[Ffmpeg-devel-irc] ffmpeg.log.20150124

[burek]

[00:42] <rock> not sure if this the right forum to ask this question , but I do not know any other forum who can address this
[00:43] <rock> Which is the best Mpeg2 TS analyzer ?
[00:43] <rock> I would like to pass in Apple HLS URL and analyze the stream
[01:41] <squeegily___> Real quick, I have a video that I want to copy and I want to preserve (copy) both audio streams
[01:41] <squeegily___> -map 0:a makes it not copy the video
[01:43] <rock> try codec:a copy codec:v copy  or -vcodec copy  and -acodec copy
[01:43] <squeegily___> I want to apply some filters, etc to the video
[01:43] <squeegily___> And I want to copy both audio tracks
[01:43] <pzich> then map both audio tracks?
[01:44] <squeegily___> Then it drops the video track for some reason
[01:44] <squeegily___> But I think I got it working: ffmpeg -i input -vf <stuff> -map 0:v -map 0:a output
[01:45] <squeegily___> DANGIT that still reencodes them
[01:45] <c_14> add -c:a copy
[01:48] <squeegily___> c_14: That makes it only copy 1 track
[01:48] <squeegily___> For some reason -c:a:0 copy -c:a:1 copy does the same
[01:48] <c_14> It should copy all audio tracks
[01:49] <squeegily___> c_14: https://bpaste.net/show/03aadc968337
[01:49] <squeegily___> I want to burn in the first 2 subtitle streams and copy all audio tracks without reencoding (the audio)
[01:50] <c_14> You got rid of the maps
[01:50] <c_14> You shouldn't have
[01:50] <squeegily___> c_14: So I should do -c:a copy -map 0:a
[01:50] <squeegily___> Right?
[01:50] <c_14> yes
[01:50] <c_14> And -map 0:v
[01:50] <squeegily___> AH
[01:51] <squeegily___> Here we are! Re-encoding the video and copying both audio tracks!
[01:51] <squeegily___> Thanks!
[03:16] <StFS> Hi. I'm getting nasty antialiasing noise when I convert from flac to wav. Are there any common pitfalls I should be aware of? If I play the flac file in VLC it sounds fine but if I convert that flac file to wav (ffmpeg -i foo.flac foo.wav) and then play the wav file I get this antialiasing artifact.
[03:17] <StFS> I've tried using avconv, sox and just now I downloaded the git binary tarball of ffmpeg... all have the same effect
[03:18] <c_14> Tried pcm_s32le ?
[03:19] <StFS> uhm... noooo?
[03:19] <StFS> is that a parameter I should be passing to ffmpeg?
[03:19] <c_14> -c:a pcm_s32le
[03:19] <StFS> testing
[03:20] <StFS> should that be after the input file and before the output file right?
[03:20] <c_14> yep
[03:21] <StFS> well I'll be! :)
[03:21] <StFS> thank you!
[03:23] <StFS> I have multiple flac files that I want to convert... is there some way to autodetect whether this is needed? I'm assuming that it may or may not be needed depending on the flac format.
[03:24] <c_14> probably
[03:24] <c_14> Check the ffprobe for that flac file
[03:24] <c_14> If my guess is correct it'll say s24 on the Stream line
[03:26] <StFS> well... for the one file I really noticed this for it has s16
[03:26] <klaxa> have you checked the samplerate?
[03:26] <StFS> 44.1 k
[03:27] <klaxa> for both files, right?
[03:30] <StFS> klaxa: you mean for the resulting WAV file? Well.. the original (flac) is s16 / 44,1k, the bad-sounding wav is also s16 / 44,1k. The good sounding wav (created with the '-c:a pcm_s32le' option) is s32 / 44.1k
[03:30] <klaxa> ok
[03:31] <c_14> What do you use to test the quality? What player?
[03:31] <StFS> not quite sure what to make of this though :/
[03:31] <StFS> c_14: vlc
[03:31] <c_14> Can you try with ffplay ?
[03:32] <StFS> hmm... that doesn't seem to be included with the binary package I got... is there a binary somewhere I can grab?
[03:34] <c_14> try ffmpeg -i file -f alsa default
[03:34] <c_14> or -f pulse default
[03:34] <c_14> Whichever you use
[03:35] <StFS> and you want me to play the "bad wav" file?
[03:35] <c_14> The flac and the "bad wave"
[03:36] <StFS> Requested output format 'pulse' is not a suitable output format
[03:36] <StFS> same for alsa :/
[03:36] <StFS> I could put the files somewhere... but they're 24 Mb and 80 Mb
[03:37] <c_14> You can put the 24MiB one here: https://pomf.se/
[03:40] <StFS> flac: http://a.pomf.se/cijytk.flac
[03:40] <grepper> c_14: damn, after getting that -vf select='between(....  filter working,  turns out it bombs on avconv :(
[03:41] <c_14> Have you found a seek method that works reliably with avconv?
[03:42] <StFS> c_14: and I believe the bad flac is fully uploaded to https://www.dropbox.com/sh/h0elblnmnx0glja/AAB6KCww4f452k2j54Sd2j_Qa?dl=0
[03:42] <grepper> c_14: :)
[03:42] <grepper> after the -i FILE I guess
[03:43] <StFS> wow... if I play the bad wave file from the dropbox page it sounds fine!?
[03:44] <StFS> I'm confused
[03:48] <c_14> Are the 2 bit-identical?
[03:48] <c_14> If you've only tested with vlc, it might be vlc messing up.
[03:49] <StFS> which two?
[03:49] <c_14> The bad wave file and the "bad wave" file from dropbox.
[03:50] <StFS> umm... there is just one bad wav file... the wav file that came out of me trying to convert the flac... and that is the file that I put on Dropbox
[03:50] <c_14> ye, but you said that if you play the bad file from the dropbox page it sounds fine
[03:51] <StFS> oh... yeah sorry.. I meant if I press the "speaker icon" on the page...
[03:51] <c_14> aaah
[03:51] <StFS> if I download the file and play it in VLC it's still bad
[03:51] Action: c_14 starts pointing the blame finger at VLC
[03:51] <StFS> I don't know what exactly happens when I press that button though
[03:51] <StFS> c_14: yeah... it's looking a bit like that...
[03:52] <StFS> but I noticed this first when I burned a CD from the flac files with k3b... I'm going to try to find something else to play that CD
[03:53] <StFS> and see if I hear it that way... but it's still a bit weird that this problem doesn't seem to be affecting all the flac files I'm converting... I can play some of the resulting wave files just fine in VLC
[03:59] <StFS> yup... seems to be VLC screwing things up
[04:00] <StFS> sounded fine on a cd player
[04:00] <StFS> (well.. a PS3)
[04:01] <StFS> c_14: thanks for your help
[04:02] <c_14> np
[04:02] <StFS> just curious... did you download the -bad.wav file?
[04:03] <c_14> ye
[04:03] <StFS> and it sounds fine?
[04:03] <c_14> Sounds fine to me, yep.
[04:03] <StFS> in what player?
[04:04] <c_14> Both ffmpeg itself and mpv
[04:04] <StFS> ok... do you have vlc?
[04:05] <c_14> No
[04:05] <StFS> allright.. nevermind then :)
[08:04] <grepper> c_14: looks like I have one filter to rule them all, it turned out to be your original suggestion of lt and gt with the select filter :P Works with both avconv and ffmpeg.
[08:05] <grepper> I ended up learning 3 ways to do pretty much the same thing ...
[08:18] <YaMoonSun> If someone were to utilize ffmpeg in their own application, would said application have to be release it's source code?
[09:21] <Guest_> Hello. I want to create a silent audio file and then concatenate it with an existing aac file, but it's not working.
[09:21] <Guest_> This is the instruction I'm using to create the empty 3 seconds file: ffmpeg -f lavfi -i aevalsrc=0:0::duration=3 -ab 23k -ar 44100 test.aac
[09:21] <Guest_> This is the instruction I'm using to merge both files: ffmpeg -i concat:"test.aac"^|"orig.aac" -i "orig.aac" -map_metadata:g:0 1:g -c copy -f adts new.aac
[09:22] <Guest_> I run ffmpeg -i on the original file, and it gives me that the bitrate is 23 kb/s, that it's 44100 Hz, so I replicated that; but it also says in the audio channels info "2 channels (FC)", and I don't know how to replicate it. I tried making the silent file a stereo file with the instruction -ac 2, but that doesn't work either. Does anyone know how to help me? Thanks a lot.
[09:23] <Guest_> Oh, btw, the files do merge, but the problem is that after the 3 seconds silence all I hear is interference.
[09:57] <filterSAC> i'm writing a simple width height change filter of video can anyone give me some suggessions about that how i can start this,and how i design its algorithm?
[10:56] <hpuser4466> how do I make video 1/2 slower and keep the same quality x264 codec ?
[11:00] <hpuser4466> sorry i mean 2x faster
[11:02] <hpuser4466> The command on the ffmpeg website isn't working and ruins the encoding (pixelates the video) and when I specify x264 (hoping to keep the same quality) it doesn't change the speed...grrrr
[12:08] <DelphiWorld> yo
[12:08] <DelphiWorld> any channel / room for mlt framework
[12:49] <relaxed> filterSAC: wouldn't that be the scale filter?
[12:50] <relaxed> DelphiWorld: I think you can do it losslessly with ffmpeg -r <double framerate here> -i input ...
[12:51] <relaxed> DelphiWorld: sorry, that was meant for hpuser446 (who parted :/ )
[12:51] <DelphiWorld> relaxed: ;)
[13:03] <filterSAC> can anyone suggest me how i can understand vf_scale.c?
[13:06] <filterSAC> relaxed:yes that we a scale filter,can you suggest me about scale how i understand this filter?
[13:07] <DelphiWorld> need someone who used mltframework
[13:36] <relaxed> filterSAC: it's written in C, so you would have to learn it or know someone that does
[14:21] <cousin_luigi> JEEB: You still around?:)
[14:21] <cousin_luigi> c
[14:21] <JEEB> never
[14:41] <cousin_luigi> JEEB: is fdk-aac LGPL compatible?
[14:41] <JEEB> no
[14:42] <JEEB> that's why it's enable-nonfree for both
[14:43] <JEEB> the only way you can use fdk-aac in an app that you distro binaries for is to a) be compliant with both licenses separately (LGPL and fraunhofer's) b) use fdk-aac SEPARATELY from FFmpeg's libraries. as in, you can use LGPL libavcodec to decode audio, and then separately use fdk-aac's APIs to encode it, not within lavc
[14:43] <JEEB> but as I said, that requires being compliant with fraunhofer's license
[14:44] <JEEB> but IANAL
[14:45] <JEEB> so don't take this as any kind of legal recommendation
[14:47] <cousin_luigi> JEEB: I'm not sure I understand the difference between solution a and b.
[14:48] <JEEB> they're part of a single solution
[14:48] <JEEB> you have to be both compliant with both licenses, and have fdk-aac and FFmpeg's stuff be completely separate. That is similar to a proprietary app using LGPL FFmpeg
[14:49] <JEEB> in other words, as soon as you put fdk-aac into FFmpeg, their licenses clash
[14:49] <JEEB> so if you can be compliant with fdk-aac's license
[14:49] <JEEB> (you have licensed the related things etc etc)
[14:49] <JEEB> you just link fdk-aac to your app and use it separately
[14:50] <JEEB> and the FFmpeg part will have nothing to do with the audio encoding part of things :P
[14:51] <cousin_luigi> Is there a distinction between using static and dynamic libraries in this case?
[14:51] <cousin_luigi> I mean, can ffmpeg keep fdk-aac separate?
[14:51] <JEEB> nononono
[14:51] <JEEB> you misunderstand
[14:51] <JEEB> FFmpeg will not be using fdk-aac
[14:52] <JEEB> as soon as that happens you will get clashes
[14:52] <cousin_luigi> in any form?
[14:52] <JEEB> get that out of your head
[14:52] <JEEB> yes, in any form
[14:52] <cousin_luigi> Damn, then I need to have a talk with marge:)
[14:52] <JEEB> basically I'd rather see people investing in improving the libavcodec encoder
[14:52] <JEEB> if they want an LGPL encoder
[14:53] <JEEB> https://trac.ffmpeg.org/ticket/2686 the longest thread on the trac has a patch that improves the AAC encoder already
[14:53] <notToday3> hi just a short question is there any update on the adobe style auth?
[14:54] <cousin_luigi> So the package is in violation then. I'll have to tell someone.
[14:54] <JEEB> what package?
[14:54] <cousin_luigi> the one for opensuse
[14:54] <JEEB> they don't package ffmpeg at all
[14:54] <JEEB> just like fedora
[14:54] <cousin_luigi> on packman I mean
[14:55] <cousin_luigi> packman is the equivalent of rpmfusion I think
[14:55] <JEEB> yes, an launchpad PPAs for ubuntu
[14:55] <JEEB> *and
[14:55] <JEEB> place where people push their random 3rd party things
[14:55] <cousin_luigi> well, something slightly more official
[14:55] <JEEB> not really
[14:56] <cousin_luigi> well, you don't have private repositories à la PPA on packman
[14:56] <cousin_luigi> and packages are still subject to approval
[14:56] <JEEB> yes, but it's still third parties maintaining them
[14:56] <cousin_luigi> There's some overlap between OBS and PMBS maintainers.
[14:57] <JEEB> anyways, FFmpeg itself won't care unless someone is not following its part of the license, but yes - if we go by the licenses together it's a violation
[14:57] <JEEB> same for faac and libaacplus
[14:57] <cousin_luigi> in fact faac is disabled
[17:44] <SimAV> hi, is someone aware of some figures about how much certain CPU instructions help to speed up specific encoders?
[17:46] <SimAV> in my specific case i would be interested in whether (and if so, how much) AVX instructions speed up encoding e.g. when using libx264
[17:48] <SimAV> I would like to know whether there is an substantial (except for GHz-numbers) speed up when switching from a 820QM (previous generation i7, which has no AVX) to a 2820QM (second generation i7, which has AVX).
[17:49] <__jack__> you want a benchmark with all "--disable-asm" toggled ? I may do that, but the libx264 cannot disable AVX (at least on configure time)
[17:50] <SimAV> __jack__, yes, that sounds already quite close to what i would like to have
[17:50] <__jack__> I don't think enable / disable AVX for just ffmpeg will make any change (except for internal encoding / decoding / (de)mux
[17:51] <SimAV> its probably hard to simulate how ffmpeg would run on the older i7 if you have the newer one
[17:51] <SimAV> but still as you pointed out, the "problem" is probably libx264...
[17:51] <SimAV> alternatively I would be interested what's about vp8 (libvpx)
[17:52] <JEEB> <__jack__> you want a benchmark with all "--disable-asm" toggled ? I may do that, but the libx264 cannot disable AVX (at least on configure time) <- no
[17:52] <foonix> i think it would be doable under virtualization (ie with KVM) and remove cpu flags .. but myself im on old core2 ...
[17:52] <JEEB> there is an option during runtime that lets you enable and disable specific asm
[17:52] <SimAV> (vpx is horribly slow on the 820QM, as only singlethreaded, but i was told that it isn't that fast on other CPUs either...)
[17:53] <SimAV> foonix, I'm "working" as well on core2duo at the moment
[17:53] <JEEB> http://mewiki.project357.com/wiki/X264_Settings#asm
[17:53] <JEEB> I think this was a bit mask
[17:53] <SimAV> I'm just wondering which used PC i should take
[17:53] <JEEB> you want as new as possible intel
[17:53] <JEEB> AVX is used in x264
[17:54] <JEEB> and each generation gives you a few % of speed in general
[17:54] <SimAV> or whether both is really wasted money
[17:54] <SimAV> JEEB, i agree with you if speed it the most important variable that you want to optimize
[17:55] <JEEB> well, that gives you a pointer when you have a couple of alternatives within your budget :P
[17:55] <SimAV> but a new workstation is > 4x a used one...
[17:56] <SimAV> JEEB, thanks for your link
[17:56] <__jack__> JEEB: interesting, so a runtime option does exists
[17:56] <JEEB> yes, it will override what the cpu capability check finds
[17:56] <SimAV> but for testing that i would need the 2820QM first...
[17:56] <__jack__> however, it disable all asm-stuff
[17:56] <JEEB> no it does not
[17:56] <JEEB> --asm BIT_MASK
[17:57] <__jack__> oky
[17:57] <__jack__> SimAV: what is your really needs ?
[17:57] <__jack__> s/ly//
[17:59] <SimAV> __jack__, if i know that AVX increases x264 speed significantly (allowing realtime compression of PAL footage to sth like 2000kbit/s x264) that would mean that getting a 2820QM is the way to go
[17:59] <SimAV> if the increase is just a factor of 1.1 or so, this does not suffice and i shouldn't spend any money now
[18:02] <SimAV> __jack__, but the same applies to webm/vp8. So if either codec is running significantly faster (e.g. twice as fast) when having AVX, this would be interesting to me
[19:17] <baegle> I have collection of AVIs of the BBC Connections series but they're all busted. VLC reports broken indexes and can't play most movies past 29s.
[19:18] <baegle> I tried to convert them to mpeg, and the index is fixed but the video is pretty damaged. Is there anyway I can try to repair these files?
[19:27] <kc8hfi> how do you get the tee command to work/
[19:27] <kc8hfi> ?
[19:27] <c_14> The pseudo-muxer?
[19:28] <kc8hfi> yeah,
[19:28] <c_14> https://trac.ffmpeg.org/wiki/Creating%20multiple%20outputs#Teepseudo-muxer
[19:28] <kc8hfi> i think i just got something out of order ffmpeg -f v4l2  -video_size 720x480  -i /dev/video0  -f pulse -i default -pix_fmt yuv420p -f tee   output.mkv | udp://localhost:123
[19:29] <c_14> needs more quotes
[19:29] <c_14> and the udp stream needs a format specifier
[19:31] <kc8hfi> is there a page somewhere that will tell me in plain english what the parameters mean,  like what does -c mean?
[19:39] <kc8hfi> this is what i got,  http://fpaste.org/174038/22124752/  let me try and explain it.  line 2 is the video,  its a v4l2 input file,  specified the size, and its coming from /dev/video0.   its named video.
[19:40] <kc8hfi> line 3,  is another input,  pulse,  supposed to be coming from the default pulse source and its named audio.
[19:41] <kc8hfi> line 5 is where the output will be teed,  line 6 and 7 sets up a map of the video and audio to be put into the output.  line 8 is the 2 things that the output needs to be sent to
[19:43] <kc8hfi> theres no file or directory called default in the audio line.  and there should be since its there without the tee
[19:48] <grepper> Never ssen -c audio etc. -c is codec, so -c:a mp3 uses mp3 for audio. -c:a copy means don' t encode that stream ...
[19:49] <pseudonomadic> I'm trying to re-encode a video to crop & shrink file size, but I only get a single frame of video in the output file.  The same command line has worked properly every other time I've re-encoded, so I assume the input file is partially corrupted enough to throw off ffmpeg, although it plays properly in VLC.
[19:49] <kc8hfi> where did yiou see -c meaning codec?
[19:50] <c_14> kc8hfi: it's in the manpage
[19:50] <c_14> and the command as you have it in that paste should throw errors
[19:50] <kc8hfi> http://linux.die.net/man/1/ffmpeg  no -c here,  is this not the right man page or something?
[19:50] <pseudonomadic> When I ran ffprobe on the problematic input file, frame number 7 has a pkt_duration of 34375
[19:50] <pseudonomadic>  (vs ~1000 for all other frames)
[19:50] <kc8hfi> it does, just trying to figure out why
[19:51] <c_14> https://ffmpeg.org/ffmpeg.html#Main-options
[19:57] <pseudonomadic> c_14: re-running it to get a clean output log
[19:58] <pseudonomadic> c_14: http://pastebin.com/KxDigE6n
[19:59] <pseudonomadic> I pressed 'q' to stop it early, let me know if you want me to let it finish
[20:00] <kc8hfi> so you just have 2 choices for the stream_specifier,  a for audio and v for video?
[20:00] <c_14> https://ffmpeg.org/ffmpeg.html#Stream-selection
[20:00] <c_14> pseudonomadic: and that output only has 1 frame?
[20:01] <pseudonomadic> c_14: possibly 2, it starts with a black screen, then displays one frame for the rest of the video
[20:01] <pseudonomadic> audio plays back as expected
[20:03] <c_14> How many frames long is R11.mp4 ?
[20:04] <pseudonomadic> c_14: "coded_picture_number=39449" is the last line from ffprobe
[20:07] <cousin_luigi> Is arm6l supported by ffmpeg?
[20:08] <pseudonomadic> c_14: here's the full output of 'ffprobe -show_frames R11.mp4'  https://drive.google.com/file/d/0B7ShDeQ59wCTS09zZXlBclpuZEk/view?usp=sharing
[20:09] <pseudonomadic> and here's the beginning of the output file R00.mp4: https://drive.google.com/file/d/0B7ShDeQ59wCTa0h1aE43Vl93SVE/view?usp=sharing
[20:09] <JEEB> cousin_luigi, I would be surprised if it wasn't buildable for that arch
[20:10] <JEEB> although I have mostly built it for stuff that supports NEON because without that it's very, very slow
[20:10] <cousin_luigi> JEEB: I wanted to see if the patch was still necessary and I see it is.
[20:10] <JEEB> what kind of patch?
[20:11] <c_14> pseudonomadic: I get a lot of frames for that
[20:11] <c_14> It's very laggy, but it plays
[20:11] <c_14> What video player are you testing with?
[20:11] <pseudonomadic> VLC
[20:11] <c_14> Try using ffplay
[20:11] <cousin_luigi> JEEB: https://pmbs.links2linux.de/package/view_file/home:Aloysius:branches:Essentials/ffmpeg/ffmpeg-2.2.4-arm6l.patch?expand=1
[20:12] <JEEB> VLC IIRC doesn't use lavf by default for mp4
[20:12] <cousin_luigi> JEEB: although it's not like it doesn't build...
[20:12] <JEEB> and if one wants to try a lavf based player I think mpv would be the least retarded
[20:12] <cousin_luigi> I wonder if this has something to do with symbol visibility
[20:12] <JEEB> cousin_luigi, oh for fuck's sake with custom TLS certs
[20:12] <JEEB> and HTTP doesn't work
[20:12] <JEEB> yeah, I'm not going to do an exception for a place I have no idea of
[20:13] <pseudonomadic> c_14: yeah, ffplay gives me the same result as you - very low output frame rate
[20:13] <pseudonomadic> source file is ~60fps
[20:13] <c_14> The output fps is apparently 1.33...
[20:14] <cousin_luigi> JEEB: it's not self-signed, it just uses a nonstandard CA
[20:14] <c_14> Can you either build from git HEAD or try with a static build?
[20:14] <JEEB> cousin_luigi, same end result
[20:14] <JEEB> when you self-sign you are using yourself as the CA :P
[20:15] <cousin_luigi> JEEB: http://pastebin.com/raw.php?i=KBHWS6e2
[20:15] <pseudonomadic> c_14: I'm running the static build for windows
[20:15] <JEEB> oh, it enables some timer stuff
[20:15] <cousin_luigi> JEEB: some distros package it
[20:15] <cousin_luigi> JEEB: timer?
[20:16] <JEEB> > /libavutil/arm/timer.h
[20:16] <cousin_luigi> oh sorry:)
[20:16] <pseudonomadic> c_14: ffmpeg-20130329-git-551f683-win64-static
[20:16] <JEEB> cousin_luigi, poke ffmpeg-devel or so with it if you think it's worth it
[20:16] <cousin_luigi> JEEB: it has been on trac for years I believe
[20:16] <c_14> pseudonomadic: sure, but that's from march of last year
[20:16] <JEEB> well trac isn't very much looked upon for patches
[20:16] <JEEB> patches go to the mailing list
[20:17] <pseudonomadic> c_14: yeah, just noticed that.  updating now & trying again
[20:17] <JEEB> also poking the -devel IRC channel with it is often useful if it's correct as far as the patch goes
[20:17] <JEEB> of course first check if that patch hasn't been merged into current master
[20:18] <ac_slater> hey all. I have some video data that I was to mux into a format via libavformat. Precisely, I have some h264 and aac data that I want to mux into an MPEGTS bitstream. Any advice?
[20:18] <ac_slater> s/into a format/into mpegts/g
[20:18] <cousin_luigi> JEEB: no, it hasn't.
[20:18] <kc8hfi> absolute barebones,  http://fpaste.org/174048/22127083/   output file #0 doesn't contain a stream.  it don't contain a stream at first, because thats the place where it gets stored.  the second file is the udp stream
[20:19] <c_14> kc8hfi: What's the problem?
[20:20] <kc8hfi> c_14:   output file #0 does not contain any stream
[20:20] <kc8hfi> that command doesn't look wrong to me
[20:20] <c_14> try adding -map 0 ?
[20:22] <kc8hfi> if you do that, don't you hafta specify a codec somewhere?
[20:22] <kc8hfi> i don't really care what the codec is,  whatever ffmpeg uses is just fine with me
[20:22] <c_14> no
[20:23] <ac_slater> I really wish there was a libavformat/codec/etc channel. :9
[20:23] <ac_slater> :( *
[20:23] <pseudonomadic> c_14: current static build works without issues, sorry for wasting your time :(
[20:23] <pseudonomadic> and thanks for the help!
[20:26] <kc8hfi> http://fpaste.org/174050/27421142/   gives error about encoder (codec none) not found for output stream #0:0.  thats why I asked about the codec
[20:26] <c_14> try "output.mkv|[f=matroska]udp://localhost:1234"
[20:27] <c_14> And whenever possible include all console output as well in your pastes
[20:30] <kc8hfi> http://fpaste.org/174051/27767142/  kind of interesting about the error
[20:31] <c_14> Can you try with a recent git static build?
[20:32] <kc8hfi> yeah, let me go get a static build....
[20:33] <kc8hfi> this is the correct location?  http://johnvansickle.com/ffmpeg/
[20:33] <c_14> yep
[20:37] <kc8hfi> http://fpaste.org/174053/42212816/
[20:38] <kc8hfi> i put the version and then the command.
[20:39] <kc8hfi> i didn't have the vcr playing at the time, but I don't think that matters
[20:39] <c_14> Ok, not sure why it's not defaulting... just add -c:v libx264 it should be the default anyway
[20:39] <kc8hfi> it got pissed about the libx264 when i tried that before.
[20:39] <kc8hfi> but i'll try that with this static build
[20:45] <kc8hfi> that didn't crash and burn  that time.  so far so good, now time to figure out how to play the stream with something
[20:56] <kc8hfi> i'm guessing the static build doesn't have pulse in it?
[21:03] <kc8hfi> http://fpaste.org/174060/42212969/  its like the params are out of order or something
[21:05] <kc8hfi> i don't know how to add a second input
[21:06] <kc8hfi> i can do it without the tee
[21:11] <klaxa> move the -c:v libx264 after all input options
[21:11] <klaxa> i.e.: ffmpeg -f v4l2 -video_size 720x480 -i /dev/video0  -f pulse -i default -c:v libx264 -c:a:0 vorbis -f tee -map 0   "output.mkv|[f=matroska]udp://localhost:1234"
[21:12] <klaxa> kc8hfi: ^
[21:12] <kc8hfi> i haven't tried that yet, might as well...
[21:15] <kc8hfi> ok, so that worked.
[21:17] <kc8hfi> the audio is going into the first output, that file.  but I dont' think its going into the stream, because the stream's audio and video aren't in sync.  the file's audio and video are in sync, thats what makes me think that
[21:20] <kc8hfi> for the line in volume, do you turn it all the way up?  or whats a good level so the recorded audio sounds good?
[00:00] --- Sun Jan 25 2015