[Ffmpeg-devel-irc] ffmpeg.log.20130211

[burek]

[00:34] <rbrauer> I'm piping raw video (rgba) to h264 in mp4 via Java which produces a valid video if I use -t to set duration. Does anyone know how I would signal ffmpeg to close the file without an explicit duration? I tried sending 0x1a, but that didn't work. I'm doing this on Windows 7.
[01:00] <fenduru> rbrauer: I think 0x1A is only for text files. Are you closing the stream properly?
[01:01] <rbrauer> I tired closing both input and output streams on the ffmpeg process, but it still didn't complete the file correctly
[01:03] <rbrauer> after closing the video pipe it just sits there with no messages
[01:03] <fenduru> are you closing the stream without sending an EOF character?
[01:04] <FergusL> anybody with a mp3->ogg ffmpeg command at hand ?! I can't seem to find one working...
[01:06] <rbrauer> fenduru: what byte am I supposed to send for EOF?
[01:08] <fenduru> rbrauer: Honestly I'm not sure. But I would try just closing the stream without one.
[01:08] <rbrauer> I tried closing it from the Java side (ffOut.close()), wasn't working
[01:09] <rbrauer> I mean I get no repsonse from ffmpeg and the file isn't correct
[01:09] <fenduru> What type of stream are you dealing with on the Java side? (so I can lookup javadoc)
[01:10] <fenduru> FergusL: try using -acodec libvorbis and have your output file be NAME.ogg
[01:10] <FergusL> fenduru: I have vorbis but not libvorbis, weird
[01:11] <FergusL> oh wait
[01:11] <FergusL> it's in avconv now...
[01:11] <FergusL> Ubuntu here
[01:12] <rbrauer> fenduru: its a plain outputStream coming out of a Process wrapped with a bufferedOutputStream
[01:12] <rbrauer> I'm calling ffOut.flush() then ffOut.close()
[01:22] <fenduru> I'm out of ideas. I'm not familiar with piping/stdin on windows.
[01:22] <fenduru> Looking around it seems that there is no standard EOF in the spec
[01:22] <fenduru> Others suggest using -1
[01:23] <fenduru> I'm piping data from node.js into ffmpeg's stdin
[01:24] <fenduru> and using stream.end() to send the eof, which works great on Linux
[01:24] <fenduru> Perhaps try that, and if it works take a look at node.js' source for what it sends for EOF? (or if not node.js something else that has a similar .end() function)
[01:26] <fenduru> FergusL: did you figure out your problem?
[01:26] <FergusL> fenduru: yes I used avconv and there I could use -acodec libvorbis
[01:27] <FergusL> that's just crazy, avconv is a wrapper
[01:27] <FergusL> could just be a symlink
[01:27] <FergusL> to ffmpeg, but well... enough arguing over this I think
[01:33] <rbrauer> I figured out what was happening I think. Apparently when I closed the BufferedOutputStream it didn't close the actual one. I did ffProc.getOutputStream.close() after closing and flushing the buffered stream and it seemed to work.
[01:35] <rbrauer> well I thought that worked...maybe not
[01:37] <fatpony>  /join #avcon
[01:37] <fatpony> duh
[01:41] <rbrauer> yeah that worked my code is just really messy now and I keep putting stuff in the wrong place, thanks for the help fenduru
[01:43] <fenduru> Sure thing rbrauer.
[02:14] <fenduru> Any reason that ffmpeg would return successfully if it isn't segmenting the output as specified?
[02:43] <fenduru> Oh, the ffmpeg in the Ubuntu repo is very out of date
[02:44] <klaxa> the ffmpeg in the ubuntu repo isn't ffmpeg
[02:46] <fenduru> yeah, but even the one listed on the ffmpeg site is old
[02:50] <klaxa> huh?
[02:50] <klaxa> fenduru: elaborate
[02:52] <fenduru> klaxa: https://launchpad.net/~jon-severinsson/+archive/ffmpeg is listed as the repo to use for ffmpeg
[02:52] <fenduru> but is at version 0.10 instead of 1.x
[02:53] <klaxa> oh true that
[02:54] Action: fenduru is compiling from source now
[02:54] <klaxa> the only sane thing to do really :)
[02:55] <fenduru> yep :) I've been spoiled with arch's rolling release / package system
[02:56] <fenduru> although I should really just use one of the static builds that are online... this amazon EC2 server is quite slow
[03:09] <asher^> is there a way to have ffmpeg add metadata to an existing file?
[03:17] <klaxa> asher^: https://www.ffmpeg.org/ffmpeg.html#Main-options see: -metadata
[03:17] <klaxa> to not touch the rest of the streams, do: -map 0 -c copy
[03:18] <asher^> great, thanks, ill try that out
[03:18] <fenduru> klaxa, any idea how to avoid choppiness when segmenting your output?
[03:18] <klaxa> what do you mean with choppyness?
[03:19] <fenduru> if i recombine sequential output segments, sometimes there are clicks between the segments
[03:19] <fenduru> I assume there is some sort of padding going on here
[03:19] <klaxa> i.e. audio clicks?
[03:19] <fenduru> yep
[03:19] <asher^> klaxa, Unrecognized option 'c'. Any idea why this would happen?
[03:19] <klaxa> asher^: ffmpeg too old
[03:20] <klaxa> fenduru: that's probably because the last sample of the prepending segment, or the first sample of the following segment are not at 0 amplitude
[03:20] <klaxa> at least that's what i would think
[03:21] <klaxa> those things produce audio-clicks
[03:21] <klaxa> open the audio stream in a graphical audio editor and you should be able to see that
[03:21] <klaxa> audacity is pretty well suited for that
[03:22] <fenduru> I'll take a look. I'm segmenting for streaming on the web so I'd like to be able to reconstruct the segments losslessly
[03:22] <klaxa> asher^: install a newer version from your distributions repository, if they don't offer a newer one, compile from source or grab a static build
[03:22] <asher^> yeah ive got the latest from my distro, looks like ill have to compile. thanks
[03:25] <asher^> also, anyone know an alternative to ffmpeg-php? it causes 502 errors in some circumstances on my machine
[13:19] <veleno> hello everyone. is it possible to use ffmpeg/ffprobe to statically analyze a file and detect lost packets/lost frames ?
[13:25] <durandal_1707> veleno: i guess it depends on container/codecs
[13:25] <durandal_1707> which means more info is needed
[13:26] <durandal_1707> some containers simply are not designed that way
[13:26] <durandal_1707> also lost packets/lost frames is not same as corrupted packets/frames that fail to decode
[13:27] <veleno> durandal_1707: what would be the ideal container/codecs to detect those info? i have the liberty to chhoose, a this point.
[13:28] <veleno> lost packets/frame do have a mapping to corrupeted packets/frame, do they ?
[13:28] <durandal_1707> corrupted/incomplete frames are reported by lavc when decoding fails
[13:29] <durandal_1707> this depends on decoder, because some decoders may try to recreate data
[13:29] <veleno> (when I refer to lost/corrupted packets, it imply an input file received by a network streaming, not rtp-based)
[13:30] <durandal_1707> .. but than usually it is displayed in log what is happening ..
[13:30] <veleno> i see.
[13:31] <durandal_1707> veleno: than you need to look at that stream protocol/format
[13:32] <veleno> durandal_1707: that i know that is udp. the stream is sent over the network with somehting like:  ffmpeg -re -i video.mp4 -f mpegts rtp://127.0.0.1:60000 -loglevel debug -report -vstat
[13:32] <veleno> with udp instead of rtp in the streaming url
[13:35] <durandal_1707> then you need to find out if mpegts support some kind of packet drop detection
[13:36] <mf_> I've read it somewhere, but has forgotten what it stated, but what does max-rate really make for a difference when you run 2-pass encoding? Did it distribute the bitrate more evenly across all frames, or did it "move" the unused bits from a non/barely-moving sequence to the one that has more movement? How can I optimize this further? (i.e. it it possible to look ahead a lot longer for it to calculate more accurately?). I have a s
[13:36] <JEEB> in mpeg-ts there are continuity bits IIRC
[13:36] <JEEB> 0-15 was it?
[13:36] <JEEB> and if it doesn't go +1 then you've lost a packet
[13:37] <JEEB> (an mpeg-ts packet)
[13:37] <veleno> JEEB: that's useful indeed. i'm checking http://en.wikipedia.org/wiki/MPEG_transport_stream#Packet
[13:37] <JEEB> then you've got the 192 and 204 byte mpeg-ts packets that contain checksums
[13:37] <JEEB> I have no idea what libavformat supports and how
[13:41] <veleno> and i guess there's a way to see which are the stream transport protocols available
[13:41] <durandal_1707> there is some log under debug: "Continuity check failed for pid %d expected %d got %d\n"
[13:42] <veleno> durandal_1707: under the logs of ffmpeg or ffprobe ?
[13:42] <durandal_1707> veleno: it is in lavf, so anything that use lavf and care for its logs will get it
[13:43] <JEEB> what about the checksums for >188 byte mpeg-ts?
[13:47] <durandal_1707> i see some check_crc but it its result is not loged, i'm not mpegts expert and mpegts in lavf can be improved for sure
[13:50] <JEEB> yeah, lavf lacks support for PID switching, too
[13:53] <veleno> if a transport other than mpegts is preferrable for my goals, i'd rather switch to it. open for suggestions.
[13:57] <Mavrik> JEEB: PID switching?
[14:03] <JEEB> Mavrik, a PMT says that one or all of the streams will switch the PID they will be transferred over
[14:03] <Mavrik> ah, yes, that's not supported :)
[14:03] <JEEB> oh I know it's not supported
[14:04] <JEEB> which is why I use tvtest for Tokyo MX recordings :)
[14:04] <JEEB> (Tokyo MX switches between PIDs when they switch between SD and HD)
[14:04] <JEEB> NHK on the other hand keeps the video PID the same afaik
[14:06] <veleno> JEEP: in the mpegts packets i don't see CRC fields . where did you see the infos about checksum transported by byte 192 and 204 ?
[14:06] <veleno> JEEB: ^
[14:06] <Mavrik> veleno: there's an option of adding CRC fields to the 188 byte packets
[14:07] <Mavrik> if your MPEG-TS packets are 188B then you don't have those :)
[14:07] <JEEB> http://wiki.multimedia.cx/index.php?title=MPEG-2_Transport_Stream#Packet
[14:08] <JEEB> "Packets are most often 188 bytes in length, but some transport streams consist of 204-byte packets which end in 16 bytes of Reed-Solomon error correction data."
[15:32] <iive> isn't that part of the FEC?
[16:15] <skarm> Hi to all. Is it possible to stream /dev/video0 to local network, without using ffserver?
[16:24] <veleno> I guess there are CRC infos in the TS packets of the stream being analyzed http://codepad.org/GrvX6094 ..
[16:24] <veleno> i wonder if when using ffmpeg to stream the video it's possible to add those
[16:26] <durandal_1707> veleno: yes, modify mpegts muxer and send patch
[16:40] <veleno> durandal_1707: ok
[17:53] <espr3ss0> hilo
[17:54] <espr3ss0> need some help, is anyone here ?
[17:55] <Fjorgynn> duno
[17:55] <stqn> Cant help, but Im here. Hi.
[17:55] <Fjorgynn> where's the cake?
[17:56] <espr3ss0> lol
[17:57] <Fjorgynn> espr3ss0: shoot?
[17:57] <espr3ss0> anyone know how to set the timescale ?
[17:57] <Fjorgynn> what?
[17:57] <espr3ss0> mp4 works with QuickTime but to iPad complains
[17:57] <espr3ss0> iTunes even plays fine :-/
[17:59] <klaxa> espr3ss0: a bit more information about the problem would be great
[17:59] <klaxa> i.e. information about the produced file and ffmpeg output logs
[17:59] <espr3ss0> " [mov,mp4,m4a,3gp,3g2,mj2 @ 0x10180e800] stream 0, timescale not set "
[17:59] <klaxa> paste everything on pastebin
[18:00] <espr3ss0> not wanting to do that soz ;-p
[18:01] <espr3ss0> does ffmpeg have a timescale parameter ?
[18:21] <durandal_1707> espr3ss0: yes
[18:54] <iamtakin1iteasy> hi, which codec (and/or settings) do you recommend for recording desktop?
[18:55] <iamtakin1iteasy> from x11grab for instance
[18:55] <iamtakin1iteasy> so text won't get blurred
[18:55] <JEEBsv> ffvhuff is a fast and simple lossless encoder
[18:55] <iamtakin1iteasy> after a couple of seconds
[18:55] <iamtakin1iteasy> thank you
[18:56] <Guest94> in FLV transcoding using ffmpeg+libx264 i could see CPU usage is around 115%
[18:56] <Guest94> any suggestions?
[18:56] <Guest94> i am using single thread option
[18:57] <durandal_1707> use faster preset/more threads/faster hardware/ ....
[18:57] <Guest94> here is the commandline http://pastebin.com/nJpGpt26
[18:58] <Guest94> <durandal_1707> do you think using faster presets CPU usage will go down?
[18:59] <durandal_1707> try it
[19:00] <Guest94> let me give a shot
[19:02] <Guest94> i tried superfast/ultrafast and veryfast but nothing seems to take the cpu usage below 100%
[19:02] <Guest94> its always going >108%
[19:05] <durandal_1707> those preset and without extra options?
[19:06] <Guest94> here is the command i am using now ffmpeg -y -i  /root/fGFNmEOntFA.flv -vcodec libx264 -preset superfast -threads 1 /tmp/test.3g1_super.flv
[19:09] <klaxa> there is a superfast preset?
[19:09] <klaxa> wtf?
[19:11] <iamtakin1iteasy> also, can ffmpeg stream over the network?
[19:11] <iamtakin1iteasy> without medium of local file
[19:11] <klaxa> yes
[19:11] <iamtakin1iteasy> how it looks like?
[19:12] <Guest94> <klaxa> Read the article http://dev.gentoo.org/~beandog/x264_preset_reference.html
[19:12] <klaxa> write to udp or rtp, if you have an rtmp server, you can stream to that too
[19:12] <klaxa> oh cool didn't know that it had that many presets
[19:13] <Guest94> now if you can suggest something which is causing high CPU usage would be great
[19:13] <iamtakin1iteasy> hmm
[19:14] <klaxa> well too much computation needed for compression, also ffmpeg will use as much cpu as possible to decrease encoding time
[19:14] <Guest94> right but this will reduce the throughput of the system when we do multiple transcode
[19:15] <klaxa> you can increase the nice-value for ffmpeg
[19:15] <klaxa> however, it will still use as much cpu time as it can get
[19:16] <durandal_1707> and need
[19:17] <Guest94> any other better thought?
[19:17] <klaxa> no
[19:18] <klaxa> if you have multiple encodes at the same time obviously all encodes will be slower than if they are run sequentially
[19:18] <klaxa> if you increase nice values for selected ffmpeg instances, those should work slower than those instances with unchanged nice values
[19:18] <Guest94> all the encode in parallel, just like independent thread
[19:19] <klaxa> yes, that doesn't increase productivity of a system
[19:19] <iamtakin1iteasy> i am trying to stream x11grab over network, but apparently doing something wrong
[19:19] <iamtakin1iteasy> here is my commandline output: http://paste.eientei.org/show/39/
[19:19] <iamtakin1iteasy> any ideas where i got wrong?
[19:19] <iamtakin1iteasy> might it be wrong USE-flags i built ffmpeg with?
[19:20] <klaxa> iamtakin1iteasy: you have to specify a container-format
[19:20] <klaxa> for example: -f matroska
[19:20] <klaxa> i *think* that should work, i'm not sure though
[19:20] <klaxa> also, note that you will have rather high latency
[19:20] <klaxa> and make sure you have high network throughput
[19:21] <iamtakin1iteasy> i have added -f matroska to options list and now i am getting  udp://192.168.1.4:6000: No such file or directory
[19:21] <iamtakin1iteasy> 192.168.1.4 is the ip of current machine
[19:21] <iamtakin1iteasy> should it be some remote listening server?
[19:21] <klaxa> --disable-network
[19:21] <klaxa> heh
[19:22] <iamtakin1iteasy> ah
[19:22] <iamtakin1iteasy> i see
[19:22] <klaxa> that will be it
[19:22] <iamtakin1iteasy> thank you
[19:22] <klaxa> you could get it working with pipes and netcat though :V
[19:22] <iamtakin1iteasy> nah, i'll just re-emerge the ffmpeg with proper use-flag
[19:23] <iamtakin1iteasy> thanks for noticing
[19:23] <klaxa> i can't say how reliable udp streaming is, i don't think i ever got it to work really
[19:23] <iamtakin1iteasy> what about rtp servers?
[19:23] <iamtakin1iteasy> is there any opensourced of them?
[19:24] <iamtakin1iteasy> so i can setup it on my server with higher bandwitch
[19:24] <iamtakin1iteasy> and stream to other users
[19:24] <klaxa> dunno i haven't really found many good streaming solutions that are open-source
[19:24] <klaxa> i think live555 is rather popular though
[19:24] <Mavrik> red5?
[19:25] <klaxa> oh yeah that sounds like what we are looking for
[19:25] <klaxa> reading that it's implemented in java makes me think it must be bloated :V
[19:26] <iamtakin1iteasy> i am java coder. i am getting paid for work with bloatware >:3
[19:26] <iamtakin1iteasy> (shouldn't be so proud of it)
[19:27] <iamtakin1iteasy> but meh, there is no way i'll be scared by any kind of java monstrocity
[19:27] <iamtakin1iteasy> thank you
[19:57] <espr3ss0> ok ... had a look for a command to set the timescale but no luck :-(
[19:58] <espr3ss0> and everybody has gone :-/
[19:59] <durandal_1707> timescale what?
[19:59] <espr3ss0> an mp4 ? o_o
[20:00] <durandal_1707> can you explain in finer detail
[20:00] <espr3ss0> warning, " timescale not set  "
[20:00] <espr3ss0> i assume it's why i can't copy it to mi' iPad
[20:01] Action: espr3ss0 shrugs
[20:01] <espr3ss0> why need the whole info ?
[20:01] <espr3ss0> Stream #0:0(und): Video: h264 (High) (avc1 / 0x31637661), yuv420p, 1920x1080 [SAR 1:1 DAR 16:9], 2027 kb/s, 23.98 fps, 23.98 tbr, 96k tbn, 47.95 tbc
[20:02] <durandal_1707> i cant help if i cant reproduce - is this so hard to understand?
[20:02] <espr3ss0> oh, no :-)
[20:02] <espr3ss0> how can u re-produce when u don't have the file ?
[20:03] <durandal_1707> so instead of providing requested info you ask questions?
[20:03] <klaxa> espr3ss0: simple: you can't, but you can try to encode a file with the same output parameters
[20:04] <klaxa> that should produce similar encodes
[20:04] <espr3ss0> ahh, i c
[20:04] <klaxa> also, what durandal_1707 said, if you don't want to provide information, we can't help
[20:04] <klaxa> "I cannot get my car to start, I think the battery might be empty, also it's read, how do i fix it?"
[20:04] <espr3ss0> fair enuff ;-p
[20:04] <klaxa> *red
[20:05] <espr3ss0> i will post some info
[20:05] <durandal_1707> klaxa: call 911
[20:07] <espr3ss0> http://pastebin.com/tdJ5tquf
[20:11] <klaxa> looks like it's something with the input file rather than with ffmpeg
[20:13] <espr3ss0> what's wrong with the input file ?
[20:14] <klaxa> timescale is not set :V
[20:14] <klaxa> no idea how to fix it myself
[20:15] <klaxa> except maybe re-encode but you want to avoid that if possible
[20:15] <durandal_1707> time scale is set from demuxer
[20:15] <durandal_1707> or muxer
[20:15] <durandal_1707> so perhaps png stream messed up something
[20:16] <espr3ss0> tried from the original file, extract (copy) to an MKV then to MP4, which usually solves just about everything so far
[20:16] <espr3ss0> oh, that's without completely re-encoding
[20:16] <durandal_1707> perhaps it copy png "stream" too?
[20:17] <durandal_1707> because timescale is set for video
[20:17] <durandal_1707> and audio
[20:17] <durandal_1707> it is just not set for cover png
[20:19] <espr3ss0> stream 0.1 ?
[20:19] <espr3ss0> unless that's the "Artwork" (meta), then it shouldn't be there :-/
[20:20] <klaxa> heh now that i see it...
[20:20] <espr3ss0> listed in Subler is only Vid & Aud
[20:20] <klaxa> 1080p video but 64 kbps 22050hz aac
[20:20] <espr3ss0> ? o_O
[20:20] <klaxa> the audio track is extremly low quality
[20:21] <durandal_1707> espr3ss0: stream 0.2 is png
[20:21] <klaxa> it's funny that one shouldn't get copied even
[20:22] <espr3ss0> why not ?
[20:22] <klaxa> but try: -map 0:0 instead of -an
[20:23] <durandal_1707> anyway if you could make smaller sample (eg it happens when partially copied into mkv) that you could report it on trac and upload sample
[20:24] <Pitfall> Hi folks
[20:24] <Pitfall> I have a question about licensing
[20:25] <Pitfall> I want to use FFMPEG in my project which is closed source
[20:25] <Pitfall> Is that allowed?
[20:25] <JEEBsv> see the LGPL license
[20:26] <Pitfall> Is there a way to distinguish code between GPL and LGPL?
[20:26] <Pitfall> I heard that it had some GPL too in it
[20:26] <JEEBsv> yes, but if you do not do --enable-gpl ffmpeg will be LGPL
[20:26] <JEEBsv> it will only compile code that is LGPL
[20:26] <JEEBsv> (or LGPL-compatible)
[20:26] <Pitfall> OK
[20:27] <JEEBsv> you can then separately get a license from x264 LLC if you want to use x264 in a closed source app
[20:27] <JEEBsv> because it otherwise is GPL
[20:27] <Pitfall> We already have a x264 license
[20:29] <JEEBsv> nice
[20:29] <JEEBsv> anyways, just keep your ffmpeg configuration as LGPL, and you should be able to use it under the LGPL license :)
[20:29] <Pitfall> OK thanks
[20:29] <Pitfall> I do not want to be blamed for my actions :)
[20:31] <JEEBsv> just keep away from enabling GPL or nonfree
[20:31] <Pitfall> What if a library actually is GPL?
[20:31] <JEEBsv> then you can't use it
[20:31] <Pitfall> You cannot purchase a license for that?
[20:32] <JEEBsv> no unless the developers give such out
[20:32] <JEEBsv> you can always ask the developers
[20:32] <JEEBsv> also, a quick go-over for how to be "OK" with LGPL
[20:32] <durandal_1707> hmm, wasn't you can use gpl but using dynamic lib and not modifying code?
[20:33] <JEEBsv> that's /very/ iffy
[20:33] <JEEBsv> I'd rather keep away from that
[20:34] <JEEBsv> LGPL: if you do dynamic linking, just the exact source code of the LGPL component(s) used in the binary
[20:34] <JEEBsv> if you do static linking, the exact source code of the LGPL component(s) as well as the object files for the proprietary part
[20:34] <JEEBsv> so that the user can switch around the librarie(s) if he/she wants
[20:34] <JEEBsv> that is LGPL
[20:34] <JEEBsv> more or less
[20:35] <durandal_1707> Pitfall: read https://ffmpeg.org/legal.html
[20:36] <Pitfall> I did read that, but it is pretty vague
[20:36] <durandal_1707> than ask lawyer
[20:37] <espr3ss0> just checked with a different file ... muxing (mp4) only a video (x264) track and audio (aac) shows as fine and i didn't realise adding meta artwork adds an extra track. but then it shows the timescale problem
[20:38] <durandal_1707> it is definitely bug so report it anyway
[20:38] <Pitfall> I guess I will first check if I can use FFMPEG in my project compiled as LGPL
[21:08] <iamtakin1iteasy> hi, i hava a low-bandwitch netbook which screen i wish to stream over http protocol and i have a higher-bandwitch server which i want to do actual streaming to end-users. how do i best transfer the stream from netbook to server and restream for each user then?
[21:08] <iamtakin1iteasy> i have tried various rtsp servers, but they don't seem to be capabel of http streaming
[21:09] <iamtakin1iteasy> so you can't simply use mplayer on http stream they are [not] providing
[21:09] <iamtakin1iteasy> maybe i missed something?
[22:02] <Pinhole> How can I determine which extra libraries are actually being used (not compiled for)?  I have two boxes with the same stream and the same ffmpeg command line.  The resulting output files are very different in size.  I'm trying to figure out why.
[22:03] <Pinhole> Both are Fedora 17 with the same version of ffmpeg installed.  One is 32 bit the other is 64 bit.  The 64 bit seems right, the 32 bit is very bad.
[22:03] <JEEBsv> what is being used should be visible from the terminal output, pastebin the commands and terminal output onto a pastebin provider of your choice, and then link them here
[22:08] <Pinhole> Ok, I'll paste these and be back in about 40 minutes (they are in progress)
[23:06] <fenduru> is there a way to use -f segment to segment audio and pipe the segments out to stdout?
[23:08] <klaxa> use pipe: as an output file?
[23:08] <klaxa> that will write to stdout
[23:43] <fenduru> klaxa, I think I figured it out. I was overlooking the fact that aac needs a seekable output
[23:43] <klaxa> kk
[23:48] <humans63> !lista
[23:48] <humans63> ciao
[00:00] --- Tue Feb 12 2013