[Ffmpeg-devel-irc] ffmpeg.log.20140721

burek burek021 at gmail.com
Tue Jul 22 02:05:01 CEST 2014


[00:05] <ac_slater> hey all, I want to make a ffmpeg plugin or something for my metadata stream. I have a custom binary format that I want to associate with some AV data. Is there a plugins guide or a 'extending ffmpeg' guide
[00:14] <c_14> phingers: try using -framerate instead of -r on the input side, also h264 does not look like a valid idct according to the manpages.
[00:29] <phingers> I was on an old server, I switched out to a new rhel6 and the code is working much better
[00:30] <phingers> This is working well now - ffmpeg -r 1/10 -i foo%03d.png -i audio.wav -r 32 -c:v libx264 -strict experimental -b:a 192k -pix_fmt yuv420p -shortest -vf drawtext="fontsize=30:fontcolor=white at 0.9:fontfile=FreeSerif.ttf:shadowy=1:shadowx=1:text='I am getting the hang of ffmpeg\!':x=(w-text_w)/2:y=(h-text_h-line_h)/2"  output.mp4
[00:30] <phingers> But I want to control when the text appears and disappears
[00:30] <phingers> start_number and n don't seem to be working
[00:32] <phingers> plus I have three images, and it seems to be only picking up 2
[00:33] <c_14> try switching from -r in the input to -framerate, using -r can cause weird things to happen
[00:37] <c_14> And judging by the examples from the manpage you might be able to control when the text appears using the enable= expcession
[00:37] <phingers> actually the opposite happened adding in framerate and dropping r caused the filesize to go from 1706623 to 353171 - it lost the audio and skips to 40 seconds in when I hit play
[00:39] <phingers> http://pastie.org/9408017
[00:40] <c_14> LIt's telling me that pastie does not exist
[00:40] <phingers> http://pastie.org/9408019
[00:40] <phingers> is the other one
[00:40] <c_14> ah, now it does
[00:41] <phingers> http://pastie.org/9408021
[00:41] <c_14> The input framerate option, not the output one
[00:41] <phingers> the only difference is the r vs framerate
[00:41] <c_14> ffmpeg -framerate 1/10
[00:41] <phingers> oh, sorry
[00:43] <phingers> that works, I have 6 images though, and it is only picking up the first 5
[00:44] <c_14> Ok, let's start by updating ffmpeg... 0.10.11 is rather how do I put it, dated.
[00:45] <c_14> try grabbing a static build
[00:49] <phingers> http://pastie.org/9408032
[00:49] <phingers> I updated ffmpeg - same issue though
[00:49] <phingers> playing with expression
[00:52] <c_14> Ok, you don't need -strict experimental if you're going to be using libvo_aacenc (though the builtin might be better, use -c:a aac if you want that one). The next thing I'd try is either removing the filter and see if that gives you all the frames and/or removing the audio/the -shortest and see if that helps.
[00:52] <c_14> If all that doesn't help then I remember somebody else having similar issues, but I can't remember how I ended up helping him...
[00:54] <phingers> audio is 4 mins long - so shortest is def needed
[00:54] <phingers> I'm all for simplifying the command.
[00:54] <phingers> I'm going to add in a text to speech web service call, any recommendations?
[00:55] <phingers> Do you mess with ffmpeg for a living?
[00:55] <c_14> Mainly just for fun.
[00:55] <c_14> What do you mean 'add in a text to speech web service call'?
[00:55] <phingers> I want to overlay custom text on top of the audio track
[00:56] <phingers> can I town down the one audio track to be background music?
[00:56] <phingers> tone
[00:57] <c_14> You can cut the audio into three sections and then fade/lower the volume of the middle section.
[00:59] <phingers> I just want the whole audio track to run quieter
[00:59] <phingers> like I do with text, opacity 0.2
[00:59] <c_14> Just use the volume filter.
[01:05] <phingers> okay, I think the -af worked, how do I add in the second audio file at 100% volume?
[01:06] <c_14> you'll probably have to switch to a filter_complex something along the lines of '[0:a]volume:foo[aquiet];[aquiet][1:a]concat:n=2:a=1:v=0[a]' -map '[a]'
[01:07] <phingers> it wont let me add two audio streams I dont think
[01:07] <c_14> hmm?
[01:07] <phingers> I put both and only the audio comes, not the speak
[01:08] <c_14> commandline?
[01:08] <phingers> http://pastie.org/9408058
[01:09] <phingers> does it serialize the audio?
[01:10] <c_14> your current commandline doesn't concatenate the audio, it'll create an output file with two audio streams.
[01:10] <c_14> Which is why I said you'll probably need a filter_complex with the concat in it.
[01:13] <phingers> nailed it. Wow
[01:13] <phingers> amazing, who knew ffmpeg was so powerful
[01:14] <phingers> so I have them combined into one output, but still need to lower the first one's volume
[01:14] <c_14> ye, look at the example I gave you.
[01:15] <phingers> you'll have to break it down for me, it seems complex.
[01:15] <phingers> I have -filter_complex amix=inputs=2:duration=first
[01:17] <c_14> You wanted to mix them, ok. Then [0:a]volume=foo[quiet] <- this'll lower the volume of the audio stream of the first input file and "store" it in quiet, then [quiet][1:a]amix=foo[a] <- this'll take [quiet] and the audio stream of the second input file, mix it and store it in [a], connect those two with a ';' in the filter complex and then add -map '[a]'
[01:19] <phingers> http://pastie.org/9408071
[01:20] <phingers> must be missing something simple
[01:20] <phingers> explanation makes sense
[01:20] <c_14> Ehh, you weren't supposed to take that foo literally.
[01:20] <phingers> lol
[01:21] <c_14> and that amix= part is off as well
[01:21] <c_14> Just to get it straight, do you want to mix the tracks or concatenate them?
[01:21] <phingers> mix
[01:21] <phingers> one is the elevator music in the background, the other is the "doctor" speaking instructions
[01:22] <phingers> what do I replace foo with then?
[01:23] <phingers> somehow I name my inputs?
[01:23] <c_14> '[1:a]volume:<options>[aquiet];[aquiet][2:a]amix:<options>[a];[0:v]drawtext=<options>[v]' -map '[a]' -map '[v]'
[01:23] <c_14> https://ffmpeg.org/ffmpeg-filters.html#volume
[01:23] <c_14> Look at the documentation of each of the filters on what to replace <options> with.
[01:24] <c_14> And yes, you'll need to put the drawtext filter into the filter_complex because I'm pretty sure you can't have a filter_complex and a -vf in the same commandline. One of them gets ignored. That's how it is with multiple -vfs anyway.
[01:28] <phingers> I dont see any options for volume
[01:29] <c_14> volume=volume=0.5 will have the volume for example
[01:30] <phingers> http://pastie.org/9408097
[01:30] <phingers> close
[01:31] <c_14> Your text= has quotations, you'll need to escape them.
[01:31] <c_14> try escaping them once, if that doesn't work, try twice etc etc. It can be a bit tricky sometimes
[01:32] <phingers> the single or double quote?
[01:32] <phingers> single escaped twice gives same error
[01:33] <c_14> The single
[01:34] <c_14> You might be able to replace the single quotes around the text with double quotes.
[01:34] <phingers> I have double quotes encompassing the drawtext?
[01:34] <phingers> those arent needed?
[01:35] <c_14> you can probably get rid of them, those are usually only for escaping on the shell level, the single quotes around the whole filtergraph should take care of that.
[01:36] <phingers> dropping the double, and adding it in around the text did it
[01:37] <phingers> crazy stuff - its working.
[01:38] <phingers> You are quite the master at this. I owe you a beer.
[01:38] <phingers> but it isnt taking the shortest bit
[01:40] <phingers> I want it to be as long as the slides are.
[01:41] <c_14> try placing the -shortest before the map? it should work though...
[01:41] <c_14> That or switch to duration=shortest in the amix options.
[01:41] <c_14> That could be interfering.
[01:44] <phingers> the shortest on the audio portion stops after the talking sound is done
[01:44] <phingers> it needs to be mapped to the length of the video somehow
[01:44] <c_14> ahh
[01:45] <phingers> almost need another complex filter to map those two for shortest?
[01:46] <c_14> The -shortest should be working though... You could always just add a -t output option and forcibly halt if you know how long the stream should be.
[01:47] <phingers> tried shortest before and after the maps
[01:47] <phingers> still taking the full audio length
[01:47] <c_14> I'm gonna try testing something on my machine.
[01:49] <phingers> I can send you my files if you'd like
[01:50] <c_14> Sure, that would help.
[01:55] <phingers> emailed
[01:56] <phingers> bounced too big.
[01:57] <c_14> Mhm, well I think I might have found the cause. If I don't apply the drawtext filter -shortest works, so I'm seeing if I can find a workaround for taht.
[01:59] <phingers> I have to walk the dogs, back as soon as their bladders are empty
[02:02] <c_14> kk
[02:04] <onyx> hey guys, anyone with experience running ffmpeg using cron?
[02:05] <onyx> I cant get it to work... I been researching all day
[02:07] <onyx> Im running in root cron tab
[02:07] <onyx> this doesnt work http://pastebin.com/YsLjzQsR
[02:08] <c_14> onyx: running ffmpeg as root is usually considered overkill/bad practice, I'll look at it in a sec though
[02:08] <onyx> root is the only user i have so far thats why
[02:08] <c_14> phingers: As soon as you add the video into the filter_complex even if you just pass it through the null filter -shortest appears to lose effect. I don't know why, but to me this is probably a bug and I'd report it as such. In your case since you know the output video is supposed to be 25 seconds you can just add -t 25 as an output option and that'll limit it.
[02:09] <c_14> onyx: Does that command work if you run it in your shell?
[02:09] <phingers> onyx: it should run in a cron the same as a command line more or less
[02:09] <phingers> onyx: specify full paths in cron
[02:09] <c_14> ^That too.
[02:10] <onyx> I read that full path thing but I dont understand what that means
[02:10] <c_14> /path/to/ffmpeg
[02:10] <phingers> no ..
[02:10] <onyx> I tried creating a record.sh file and running that
[02:10] <onyx> http://pastebin.com/XSw8JRKC
[02:11] <onyx> if I execute that record.sh it works
[02:11] <onyx> but in the cront job doesnt work
[02:11] <phingers> onyx: /usr/bin/ffmpeg parameters /pathto/var/www/vod/video.mp4
[02:11] <phingers> you need to specify full path for ffmpeg and the output file
[02:11] <phingers> which ffmpeg on your system
[02:12] <onyx> ok one sec
[02:12] <phingers> assuming you use the default one for the system
[02:12] <onyx> is showing this path /usr/local/bin/ffmpeg
[02:12] <phingers> k so put that in your cron
[02:13] <phingers> and specify full path to output
[02:13] <onyx> full path for output file will be /var/www.vod/video.mp4
[02:14] <phingers> c_14: where is the best place to report the bug?
[02:14] <onyx> sorry this path /var/www/vod/video.mp4
[02:14] <c_14> https://ffmpeg.org/bugreports.html
[02:14] <c_14> The important part is near the bottom.
[02:23] <phingers> https://trac.ffmpeg.org/ticket/3789#ticket
[02:24] <phingers> what other video creation tools do you know about?
[02:25] <phingers> basically, I need to have different text fade in and out on each slide, and even stick a video clip in between the slides, with various audio clips as well.
[02:25] <phingers> it almost seems like I would need to run multiple ffmpeg commands to build up the video needed
[02:27] <c_14> Hmm, either that or look at some sort of actual full-on video editing software. Some movie editing thing or something. Don't know of any good ones off the top of my head though.
[02:28] <phingers> needs to be dynamic running off the server
[02:28] <phingers> only other one I know of is https://developer.stupeflix.com/documentation/
[02:28] <c_14> Your best bet might be to programatically combine several ffmpeg commands.
[02:28] <phingers> but their support stinks
[02:29] <phingers> slide one, slide one text, slide one audio - save
[02:29] <phingers> rinse repeat
[02:29] <phingers> combine them all with an background audio into one big file?
[02:30] <c_14> something like that, sure
[02:32] <onyx> ok, it works now
[02:32] <onyx> the only issue is that when I add the `date +%d-%H` to the file name it doesnt work
[02:33] <c_14> I'm guessing cron doesn't like that.
[02:33] <c_14> you could always /bin/bash -c 'ffmpeg [foobar]'
[02:33] <onyx> lost me there
[02:34] <c_14> `date +%d-%H` is shell expansion, cron probably won't do that
[02:34] <c_14> so fix it by spawning a shell using cron and running the command there
[02:35] <Riviera> nah, it's not that
[02:35] <phingers> look at this example: touch "/tmp/$(date +%d-%m-%Y).sql"
[02:35] <Riviera> cron does start a shell, so the backticks won't be a problem (don't use backticks, though, use $())
[02:35] <Riviera> the issue is that % is special for many cron variants
[02:35] <phingers> ^exactly
[02:35] <Riviera> use \% instead of %
[02:35] <c_14> aah, did not know that
[02:35] <c_14> My cron commands tend not to have '%'s in them.
[02:36] <Riviera> "Percent-signs (%) in the comand, unless escaped with backslash (\), will be changed into newline characters [..]" from "man 5 crontab" on this system, which uses Vixie cron.
[02:37] <onyx> ok so I changed it to `date +\%d-\%H`
[02:37] <c_14> Hmm, yeah here too. Good to know.
[02:38] <phingers> no ticks
[02:38] <phingers> parens
[02:38] <phingers> with a dollar sign
[02:38] <phingers> see the example above
[02:39] <onyx> oh so $(date +\%d-\%H)
[02:39] <onyx> ^like that?
[02:39] <phingers> try it out
[02:39] <Riviera> (Also it's typically better to use a date format in filenames that sorts lexicographically the same as chronologically, maybe better %Y-%m-%d.)
[02:41] <onyx> ok, waiting for cron to kick in
[02:43] <phingers> okay, my attempt at the first slides are good, how do I pad it with some space at the begining or end of the audio if I want?
[02:43] <c_14> apad will pad the end
[02:43] <c_14> you can use an aevalsrc=0 and a concat to pad the front
[02:46] <onyx> found this
[02:46] <onyx> You can not use % in the command area. They will need to be escaped and if used with command substitution like the date command you can put it in backticks. Ex. `date +\%Y-\%m-\%d`. Or use bash's command substituion $()
[02:52] <onyx> that did it, cron is working now
[02:52] <onyx> thank you so much guys!
[02:52] <onyx> really appreciate it
[02:53] <onyx> ffmpeg ftw!
[02:54] <Riviera> cool
[02:55] <Riviera> $() is a feature of POSIX sh, not exclusive to bash, though.
[02:58] <Hello71> `` is posix too, right?
[02:59] <Riviera> yup, but older than that
[02:59] <phingers> I have three videos created, now I just mash them together with an audio track?
[03:00] <c_14> yep
[03:00] <phingers> seems to easy
[03:00] <phingers> black screen with audio only
[03:02] <phingers> http://pastie.org/9408208
[03:02] <c_14> You'll probably want to actually concat the videos
[03:03] <c_14> -filter_complex '[0:v][1:v][2:v]concat=n=3:v=1:a=0[v]' -map '[v]'
[03:04] <phingers> I want to keep those video's audio though, and add the new sound in as a low background music
[03:05] <c_14> Are we back at the mixing stage?
[03:05] <phingers> lol
[03:05] <c_14> -filter_complex '[0][1][2]concat=n=3:v=1:a=1[v][a];[a][3]amix[out]' -map '[v]' -map '[out]'
[03:05] <c_14> or something
[03:05] <c_14> ah, wait
[03:05] <c_14> forgot the volume filter
[03:05] <c_14> but you get the jist
[03:08] <phingers> it isnt accepting shortest again
[03:08] <phingers> same issue I'd guess
[03:08] <c_14> I feel like we've been here before.
[03:08] <c_14> Yeah probably.
[03:16] <phingers> its mixing the audio and videos up
[03:17] <c_14> with the concat?
[03:17] <c_14> you might need to specify them seperately
[03:17] <c_14> ie [0:v][0:a] etc
[03:18] <phingers> I'm double checking the raw files.
[03:20] <phingers> http://pastie.org/9408240
[03:20] <phingers> its not picking up audio from this clip generation
[03:21] <c_14> Have you tried playing/ffprobing the file to check it has audio?
[03:21] <phingers> hmm... its maybe just the sublime embeded player
[03:21] <phingers> direct in browser it has audio
[03:21] <phingers> thats weird.
[03:25] <phingers> what did you mean by separately? Sorry I think my brain is turning to mush now with these complex_filters
[03:26] <c_14> If the command is erroring out you can try specifying the inputs for the concat filter seperately, ie instead of specifying the whole file with [0] specify the video and audio streams with [0:a][0:v]
[03:28] <phingers> it showed me clip 3 video with the first video's audio
[03:28] <phingers> then video 2, and clip 1's audio
[03:28] <phingers> then video 2, and clip 2 audio
[03:29] <c_14> ohlol, try what I said with the seperately, that might fix it i hope.
[03:29] <phingers> http://pastie.org/9408257
[03:31] <c_14> And that isn't working?
[03:31] <c_14> As in messing up the audio/video order?
[03:31] <phingers> never even see clip 3 video
[03:32] <phingers> audio is coming out 2,1,3
[03:32] <phingers> video is coming out 3,2-----
[03:34] <phingers> I give up for now. ahhhh! I appreciate your help c_14
[03:34] <phingers> catch you later all.
[03:34] <c_14> np, still don't know why that wouldn't work though
[03:38] <c_14> The one thing that it might be is if the audio and video lengths of the three input videos don't match up that it might concat a bit weirdly, you could see if that might be the issue by using the apad filter in the commands you're using to generate those 3 videos to pad the audio to the length of the corresponding video track.
[08:56] <the_s3m> Hey guys, I would like to use ffmpeg on Windows Phone 8 but I couldn't find any good solution to make it work on Windows Phone, so I would like some developer feedback from someone who knows what needs to be done in order to run ffmpeg on windows phone and how much work would it take to make this work.
[09:50] <JEEBsv> the_s3m: VLC team was doing work regarding it; and FFmpeg now builds with visual c 2013's compiler
[09:51] <JEEBsv> so you could see how #videolan builds its VLC for WinRT
[09:53] <the_s3m> thanks for info
[09:53] <the_s3m> so vlc already has the working arm build?
[09:54] <the_s3m> or is it still in development?
[10:53] <Mavrik> yes, VLC has a working ARM build for awhile now
[11:26] <cdhris> Hello, is there any way to automatically include the filters such as -bsf h264_mp4toannexb if needed without importing it into the command?
[11:28] <cdhris> c_14:  do you know that? :P
[11:34] <Mavrik> no.
[11:35] <Mavrik> but that exact filter won't do anything if it cannot
[11:35] <cdhris> i think when i include it in the streams i shouldn't
[11:35] <cdhris> it doesnt work
[11:38] <cdhris> Yes, confirmed, It doesn'
[11:38] <cdhris> it doesn't work when it don't need it
[12:22] <cdhris> "Stream #0:2[0x102]: Unknown: none ([6][0][0][0] / 0x0006)"
[12:23] <cdhris> Is there any way to stip all the unknown codecs?
[12:23] <cdhris> cause i have problem with this, if i stream has this one and i use -map 0  it will try to include this as well
[12:23] <cdhris> and it will fail
[12:23] <cdhris> strip*
[12:23] <sfan5> -map 0:v -map 0:a -map 0:s
[12:23] <sfan5> that will only select all known video, audio or sub streams
[12:25] <cdhris> Can i ask a noobish question?
[12:25] <cdhris> what data codec exactly is
[12:26] <sfan5> that questions does not make any sense to me, can you rephrase it?
[12:27] <cdhris> I mean in some rtmp streams
[12:27] <cdhris> there is a Data codec as well that you can copy with -dcodec copy
[12:27] <cdhris> what is the data codec exactly
[12:27] <sfan5> hm
[12:27] <sfan5> I didn't even know there is a -dcodec option
[12:28] <sfan5> I'm probably the wrong person to ask then
[12:31] <cdhris> The command you gave me does work
[12:31] <cdhris> the only issue
[12:32] <cdhris> is that if there is a stream that has only audio /video for example and if i write -map 0:s it will fail saying that there is n subtitles to map. My point is to have a static command
[12:32] <cdhris> that can work on all streams
[12:32] <cdhris> and strip the unknown codecs
[13:07] <Diogo> ./ffmpeg -i rtmp://IP/live/STREAMA -i rtmp://IP/live/STREAMB -filter_complex '[0:0] [0:1] ] [1:2] [2:0] [2:1] [2:2] concat=n=3:v=1:a=2 [v] [a1] [a2]' -map '[v]' -map '[a1]' -map '[a2]' -f flv rtmp://STREAMFINAL
[13:08] <Diogo> i want to to a fallback system if the first stream failed put the second stream without breaks the streamfinal?
[13:08] <Diogo> this is possible?
[14:01] <Giepher> hello which version ffmpeg support -filter_complex? I rum it on ffmpeg 0.7 display ffmpeg unrecognized option 'filter_complex'
[14:05] <spaam> 0.7 is old :S
[14:06] <cdhris> Is there any way to strip ONLY the unknown Codecs from Input Stream?
[14:07] <cdhris> Like this => Stream #0:2[0x102]: Unknown: none ([6][0][0][0] / 0x0006)"
[14:36] <cdhris> Hello again
[14:37] <cdhris> [mpegts @ 0x2c5c3e0] DTS 4130328530 < 4130663717 out of order [mpegts @ 0x2c5c3e0] DTS 4130751384 < 4131134930 out of order [mpegts @ 0x2c5c3e0] PES packet size mismatch [mpegts @ 0x2d79fa0] AAC bitstream not in ADTS format and extradata missing av_interleaved_write_frame(): Invalid data found when processing input
[14:37] <cdhris> what does this means :(
[15:37] <mustaf> Hi, on the download page of ffmpeg, when I choose Linux, there is section for static builds, "Linux Static Builds" both links lead to the same location.
[15:38] <mustaf> On the old download page there where two.
[15:41] <c_14> Yeah, that's been reported on the ml. Should be fixed soon".
[15:42] <mustaf> OK
[15:43] <c_14> This is the link if you're looking for it: http://johnvansickle.com/ffmpeg/
[15:44] <c_14> If you see anything else off with the new website don't hesitate to report it.
[15:48] <mustaf> OK, I'll, nice website by the way.
[16:15] <cdhris> c_14:  my friend
[16:15] <cdhris> do you know what this error means
[16:15] <cdhris> AAC bitstream not in ADTS format and extradata missing av_interleaved_write_frame(): Invalid data found when processing input
[16:17] <cdhris> http://pastebin.com/SaRJ85jH
[16:18] <cdhris> btw +bitexact etc same if i dont write them
[16:18] <cdhris> Even the most simpkle command like this: ffmpeg -y -hide_banner -i "http://localhost:5211" -c copy -map 0 -f mpegts /tmp/test.ts
[16:18] <cdhris> fails and shows me the error above
[16:23] <cdhris> if i transcoded it
[16:23] <cdhris> it works but i want to copy the audio not transcode it
[16:24] <c_14> Try adding -flags +global_header
[16:24] <c_14> If that doesn't work there's an open bug report that has a similar issue: https://trac.ffmpeg.org/ticket/279
[16:26] <cdhris> it doesnt work
[16:26] <cdhris> but that report is 3 years ago
[16:27] <cdhris> and i am using ffmpeg version 2.3 with all the latest libs etc
[16:27] <cdhris> tried with ffmpeg 2.2.4 same
[16:28] <c_14> Yeah, but it's never been marked as closed, maybe nobody got around to it.
[16:28] <c_14> Do you get the same issue if you mux to say matroska?
[16:29] <cdhris> what is the command
[16:29] <c_14> ffmpeg -i file -c copy -map 0 -f matroska /tmp/test.mkv
[16:30] <cdhris> i think it works with that
[16:30] <cdhris> so its bug with ts
[17:13] <Katharsis> hi guys
[17:14] <Katharsis> my client CS6 adobe doesn't recognize .mp4 file which was recorded on Debian 7.5
[17:14] <Katharsis> i tried a lot of combination, but it faild
[17:19] <c_14> Can you ffprobe the mp4?
[17:20] <Katharsis> c_14: i tried but it returns empty file
[17:21] <c_14> Is the file actually correct? What's the file size, can you play it?
[17:21] <Katharsis> I can play it with VLC on linux without any problems
[17:21] <Katharsis> it has ~5mb
[17:21] <Fjorgynn> i see
[17:22] <c_14> And ffprobe says empty file?
[17:22] <Katharsis> yeap
[17:22] <c_14> What does something like mediainfo say?
[17:22] <Katharsis> i used it: ffprobe -v quiet -print_format json -show_format -show_streams test.mp4 > test.json
[17:22] <c_14> What version of FFmpeg do you have?
[17:22] <Katharsis> the newest form debian 7.5 repository
[17:23] <c_14> That'll be libav, see #libav for support or download a static build/compile from source.
[17:24] <Katharsis> what's about with this libav?
[17:24] <Katharsis> should i paste you the results of ffmpeg -i test.mp4 ?
[17:24] <c_14> You can put in on a pastebin, but I'm pretty sure you're not using FFmpeg, but Libav.
[17:25] <Katharsis> http://pastebin.com/S4jLNKQ3
[17:26] <c_14> According to that partial output, the file looks fine. Can you play it with ffplay?
[17:27] <Katharsis> yes I can
[17:27] <c_14> Wait, so ffprobe file says empty file, but ffmpeg -i file gives you output?
[17:27] <Katharsis> yeap
[17:27] <Katharsis> ;)
[17:27] <c_14> Can you pastebin the full output of ffprobe file?
[17:28] <Katharsis> there is no output as you can see
[17:28] <Katharsis> i used for it this prompt:
[17:28] <Katharsis> ffprobe -v quiet -print_format json -show_format -show_streams test.mp4 > test.json
[17:28] <Katharsis> test.json is empty
[17:29] <c_14> ffrobe -v quiet won't output anything unless it throws errors
[17:30] <Katharsis> ffprobe test_25ftp.mp4 without args works
[17:30] <Katharsis> http://pastebin.com/8cScNyKr
[17:31] <c_14> try ffprobe -hide_banner -print_format json -show_format -show_streams test.mp4 &> test.json
[17:31] <c_14> that should give you your json.
[17:31] <Katharsis> yes, it has
[17:31] <c_14> get rid of the & and it should be valid json.
[17:31] <Katharsis> but the source has "Unrecognized option 'hide_banner' Failed to set value '-print_format' for option 'hide_banner'"
[17:33] <c_14> That's probably because you're using Libav. Pastebin the full output of `ffmpeg'
[17:34] <Katharsis> on this laptop i has diffrent linux so my result will not be the same
[17:34] <c_14> But coming back to your original issue. The file looks correct, and if ffplay can play it the problem seems to be on CS6es side.
[17:35] <c_14> Especially since both vlc and ffmpeg can play it, the party that can't is usually at fault.
[18:33] <jnvsor> Anyone on here knowledgeable about audio desync?
[18:47] <Insayne> hello, does ffmpeg support EPG data reading from RTSP streams ?
[19:34] <Fjorgynn> So I found the problem
[20:01] <Baluse> Hello
[20:01] <Baluse> Does anyone know where a-law codec is in FFmpeg library?
[20:05] <jnvsor> ffmpeg -codecs | grep alaw
[20:05] <jnvsor> I you have D you can decode, if you have E you can encode
[20:05] <jnvsor> If not your ffmpeg is compiled without the codec support
[20:08] <Baluse> i got the source
[20:09] <Baluse> I know that it is supported I am looking for sauce/source
[20:14] <c_14> Presumably in libavcodec/pcm.c
[20:14] <Baluse> yep found it:)
[20:14] <Baluse> it seems nice collection of codecs, there are a hell lot of them
[20:17] <c_14> Well, there are a whole lot of variations on pcm.
[20:21] <Baluse> yep adpcm is worse
[20:22] <Baluse> However opus etc are using external lib
[20:22] <Baluse> But what would be the difference if it was embeded ?
[20:23] <c_14> hmm?
[20:25] <Baluse> https://www.ffmpeg.org/general.html#Audio-Codecs
[20:26] <Baluse> here is says  E means that support is provided through an external library.
[20:30] <jnvsor> How do I record from pulseaudio without audio desync?
[20:30] <c_14> ah yes. The main difference between internal codecs and external libraries is that with the internal codecs you don't have to worry about licensing issues or making sure everything works across version updates and the users don't have to install the library to use the codec.
[21:43] <toso> Hi there. I'm trying to record an analogic video+audio from a pci card. I'm having trouble with the audio. I can use alsa and record the audio with arecord.. but the same does not apply with ffmpeg -f alsa -i ...
[21:43] <toso> This is a simillar problem from ml: https://lists.ffmpeg.org/pipermail/ffmpeg-user/2013-March/014217.html
[21:43] <Fjorgynn> i see
[21:45] <toso> sacarasc: sorry. http://pastebin.com/H3ejanbc
[21:49] <toso> with the ffmpeg output: http://pastebin.com/KkBJtFuZ -- thanks for any help
[21:50] <c_14> Try adding -map 1
[21:52] <toso> c_14: man... thank you so much.
[21:54] <c_14> np
[22:39] <ac_slater_> hey all, is it possible to have an h264 stream in an mpegts file/container/stream?
[22:40] <c_14> yes
[22:40] <sacarasc> That's how a lot of TV is transmitted.
[22:40] <ac_slater_> that's what I thought
[22:41] <ac_slater_> also, I just posted to the libav-user mailing list, but I might as well ask on here. I'd interesting in muxing a custom binary datastream into a mpegts file/container. Does ffmpeg support this?
[22:41] <ac_slater_> (not the application itself, but libav*)
[22:42] <ac_slater_> s/I'd/I'm
[22:42] <ac_slater_> s/interesting/interested
[22:42] <ac_slater_> (talking on the phone and typing isnt easy sometimes)
[22:47] <ac_slater_> my idea is to mux together Video+data into a mpegts context, and deliver it via rtp (and to a container)
[22:47] <ac_slater_> the mpegts files in the source tree are a little large and I havent figured out how they all connect yet. Any advice is appreciated.
[00:00] --- Tue Jul 22 2014


More information about the Ffmpeg-devel-irc mailing list