[Ffmpeg-devel-irc] ffmpeg.log.20130625

[burek]

[02:58] <bencc> does the static build include rtmp support?
[03:03] <Zeranoe> FFmpeg seems to be speeding up the video when it is encoding, the input is an avisythn script
[06:49] <ilove11ven> bencc: yes
[07:17] <bencc> ilove11ven: I'm trying to save rtmp stream as flv file with
[07:17] <bencc> ./ffmpeg -loglevel debug -i rtmp://127.0.0.1/oflaDemo/test test.flv
[07:17] <bencc> but getting: Operation not permitted
[07:18] <bencc> full output: http://dpaste.com/1269876/
[07:18] <bencc> so I thought maybe rtmp support is missing
[08:27] <ilove11ven> bencc: it seems to be an accessibility error on server. Just wireshark it.
[08:29] <bencc> ilove11ven: I've tried with both rtmplite and red5
[08:29] <bencc> I'll use wireshark. thanks
[08:31] <ilove11ven> bencc: So they both worked. I have dumped rtmp stream from Adobe's server 4.0. May I know what server are you using?
[08:33] <bencc> tested rtmplite and red5
[08:34] <bencc> both don't work for me
[08:34] <bencc> I'm able to publish and play with browsers
[08:34] <bencc> but can't use ffmpeg with it
[08:37] <bencc> ilove11ven: does ffmpeg support the nellymorser codec?
[08:37] <bencc> I wasn't sure so I used speex
[08:37] <ilove11ven> ok. maybe we could check the dump, and see what happened.
[08:38] <steinchen> moin.. id like to take a shot every 3 seconds... for 5 seconds i use -r .2  .. and for 3?
[08:38] <bencc> ilove11ven: what dump?
[08:38] <ilove11ven> wireshark
[08:39] <bencc> ilove11ven: I'll install it and test
[08:42] <ilove11ven> bencc: ffmpeg -codecs tells nellymorser should be supported.
[08:44] <bencc> trying
[08:45] <ilove11ven> bencc: I have published stream to adobe's server using FFmpeg, but failed to do so, since a command FFmpeg is sending is not supported on the server side. Commenting it out solves the problem. You case might be similar.
[08:45] <bencc> what command is failing?
[08:46] <ilove11ven> half a year ago. Let me check the code.
[08:47] <bencc> ilove11ven: I see nellymoser in ffmpeg -codecs
[08:47] <ilove11ven> yes. That suppose to work, but i have not tried it.
[08:48] <bencc> http://dpaste.com/1270087/
[08:49] <bencc> this is what I'm getting
[08:49] <bencc> Proto = rtmp, path = /oflaDemo/test, app = oflaDemo, fname = test
[08:50] <bencc> what "fname = test" means?
[08:51] <bencc> ilove11ven: what syntax do you use?
[08:51] <ilove11ven> file name. That's explained in RTMP spec
[08:51] <bencc> I don't have filename
[08:52] <bencc> I only have app (oflaDemo) and stream (test)
[08:54] <steinchen> or should i use x/60 ?
[09:06] <ilove11ven> bencc: ffmpeg -i rtmp://192.168.42.19:1935/vod/mp4:sample2_1000kbps -f mp4 -vcodec copy /dev/null
[09:06] <ilove11ven> bencc: just setup fms, which has not been used for months. huuuu
[09:08] <ilove11ven> bencc: I expect a wireshark capture will reveal which command failed, then we know what to do next.
[09:53] <RSDRSDRSD> why does ffmpeg -y -i ¨rtmp://5.178.66.238:80/sexdumpertlive/sexdumpert/channel1 stop=1 live=1¨ give an error and without ¨ it is working
[09:57] <RSDRSDRSD> what is the best method to get screenshots of a live rtmp stream
[09:57] <RSDRSDRSD> can find a working solution
[09:57] <RSDRSDRSD> can´t
[10:08] <steinchen> how may i take a screenshot every 3 seconds?
[10:11] <saste> steinchen, select=gte(t, selected_n*3)
[10:44] <esperegu> Hi. I want to stream my desktop audio over rtsp but I get a high cpu load (around 27). I have the following commandline: [[[ffmpeg -vn -f alsa -i pulse -acodec libvo_aacenc -ar 16000 -ac 1 -b 28000 -f rtsp -metadata title=desktop http://localhost:554/flvplayback ]]]. when I run vlc like this [[[ cvlc -vvv pulse://alsa_output.pci-0000_00_1b.0.analog-stereo.monitor --sout '#transcode{acodec=mp4a,ab=22,channels=1}:rtp{sdp=rtsp://0.
[10:44] <esperegu> 0.0.0:8080/test.sdp}' ]]] the load is only about 2. Anyone knows how to solve this high load for ffmpeg?
[12:20] <superware> vlc udp://@:1234 works while ffplay udp://127.0.0.1:1234 shows nothing but lots of errors such as "fPES packet size mismatch" etc, any ideas?
[12:32] <superware> anyone? :|
[12:41] <superware> Mavrik: hi
[13:14] <esperegu> FYI:
[13:14] <esperegu> $ git clone --depth 1 git://source.ffmpeg.org/ffmpeg
[13:14] <esperegu> Cloning into 'ffmpeg'...
[13:14] <esperegu> fatal: read error: Connection reset by peer
[13:17] <saste> esperegu, git is dow
[13:17] <saste> *down
[13:17] <esperegu> saste: I saw that. didn't know if that was known ;-)
[13:54] <superware> Mavrik?
[14:28] <JustinInTheFlesh> Anyone familiar with stream via rtp from a direct show device?
[15:31] <superwar> Mavrik?
[15:33] <sacarasc> There's no one by that nick in the channel.
[15:46] <jure> is this still true? http://transcoding.wordpress.com/2011/11/16/careful-with-audio-resampling-using-ffmpeg/
[15:49] <ubitux> the defaults changed
[15:49] <ubitux> so it should not be the case
[15:50] <jure> does it use ssrc now?
[15:50] <ubitux> no, we improved the defaults
[15:52] <ubitux> there is a way to use sox resampler though
[15:58] <superware> "vlc.exe udp://@:1234" works while "ffplay udp://127.0.0.1:1234" shows nothing but lots of errors such as "fPES packet size mismatch" etc, any ideas?
[16:08] <superware> Mavrik: hi, you here?
[16:15] <superware> I have a H.264 TS over UDP, "vlc.exe udp://@:1234" works while "ffplay udp://127.0.0.1:1234" shows nothing but lots of errors such as "fPES packet size mismatch" etc, any ideas?
[16:48] <bencc> I'm able to get mp3 file from a live rtmp stream: ./ffmpeg -i rtmp://127.0.0.1/audio/test -rtmp_live 1 -ar 44100 -f mp3 test.mp3
[16:48] <bencc> when replacing test.mp3 with pipe:1 I'm exepcting to get the stream in real time
[16:49] <bencc> but I only see the output when the rtmp stream ends
[16:51] <klaxa> bencc: why are you using pipe:1 ?
[16:51] <klaxa> and not pipe: or pipe:0?
[16:53] <bencc> klaxa: isn't pipe:0 for stdin and pipe:1 for stdout?
[16:54] <klaxa> uh...
[16:54] <klaxa> nvm me then, lol
[16:54] <bencc> I'm getting the same issue with avconv
[16:54] <bencc> so ffmpeg doesn't know that the stream is live in real time
[16:54] <bencc> I've tried "rtmp_live 1" but I'm not sure how to use it
[17:03] <JuxTApose> hey guys, I am a blender developer and I am looking to improve our game engine ffmpeg video caching routine...
[17:03] <JuxTApose> any suggestions on some good code examples?
[17:06] <Mavrik> what do you mean by video caching?
[17:07] <JuxTApose> mavrik let me get the source code page...on the game engine you can cache aka prefetch video to be played...
[17:08] <JuxTApose> the issue to fix is the cache is dumped on looping and it creates a seconds versus frames issues that makes the loops skip frames...
[17:08] <JuxTApose> it counts seconds then dumps the cache before all the frames have been played then it starts the playback from beginning to create the loop
[17:08] <beginner> hi everybody, im having trouble with -t parameter. When im recording audio, the FFMPEG seems to ignore this parameter. What am I doing wrong? "ffmpeg -t 3 -f dshow -i audio="Mikrofon (Conexant SmartAudio H" -f dshow -i audio="virtual-audio-capturer" -y -f flv test.flv
[17:09] <Mavrik> beginner, put it after "-i" for starters :)
[17:09] <JuxTApose> let me look at the code and get you some line numbers, it's pretty simple yet creates the bad skip side effect in seamless loops...
[17:10] <JuxTApose> the operative word here is "seamless"
[17:10] <JuxTApose> such as an ocean wave loop
[17:11] <beginner> oh yea it worked thanks. I thought the order of the parameters are not taken into account
[17:14] <Mavrik> beginner, ffmpeg is rather strict when it comes to parameter order
[17:15] <Mavrik> ffmpeg <parameters applied to input> -i .. <parameters applied to 1st output> file.mp4 <parameters applied to 2nd output> file2.mp4 ...
[17:17] <beginner> I see
[17:19] <Mavrik> JuxTApose, sorry, but it's not really clear what your usecase and problem is or just how are you using ffmpeg
[17:20] <JuxTApose> I am looking at the source code to find the routine...I'm away from my normal dev computer so it's taking me a minute...
[17:21] <JuxTApose> essentially the user case is that seamless animation looping as in a loop that have ending and beginning frame aligned to create a continual loop without misalignment of motion
[17:22] <JuxTApose> in the file that is using ffmpeg to playback that loops it is evaluating seconds instead of frames to end the loop...any frames that are a fraction of a second on the end of the animation are not played because the cache is purged on a even second count...
[17:23] <JuxTApose> so a seamless animation of 5 seconds and 10 frames is looped at the 5 second mark, not the 5 second + 10 frames mark...
[17:23] <JuxTApose> so I am looking for some source code examples that properly handling ffmpeg video playback caching...
[17:24] <JuxTApose> i'm looking back at the source code now, the routine obviously shows this issue...
[17:28] <Mavrik> uh ok
[17:29] <Mavrik> you're calling out to the ffmpeg executable or are you linking against libav libraries_
[17:31] <JuxTApose> linking for those still here....
[17:31] <JuxTApose> http://www.letworyinteractive.com/blendercode/d1/d9a/VideoFFmpeg_8cpp_source.html
[17:32] <JuxTApose> offending line I think is 768
[17:33] <JuxTApose> it dumps the cache and starts playback via a seconds reference not an actual frame count reference if the source is seconds plus a fraction of a second in playback...
[17:33] <JuxTApose> normal video it will not be noticed only in seamless animation loops will this be noticable
[17:48] <JuxTApose> wb Mavrik
[17:49] <JuxTApose> blender links to the libraries
[17:49] <JuxTApose> the offending line looks like #768 in http://www.letworyinteractive.com/blendercode/d1/d9a/VideoFFmpeg_8cpp_source.html
[17:50] <JuxTApose> the test true dumps the cache and starts playback via a seconds reference not an actual frame count reference if the source is seconds plus a fraction of a second in playback...
[17:50] <JuxTApose> normal video it will not be noticed only in seamless animation loops will this be noticable
[17:53] <sobaah> hi all
[17:53] <JuxTApose> Ok, so what's the question again?  Can anyone link me to some source code that efficiently handles video playback caching at the frame level from the ffmpeg linked libraries?
[17:55] <sobaah> having a small issue with trying to sync an mp3 from an mp3 recorder and a video from a DV camera that output in mpg format.
[17:55] <sobaah> Camera:
[17:55] <sobaah> Video framerate: 59.940059, original audio sample rate: 48000kHz
[17:55] <sobaah> MP3 recorder: audio sample rate: 44100kHz
[17:56] <sobaah> Issue is that the video is 6.5hours long and by the end of the video, the audio is out of sync by 1.3seconds
[17:56] <jure> JuxTApose: why are you not in #ffmpeg-devel ?
[17:57] <sobaah> can anybody let me know how to fix this?  I've tried changing the sample rate and using other CLI utilities to change the speed but it's not working well
[17:57] <JuxTApose> O.o  cuz I didn't know it existed....i found this on a search..I'm headed there now, thanks jure
[17:58] <ubitux> that's the correct channel
[17:58] <ubitux> to ask for API
[17:58] <ubitux> API usage*
[17:58] <ubitux> (#ffmpeg)
[18:00] <sobaah> does anyone know of a channel that might deal with these issues? please
[18:02] <brontosaurusrex> sobaah, audio is longer?
[18:03] <sobaah> hey brontosaurusrex, yes the audio seems to be a bit longer
[18:03] <sobaah> I have to apply a speedup of 1300ms for it to sync
[18:03] <brontosaurusrex> ok, then align the two in your favorite editing app
[18:03] <brontosaurusrex> and play until its visible out of sync
[18:03] <sobaah> is there a common issue I am not seeing though?
[18:03] <brontosaurusrex> then just cut out the audio until it fits
[18:03] <sobaah> the issue is actually not just in the length
[18:04] <sobaah> it's that the audio is playing slower
[18:04] <brontosaurusrex> sobaah, probably dv capturing is missing a frame or two?
[18:04] <sobaah> hmm, I see
[18:04] <sobaah> the sample rates are different from the original video file, 48000 vs the one in the mp3, 44100
[18:04] <brontosaurusrex> also the devices may run at slightly different speeds
[18:05] <sobaah> hmm, is this an issue that I could research? I mean...like the common playback speed differences in PAL vs NTSC, etc
[18:05] <brontosaurusrex> well, sample rate is usually choosen depending on your final format needs
[18:05] <sobaah> the difference I am seeing is unlike any ratio I've ever seen
[18:06] <sobaah> hmm, I see bron
[18:06] <sobaah> err
[18:06] <brontosaurusrex> thats not pal vs ntsc issue
[18:06] <sobaah> brontosaurusrex
[18:06] <sobaah> right, I underatand
[18:06] <sobaah> I was just asking if there is some sort of known issue that I can look into
[18:06] <sobaah> similar to that
[18:07] <sobaah> because at this point I'm just doing it "by ear", it sync near the end of the 6.5h video with a 1300ms negative delay
[18:07] <brontosaurusrex> you did hear the pro devices are usually locked by some sort of signal right?
[18:07] <sobaah> it syncs up near*
[18:07] <sobaah> but at the beginning, the video and audio are pretty well synced
[18:07] <sobaah> or at least not off by 1.3s!
[18:07] <sobaah> =D
[18:07] <sobaah> hmm, brontosaurusrex, locked how?
[18:07] <jure> https://kb.speeddemosarchive.com/Progressive_Audio_Desync
[18:08] <brontosaurusrex> a generator of some sort
[18:08] <brontosaurusrex> used to be a called house-sync or black-burst
[18:08] <sobaah> hmm, we're just using a regular digital camera and an mp3 recorder
[18:08] <sobaah> for some training
[18:08] <brontosaurusrex> yeah, that should work well, unless your clips are 7 hours long
[18:09] <sobaah> heh, indeed, brontosaurusrex
[18:09] <sobaah> thanks, jure
[18:10] <jure> their solution is a bit too complicated though
[18:10] <sobaah> seems like that is a defined concept of what I am seeing/hearing
[18:10] <sobaah> Ive done more complicated things =D
[18:10] <jure> unnecessarily complicated
[18:10] <sobaah> ffmpeg is amazing, but it's a bit iffy at changing speed of audio
[18:11] <sobaah> hmm, what would you suggest?
[18:11] <sobaah> using setpts?
[18:11] <jure> virtualdub has a nice feature to force audio/video sync
[18:11] <sobaah> yep, seen that also
[18:11] <sobaah> thing is I dont want to have to reencode
[18:11] <sobaah> I just want to fix the audio file itself
[18:11] <sobaah> then merge it with the video file
[18:11] <jure> then open the video file in audacity, it will auto-strip the audio out of it
[18:11] <sobaah> hmm, I guess VirtualDub should met me do it
[18:12] <sobaah> met=let
[18:12] <sobaah> dont have audacity =.
[18:12] <sobaah> =/
[18:12] <sobaah> is it free?
[18:12] <jure> yes
[18:12] <sobaah> I forget
[18:12] <sobaah> ahh, I see
[18:12] <sobaah> I can strip the audio using ffmpeg too
[18:12] <brontosaurusrex> sobaah, avisynth?
[18:12] <sobaah> I have the video file and the mp3 from the recorder
[18:12] <sobaah> I just want to fix the audio to the correct length
[18:12] <jure> you can then either resample or change the audio speed in audacity
[18:13] <sobaah> mmm have never tried avisynth, brontosaurusrex, Ive heard of it though
[18:13] <sobaah> mm, ok jure
[18:13] <brontosaurusrex> sobaah, basically:
[18:13] <sobaah> I tried a resample in ffmpeg but it didnt make much of a difference
[18:13] <jure> then you save that file and remux with ffmpeg
[18:13] <sobaah> sounds good jure
[18:13] <sobaah> I figured I would need to do so
[18:13] <brontosaurusrex> v=video.avi, a=audio.wav, resample(a)ToFixLenght, resample(a)to48kOr something, dub(v,a)
[18:14] <sobaah> o right avisynth is pretty much a scripting language?
[18:14] <brontosaurusrex> but you will need to reencode
[18:14] <brontosaurusrex> yes
[18:14] <sobaah> and it ties in with VD
[18:14] <brontosaurusrex> yes
[18:14] <sobaah> yeah, never tried it but sounds cool
[18:14] <sobaah> jure, was more just wondering, initially, if there was something simple I was missing
[18:14] <jure> well, yeah, there is
[18:15] <sobaah> heh
[18:15] <sobaah> the PAD?
[18:15] <sobaah> Progressive Audio Delay?
[18:15] <jure> desync
[18:15] <sobaah> DeSync*
[18:15] <sobaah> heh yah
[18:15] <sobaah> that?
[18:15] <jure> no
[18:15] <sobaah> or was it something else?
[18:15] <sobaah> oh
[18:15] <jure> sample rate difference causes it
[18:16] <sobaah> thanks for the help, brontosaurusrex
[18:16] <sobaah> hmm, I figured as much
[18:16] <sobaah> 48000 vs 44100
[18:16] <jure> those 4000 samples get piled up eventually
[18:16] <sobaah> but Ive changed it using both eac3to and ffmpeg
[18:16] <sobaah> and still, it didnt seem to affect the delay
[18:17] <sobaah> can I resample the mp3 directly or would it be preferable to change it to a different format?
[18:17] <brontosaurusrex> samplerate shouldn't make a difference
[18:17] <jure> transcoding loses fidelity
[18:17] <sobaah> I converted the mp3 to an ac3 using ffmpeg, but when playing the ac3 in VLC, the total runtime jumps around like it's trying to be calculated
[18:18] <sobaah> right, I understand about the loss during compression/transcoding
[18:18] <sobaah> but I'm just trying to get it to work firsty
[18:18] <sobaah> first*
[18:18] <jure> probably bad container format
[18:18] <sobaah> brontosaurusrex: hmm, you think the sample rate shouldnt make a diff?
[18:18] <jure> (total runtime jumping around)
[18:18] <sobaah> mm, I see jure
[18:18] <jure> I don't know why you'd convert mp3 to ac3 though
[18:19] <sobaah> it was to try to let eac3to change the sample rate
[18:19] <sobaah> the documentation says it works with "any format"
[18:19] <brontosaurusrex> sobaah, if i fart 3 times per second or 100 times per second, its still just a second
[18:19] <jure> sure, that's why you had to transcode it from mp3 right?
[18:19] <sobaah> but I've tried various mp3s and it just ignores all options for doing stuff
[18:19] <sobaah> yep, jure
[18:22] <jure> *sigh*
[18:22] <sobaah> brontosaurusrex: I agree but I believe sample rate still has something to do with it
[18:22] <brontosaurusrex> how?
[18:22] <sobaah> also, a bit vulgar choice of an example
[18:22] <sobaah> im not sure atm but I found a few threads pointing to changing sample rate in Audacity to fix audio sync issues
[18:22] <sobaah> i'll give audacity a go, thanks both of you, jure and brontosaurusrex
[18:22] <jure> fiddling around with sample rate to fix audio sync is tricky, just so you know
[18:23] <sobaah> jure, hmm
[18:23] <sobaah> so just change tempo instead?
[18:23] <sobaah> or what should I go for?
[18:23] <jure> you should play around with audacity for a while, see what you can do
[18:24] <sobaah> but with the sample rate right?
[18:24] <jure> but work on a smaller patch of the video, not the entire 6 hours
[18:24] <sobaah> it's hard not to work with the full length
[18:24] <sobaah> simply because the audio becomes desynced  near the end
[18:24] <sobaah> or at least that's where I see it more noticeably
[18:24] <brontosaurusrex> sobaah, http://avisynth.nl/index.php/FPS , see PAL +4% conversion example
[18:24] <sobaah> 1300ms difference
[18:25] <brontosaurusrex> its basically dual-resampling
[18:25] <brontosaurusrex> but as allready mentioned your sync may not be gradually and linearly lost, but missing frames on places
[18:25] <brontosaurusrex> + unknown
[18:26] <sobaah> hmm
[18:26] <sobaah> going to go check if near the middle of video
[18:26] <sobaah> there is a 1300/2 difference
[18:26] <sobaah> should have done that
[18:26] <brontosaurusrex> yes
[18:27] <sobaah> actually, I realized...
[18:27] <sobaah> this camera writes oddly
[18:27] <sobaah> it can only have 2Gb of data before it creates a new file within the camera
[18:27] <sobaah> probably due to memory or file cluster reasons
[18:28] <jure> maybe the file can't be larger than 2GB
[18:28] <sobaah> right, due to the reasons above
[18:28] <jure> fat16?
[18:28] <brontosaurusrex> maybe the drives are formated to fat32 or something
[18:28] <sobaah> I think, perhaps, when it's finalizing one file and starting the next one, it loses some frames there
[18:28] <jure> yeah, fat32
[18:28] <sobaah> that's what I was thinking too
[18:28] <sobaah> file cluster reasons
[18:29] <brontosaurusrex> nah, i did a lot of video with a sony camera like that, never missed a frame
[18:29] <sobaah> when I export the file out, the program merges the files and outputs them into an .mpg container
[18:29] <Mavrik> uh
[18:29] <sobaah> mm, yep, this is a SONY camera
[18:29] <Mavrik> do you have a camera that records into AVCHD?
[18:29] <sobaah> it's an older camera, Mavrik
[18:29] <Mavrik> (sorry, came slightly late)
[18:29] <sobaah> so prob not AVCHD
[18:29] <sobaah> heh, np always appreciate any help
[18:30] <sobaah> im just having some desync issues
[18:30] <Mavrik> anyway, I had problem with Sony cameras which recorded into AVCHD which was also split into 2GB chunks
[18:30] <sobaah> 6.5hour long video has a 1300ms difference between it and an mp3 recorder that we used at the same time
[18:30] <Mavrik> and merging it with anything that wasn't proprietary Sony software caused some video frames to be lost on file boundaries
[18:30] <sobaah> this is SONY's PMB software though
[18:30] <sobaah> I'm not doing it on my own
[18:31] <sobaah> though I would if needed
[18:31] <sobaah> it automatically exports the 2GB files and puts them all together
[18:31] <sobaah> into 1 large 15GB file (for this particular vid)
[18:31] <sobaah> lunch
[18:31] <sobaah> thanks everyone!
[21:19] <z_sat> I'm trying to set up some options for a AVCodecContext using AV_CODEC_ID_H264 and I can't seem to find av_opt_set function.  Is that fcn avail in V1.2 or has it been replaced?
[21:31] <superware> this is what I get when trying to ffplay an h.264-ts-udp stream http://pastebin.com/bFUvBwHU which actually plays nicely with VLC... any ideas?
[22:09] <nwave> is it possible to take an rtsp stream and store it to disk in 5 second segments with the timestamp of the file as the name of the segment file?
[22:14] <saste_> nwave, not yet, but check the segment muxer
[22:17] <nwave> so can I do http live streaming HLS with ffmpeg, like I can with VLC?
[22:18] <saste_> HLS braindead streaming? yes
[22:18] <JEEB> ffmpeg can do the HLS part of it, and then you will have to have a server serving those files it made
[22:19] <nwave> cool.  are there any examples or some documentation on that?
[22:19] <saste_> nwave, segment, hls muxers in man ffmpeg-formats
[22:19] <JEEB> and just output to a directory that is then shared by a http server
[22:19] <JEEB> s/shared/served/
[22:19] <sobaah> saste_: man, what did you write there?
[22:19] <sobaah> =P
[22:19] <sobaah> I am a tech geek but wth?
[22:24] <sobaah> oh, thought that said "his muxiers"
[22:24] <sobaah> was like...what?
[22:24] <sobaah> xD
[22:24] <sobaah> muxers*
[22:25] <nwave> is mlaw audio not able to be encapsulated in mp4?
[22:25] <JEEB> no
[22:25] <saste_> nwave, from what i know no raw audio (mlaw counts more or less as raw audio)
[22:26] <nwave> interesting
[22:26] <JEEB> yeah
[22:26] <JEEB> raw audio is only being added to 14496-12 IIRC
[22:26] <JEEB> if only itscj.ipsj.or.jp was alive
[22:26] <JEEB> has been dead for a week now
[22:26] <JEEB> could've checked
[22:26] <nwave> so is there a way I can take this rtsp stream I have that has h264 video and mlaw audio and dump it to an mp4 file and encode the audio in an acceptable format?
[22:27] <JEEB> sure, if you re-encode it into a format that fits then all's fine
[22:28] <nwave> is AAC a normal audio format for mp4?
[22:30] <JEEB> fits just fine, yes
[22:31] <nwave> should this work.... ffmpeg -i out.avi -c:v copy -c:a libfdk_aac -vbr 3 out.mp4
[22:32] <JEEB> not sure what the vbr option is, but if you set some kind of rate control mode and -afterburner 1 it should be fine, yes
[22:34] <nwave> http://pastebin.com/PxwjSbze
[22:34] <nwave> not sure what's it complaining about
[22:36] <JEEB> libfaac doesn't support this output format!
[22:36] <JEEB> is the actual error
[22:36] <LithosLaptop> try libmp3lame
[22:36] <JEEB> basically it's most probably the sample rate
[22:36] <LithosLaptop> should give better quality than libfaac and also fits in mp4 container
[22:36] <axorb> hey guys, I'm trying to segment a video, but it's not working for one of my test files: http://pastebin.com/2f3FEsCh
[22:37] <JEEB> yeah, but if we go there he might as well use fdk-aac
[22:37] <LithosLaptop> yes
[22:37] <JEEB> of course he'd have to compile it first, though
[22:37] <axorb> [nut @ 0x1b17380] Negative pts not supported stream 0, pts -9223372036854775808
[22:37] <nwave> track 1: muxing mp3 at 8000hz is not supported
[22:37] <LithosLaptop> oops
[22:37] <LithosLaptop> you could resamle first
[22:37] <LithosLaptop> resample
[22:38] <nwave> how do I do that
[22:38] <LithosLaptop> -ar
[22:38] <LithosLaptop> -ar 11025
[22:38] <LithosLaptop> I don't know what sample rates are supported
[22:39] <LithosLaptop> think 22050 should work
[22:41] <nwave> ffmpeg -i out.avi -c:v copy -c:a libmp3lame -ar 22050 -b:a 128k out.mp4
[22:41] <nwave> this created the file fine, but I don't hear audio
[22:43] <superware> I have an h264-ts-udp stream being played by VLC --demux ffmpeg udp://@:1234 but not by ffplay udp://127.0.0.1:1234 (log at http://pastebin.com/bFUvBwHU)
[22:43] <llogan> nwave: in what player?
[22:48] <superware> a better log: http://pastebin.com/4navfRNM
[22:49] <superware> only on line 383 it detects the stream #0:0, not still no video
[22:58] <nwave> how can I get ffmpeg to be an rtmp server
[22:58] <nwave> translate a rtsp stream to rtmp?
[23:02] <ThePendulum> Greetings
[23:03] <nwave> llogan: sorry, didn't see you were addressing me.  I figured it out needed to use sample rate 44100
[23:07] <nwave> I'm tryingto set up a HLS segmenting from an RTSP stream but I keep getting this error
[23:07] <nwave> cmd: ffmpeg -i "rtsp://10.2.2.20:81" -c:v copy -c:a libmp3lame -ar 44100 -f segment -segment_list out.list -segment_time 2 -segment_wrap 20 hls5.mp4
[23:07] <nwave> error Output file #0 does not contain any stream
[23:42] <superware> how should I ffplay a h264-ts-udp stream?
[23:50] <superware> I guess probing is quite poor for my network stream
[00:00] --- Wed Jun 26 2013