:22] <necromancer> khali: it is. it's encoded from an FMS stream
[17:22] <necromancer> khali: i read -t wav in the docs...
[17:22] <necromancer> what's -t wav do? I thought "-t" was "type"...
[17:23] <khali> -t is for duration
[17:23] <khali> -f is for format
[17:24] <khali> -t wav does nothing, except an error
[17:24] <necromancer> ahhh
[17:24] <necromancer> must have read it wrong
[17:24] <necromancer> khali: now i got "sox FAIL formats: can't open input pipe `|ffmpeg -i audioStream_1321399534128_21.flv -ar 44100 -ac 2 -f wav pipe:1': WAVE: RIFF header not found"
[17:26] <khali> necromancer: if you redirect the output of ffmpeg to a file, leaving sox alone for the moment, do you get a valid wav file?
[17:26] <necromancer> yes
[17:27] <khali> I don't have sox installed, so I don't think I'll be able to help further, sorry
[17:29] <khali> try adding -t wav before - in the sox options
[17:30] <khali> err; sorry, before the "|ffmpeg " construct
[17:30] <khali> necromancer: the man page for sox gives this as an example:
[17:30] <khali> sox -M -t wav "|genw --imd -" -t wav "|genw --thd -" out.wav
[17:31] <necromancer> ah thank you
[17:32] <khali> (I don't quite understand the error message you got though...)
[17:32] <necromancer> neither do i :)
[17:32] <necromancer> doing a test run now
[17:32] <necromancer> i'm just getting "Press [q] to stop encoding"...that looks good right? :)
[17:38] <bcardarella> Is there any way to rotate a few by a few degrees? I see the transpose option but that appears to only work on 90 degree increments.
[17:42] <khali> necromancer: yes it seems good
[17:42] <khali> bcardarella: I don't think this is possible, no
[17:42] <bcardarella> ok
[18:06] <fris> anyone know what could be causing this WARNING: library configuration mismatch. avutil
[18:21] <grendal-prime> oi...ok i got the streaming of the desktop working with grabx11/vlc  and a flash player (than you burek)  however in the interest of science..i would like to set up a stream with ffserver.  (because it looks much cleaner.)
[18:22] <grendal-prime> and also because i like working with burek...(luuuuky uuuuuuuu)
[18:22] <grendal-prime> hahahaha
[18:22] <grendal-prime> found a couple of howtos andfor the most part i gather i have the ffserver installed and i have  a basic config file.
[18:23] <grendal-prime> when i try and start it though (that being the server) i get an error. bind(port 8080): Invalid argument
[18:24] <grendal-prime> netstat does not report anything on that port
[19:14] <burek> grendal-prime, thats a bug in ffserver
[19:14] <burek> its related to ipv6
[19:14] <burek> which version of ffmpeg/ffserver are you using?
[19:14] <burek> latest from git?
[19:14] <khali> burek: to answer the question in ticket #660: no, I don't think this is a regression
[19:14] <khali> at least I have no reason to believe so
[19:15] <burek> fris, do make distclean, and then do make again
[19:15] <burek> ok :)
[19:24] <grendal-prime> burek,  version on my ffserver is...SVN-r0.5.1-4:0.5.1-1ubuntu1.
[19:25] <grendal-prime> says it was build on sep 16-2011  but that might be when it was installed on the server here.
[19:27] <burek> thats ancient
[19:27] <burek> it doesnt matter when its been actually compiled
[19:27] <burek> because its an old source code
[19:27] <burek> use the latest git if you can
[19:47] <mantas322> anyone know anything about
[19:47] <mantas322> "invalid data found when processing input"
[19:47] <mantas322> its an output I get from some of the video I convert
[19:51] <ubitux> the file is considered invalid at some point for some reason
[19:51] <mantas322> but why
[19:51] <alyawn> mantas322, that error depends highly on the format of the file
[19:51] <mantas322> yeh
[19:51] <mantas322> I notice it more with mpg
[19:52] <mantas322> is it just a bad currupt mpg?
[19:53] <mantas322> besides that, ffmpeg is GREAT
[19:53] <mantas322> Thanks guys for making ffmpeg
[19:55] <alyawn> unfortunately, AVERROR_INVALIDDATA is thrown in about 600 places, so it could be a lot of things. I ususally see that error when a file wasn't muxed properly, etc
[19:55] <mantas322> muxed?
[19:56] <mantas322> im totally new to video conversion so can you please explain what that means
[19:56] <alyawn> wrapped, in a container .mp4 .ts, avi, etx
[19:56] <alyawn> etc*
[19:56] <mantas322> oh i see
[19:56] <mantas322> so what would cause a person to create a video file which wasnt properly muxed?
[19:57] <Zeqadious> does anyone know if ff_vdpau_vc1_decode_picture symbol is supposed to be missing from GIT head's libavcodec.so?
[19:59] <alyawn> mantas322, it really could be a lot of things... corruption in the transfer of the file is probably most common for me. I've also seen it when certain streams were wrapped in older mpg or m2ts files
[19:59] <HrznDefeated> Hi all.  I'm currently evaluating ffmpeg for use in our application.  Can anyone comment about whether ffmpeg streaming (using RTP/UDP, MPEG2-TS/UDP, and MPEG2-TS/RTP/UDP) is ready for prime time?
[20:00] <HrznDefeated> PS Platform=Windows
[20:00] <fris> what would cause WARNING: library configuration mismatch avutil
[20:01] <mantas322> hmmmm
[20:01] <mantas322> that makes sense alyawn
[20:01] <mantas322> thanks
[20:01] <mantas322> alot of these videos are uploaded to us
[20:01] <mantas322> from all over the world
[20:02] <alyawn> well, there are a lot of crappy FTP clients out there
[20:02] <mantas322> and theres a small fraciton of them that I guess, arn't "muxed" correctly
[20:07] <alyawn> fris, I believe that means that ffmpeg is using a version of avutil that is configured differently than the ffmpeg executable
[20:10] <mantas322> lol I just unleashed my conversion servers on my all videos
[20:10] <mantas322> my workstation looks like the matrix
[20:11] <mantas322> no one questions you when you've go 4 remote pcs running ffmpeg programmatically on your displays
[20:12] <alyawn> nice
[20:14] <necromancer> what happens if "Press [q] to stop encoding" just stays there indefinitely
[20:17] <alyawn> necromancer, are you sure it's indefinitely? usually it just takes a while to finalize the file then exits as usual
[20:19] <necromancer> alyawn: yeah. i'm piping data from ffmpeg -> sox -> ffmpeg, and when i just run the ffmpeg conversion command it works fine
[20:19] <necromancer> i get size=    1690kB time=9.81 bitrate=1411.2kbits/s
[20:19] <necromancer> video:0kB audio:1690kB global headers:0kB muxing overhead 0.002543%
[20:19] <necromancer> like in a second
[20:21] <alyawn> ah... could it be that the first ffmpeg instance is getting the `q' and the last instance isn't?
[20:22] <necromancer> alyawn: i'm still waiting for those msgs on the one i have all piped through
[20:23] <mantas322> 882 videos to go!
[20:24] <mantas322> lol I didnt even know this channel existed
[20:24] <mantas322> let alone had a community of such massive scale
[20:24] <mantas322> u guys put #extjs to shame
[20:25] <mantas322> I guys alot of people are into converting videos for free
[20:25] <alyawn> so, when you press q, necromancer, does ffmpeg stop processing samples, or does it keep going?
[20:26] <alyawn> I should say, the last ffmpeg in the chain
[20:26] <necromancer> alyawn: it does, because i get something from SoX
[20:26] <necromancer> sox FAIL formats: can't open input  `-': can't find sox file format identifier
[20:28] <alyawn> ah.. so the first ffmpeg in the chain stops, but does the second one?
[20:29] <necromancer> alyawn: i don't think it gets there. here's what i see: https://gist.github.com/1374184
[20:30] <mantas322> 701 videos left
[20:32] <necromancer> alyawn: when i hit "q" sox gives me that error, so it must still be running
[20:32] <necromancer> perhaps it's waiting for a response?
[20:34] <alyawn> hmm I'm not familiar with how sox handles the stdin, out in this case, but maybe try redirecting the STDIN of the first ffmpeg from  /dev/null
[20:35] <alyawn> but I don't see the output from the second ffmpeg at all
[20:39] <mantas322> 517...
[20:46] <mantas322> 335
[20:46] <mantas322> :D
[20:47] <mantas322> I know u all need to know about my video conversion count down
[20:57] <mantas322> 114
[20:59] <ubitux> you will soon realized you can find better settings and will need to start over
[21:00] <mantas322> NO!
[21:00] <mantas322> settings are great
[21:00] <mantas322> the videos that ARE converting successfully a good enough
[21:00] <ubitux> :)
[21:00] <mantas322> its the bad ones that get created into 0kb broken mp4s that are the problem
[21:01] <mantas322> older versions of ffmpeg used to crash on those exceptions
[21:02] <mantas322> this newer version hot ffmpeg just creates broken ones, and moves on
[21:02] <mantas322> like a drunk drvier who hit a car and drove away
[21:02] <mantas322> 45 left
[21:03] <bsterne> before I type a detailed question, wondering if "I'm having trouble converting portrait-oriented videos taken on iPhone" is a common topic?
[21:04] <bsterne> I spent the morning searching for a solution and keep finding -transpose, but that seems to be the solution to a different problem
[21:04] <bsterne> ffmpeg detects the input dimensions incorrectly as far as I can tell
[21:05] <bsterne> and the output is a landscape video that has the correct rotation, but has clipped the top and bottom of the input video
[21:05] <bsterne> (I guess I've typed the details at this point)
[21:06] <mantas322> 4 ....
[21:06] <mantas322> finshed!
[21:11] <bsterne> ubitux: have you seen that issue? ^
[21:13] <ubitux> huh? why are you asking me?
[21:13] <ubitux> ah.
[21:13] <bsterne> oh, because you have mod and you're one of the only people talking in the channel
[21:13] <mantas322> clearly cause you have an @ sign in front of your name
[21:13] <ubitux> fix'd that for you
[21:14] <mantas322> he cant give advice any more
[21:14] <mantas322> gl 
[21:14] <bsterne> who should I ask then?
[21:14] <mantas322> ubitux why are some of my videos currupt
[21:14] <mantas322> ?
[21:14] <mantas322> ANSWER ME
[21:14] <mantas322> >:O
[21:14] <ubitux> :(
[21:15] <mantas322> ffmpeg wont convert currupt videos
[21:15] <ubitux> bsterne: ffmpeg-users mailing list if you don't get enough replies here
[21:15] <ubitux> bsterne: -vf transpose and various flip filter might help you
[21:15] <ubitux> about the video flip properties not being handled, i don't know
[21:16] <ubitux> you might want to open a trac issue for this
[21:16] <bsterne> it's not flipping that is my problem, it is cropping (assuming I have the terminology correct)
[21:17] <ubitux> also, maybe you can share a sample
[21:18] <ubitux> but i don't think i'll be able to help much
[21:18] <ubitux> surely ppl here might be :)
[21:18] <bsterne> that's why I'm here :)
[21:18] <mantas322> not me
[21:18] <mantas322> im only interested in resolving my own problems
[21:18] <ubitux> :)
[21:19] <necromancer> alyawn: what do you mean by the STDIN of the first ffmpeg is from /dev/null?
[21:21] <necromancer> i thought /usr/share/red5/webapps/audiorecorder/streams/audioStream_1321399534128_21.flv was the input file specified...why would it read from STDIN?
[21:34] <mantas322> welp
[21:35] <mantas322> im done
[21:35] <mantas322> thansk for all your help today guys
[21:35] <mantas322> and teachign me what muxed means
[21:35] <mantas322> and everything else
[21:58] <grendal-prime> hey burek 
[21:58] <grendal-prime> can you tell me if this looks like it will be current enough..
[21:58] <grendal-prime> https://launchpad.net/~jon-severinsson/+archive/ffmpeg
[22:02] <grendal-prime> also since this is such an old version i wonder if thats why i was having such a hard time with vlc as well.
[22:02] <grendal-prime> hmm
[22:04] <burek> grendal-prime, read the topic :)
[22:04] <burek> http://bit.ly/3xSE5
[22:05] <grendal-prime> heheeh oops sorry
[22:07] <burek> or this
[22:07] <burek> http://pastebin.com/rTihbf0z
[22:07] <burek> if you want to do it quickly
[22:07] <burek> without fancy stuff
[22:07] <burek> :)
[23:00] <grendal-prime> sure wish there was just a ppa for this..
[23:00] <grendal-prime> because this part sucks
[23:00] <grendal-prime> Updating FFmpeg and x264
[23:00] <grendal-prime> Development of FFmpeg and x264 is active and an occasional update can give you new features and bug fixes. To update FFmpeg and x264 you will need to remove the packages, make distclean, update the source, recompile, and install. To update x264:
[23:00] <grendal-prime> yuk
[23:02] <relaxed> grendal-prime: what do you need?
[23:03] <grendal-prime> nothing ...just saying i wish there was an updated ppa for ubuntu.  (complaining like a wining ass end user for the most part)
[23:04] <relaxed> grendal-prime: I have recent static builds here --> http://sites.google.com/site/linuxencoding/builds
[23:06] <grendal-prime> unfamiliar with how to use those actually
[23:06] <relaxed> download, uncompress and run ./ffmpeg
[23:19] <WillMarshall> Hey! How can I convince ffmpeg to convert a .wav to a 320kbps mp3?
[23:19] <WillMarshall> ffmpeg -ab 320k -i Psybreaks\ 03.wav Psybreaks03.mp3 seems to be doing 128kbps mp3s...
[23:20] <sacarasc> It is ffmpeg [input options] -i input [output options] output
[23:21] <WillMarshall> So
[23:21] <WillMarshall> ffmpeg  -i Psybreaks\ 03.wav -b 320k Psybreaks03.mp3
[23:21] <sacarasc> Yeah, more or less.
[23:21] <WillMarshall> Sweet
[23:21] <WillMarshall> Anything else I should know?
[23:22] <sacarasc> You're using an old ffmpeg! :p
[23:22] <WillMarshall> It's a 44.1khz/24bit WAV
[00:00] --- Fri Nov 18 2011


More information about the Ffmpeg-devel-irc mailing list