[Ffmpeg-devel-irc] ffmpeg.log.20160419

burek burek021 at gmail.com
Wed Apr 20 02:05:01 CEST 2016


[00:46:52 CEST] <Guest20457> guys, anywhere with examples i can learn more about codecs , bitrates and encoding using ffmpeg , its working but seems i have pixel issues, i need to understand the process better, any good documentations around?
[00:48:43 CEST] <llogan> Guest20457: you mean anything other than the official documentation and the wiki?
[00:49:04 CEST] <Guest20457> yep
[00:49:20 CEST] <Guest20457> i can see alot about commands , how to stream and restream
[00:51:30 CEST] <Guest20457> but for example, if i choose to stream h264 over udp , then recieve and encode and restream over rtmp or any other, what parameters should i worry about , from the source to destination
[00:52:33 CEST] <Guest20457> i read about the latency at he wiki, still reading, but i guess there are to many explanations wish i do not feel i understand pretty well
[00:55:27 CEST] <Guest20457> for example, i am using this at the capture pc side ( win 7 32) , ffmpeg -f dshow -i video="webcam" - video_size 1280x960 -framerate 18 -codec:v libx264 -pix_fmt yuv420p -qp 0 -preset fast -f mpegts udp://ip-of-local-server:1500
[00:56:07 CEST] <Guest20457> at the local server side i am playing with ffplay ( linux debian wheezy ) ffplay -i udp://ip-of-source:1500
[00:56:53 CEST] <Guest20457> but still having issues with pixel gettin malformed sometime, however i added -pix_fmt option for compatibility
[00:58:35 CEST] <Guest20457> now if i get the video with bad pixel and shapes, as shown on ffplay, should i adjust something from the source side , were the capture is, or i can do it at the server sid ( fix, encode, stream ) ?
[01:20:39 CEST] <DHE> as a rule udp:// is best suited for fixed LAN links with matched speeds, or for receiving only. ffmpeg will just spew UDP packets and the risk of packet loss is actually quite high
[01:21:06 CEST] <DHE> start by using TCP instead and see if that helps
[01:22:36 CEST] <Guest20457> thank you will do
[01:33:43 CEST] <Guest20457> is ffplay reliable for such test? or its better to record with ffmpeg and then play to check?
[01:43:52 CEST] <DHE> files on disk are always reliable, but usually a pain for streaming
[01:45:05 CEST] <Guest20457> i guesse the best way is to know the source file/capture device verry well to optimise for best streaming, there is no general rule for this
[01:47:28 CEST] <Guest20457> seems network issue, using same pc does work alot better
[03:28:40 CEST] <FlyingJester_> How would I specify an offset to begin an input file at in the output?
[03:29:08 CEST] <FlyingJester_> Not like -ss, but the opposite. Let's say I want an audio file to start 10 seconds in to the output file.
[03:30:51 CEST] <furq> FlyingJester_: -itsoffset
[03:31:15 CEST] <FlyingJester_> Ah, thanks, furq
[03:37:24 CEST] <FlyingJester_> I can use values like "1:23" as time offsets, right?
[03:41:40 CEST] <FlyingJester_> Let's say I have an audio file `file1.ogg' and a video file `file2.mp4', and I want the audio from file1 to start 4 minutes and 45 seconds into the output file. What's wrong with this: https://gist.github.com/FlyingJester/6b9996f7cffaf0b1024e438b5966e08a
[03:41:51 CEST] <FlyingJester_> I get the audio from file1.ogg at the very start of the output file.
[03:55:29 CEST] <furq> FlyingJester_: iirc you need to set a negative itsoffset on the video file
[03:55:38 CEST] <furq> or use a positive ss on the audio file
[03:56:02 CEST] <FlyingJester_> Hmm. Okay, I'll try that.
[03:56:18 CEST] <FlyingJester_> Although setting ss on the audio file sounds like not what I want.
[03:56:20 CEST] <furq> you can also replace -t 5:40 with -shortest
[03:56:30 CEST] <FlyingJester_> I want the audio to come in after 4:45 of video.
[03:58:50 CEST] <FlyingJester_> The -t in that should clip a few seconds off the audio before the video ends, which is what I want.
[03:59:21 CEST] <furq> nvm then
[03:59:44 CEST] <FlyingJester_> Setting a negative itsoffset just makes the video start in the middle.
[04:03:09 CEST] <furq> oh fun
[04:03:11 CEST] <furq> https://trac.ffmpeg.org/ticket/1349#comment:6
[04:03:51 CEST] <FlyingJester_> Well, at least it
[04:03:58 CEST] <FlyingJester_> Well, at least it's not just me.
[04:04:15 CEST] <FlyingJester_> I can just prepend some silcence to the audio track in audacity or something.
[04:04:34 CEST] <furq> https://ffmpeg.org/ffmpeg-filters.html#adelay
[04:05:28 CEST] <furq> i guess `-af adelay=285000|285000` should work
[05:37:38 CEST] <johnnny22-afk> How can I make ffmpeg render a frame better, I often get a gray'ish image
[05:38:59 CEST] <johnnny22-afk> on hevc content that is.
[05:51:00 CEST] <pfelt> evening all.  i'm looking at the movie filter and it looks like it only pulls video content from the specified file.  if i want both video and audio content do i have to use both movie and amovie?
[05:53:21 CEST] <pfelt> nm.  i read further down in the docs.  /me hangs his head in shame
[11:56:07 CEST] <morissette> Any thoughts on how to convert this file to mp4 or something usable? https://bpaste.net/show/e5b0672633e2
[12:02:21 CEST] <Snoo20664> ffmpeg -i raw_video.svr libx264 out.mp4
[12:04:03 CEST] <furq> morissette: svr?
[12:04:24 CEST] <morissette> furq, i have no idea; the only context I have is it is from a DVR
[12:04:28 CEST] <morissette> Snoo20664, trying that
[12:04:53 CEST] <morissette> Snoo20664, raw_video.svr: Invalid data found when processing input
[12:05:20 CEST] <Snoo20664> yeah didn't think it would work heh
[12:06:02 CEST] <Snoo20664> open raw_video.svr with mediainfo, click view -> text, and then copy the text and paste it to pastebin
[12:06:10 CEST] <Snoo20664> or bpaste or whatever
[12:06:26 CEST] <morissette> kk
[12:07:16 CEST] <morissette> https://bpaste.net/show/b9294a3bc52a
[12:07:24 CEST] <morissette> Thats a bloody useful tool
[12:09:50 CEST] <morissette> So thoughts?
[12:12:23 CEST] <morissette> Here is the file also if that helps for actual testing purposes: https://s3.amazonaws.com/observint/raw_video.svr
[12:13:53 CEST] <furq> $ file raw_video.svr
[12:13:53 CEST] <furq> raw_video.svr: data
[12:13:55 CEST] <furq> well that's useful
[12:16:44 CEST] <morissette> yeah thats what i started with a couple hours ago
[12:16:45 CEST] <morissette> haha
[12:16:59 CEST] <morissette> its raw bait car data
[12:17:23 CEST] <morissette> and i was given one night to integrate into our existing product and convert to something playable :)
[12:17:36 CEST] <morissette> b/c developers are "magic"
[12:17:54 CEST] <morissette> sad thing is when I downloaded from dropbox on my iphone
[12:17:59 CEST] <morissette> it works in vlc app
[12:18:01 CEST] <morissette> no issue
[12:18:22 CEST] <morissette> but vlc on linux says "missing header"
[12:22:06 CEST] <furq> huh
[12:22:15 CEST] <furq> vlc 2.0.5 plays the video, vlc 2.2 plays a silent audio track
[12:22:24 CEST] <furq> neither of them say what container it is
[12:23:17 CEST] <morissette> https://bpaste.net/show/caf6a76a568f
[12:23:21 CEST] <morissette> hmmm
[12:24:10 CEST] <Snoo20664> https://forum.videolan.org/viewtopic.php?t=23305
[12:24:23 CEST] <Snoo20664> people having the same issues a decade ago with no solution heh
[12:24:30 CEST] <Snoo20664> everything i tried failed :Z
[12:26:03 CEST] <morissette> :(
[12:26:24 CEST] <morissette> ...must not break down and implement a windows solution...
[12:28:17 CEST] <morissette> looks like i get to learn windows server 2012 in 2 hours and then implement some sort of microservice on it to transcode
[12:28:46 CEST] <morissette> but its all clicking so shouldn't be too hard
[12:29:03 CEST] <morissette> outside of figuring out how to make python run as a daemon on windows
[12:30:31 CEST] <morissette> hmmm
[12:30:32 CEST] <morissette> hopefuly
[12:30:35 CEST] <morissette> http://stackoverflow.com/questions/1798429/any-open-source-fast-h263-decoder-library-for-arm9-than-ffmpeg
[12:36:58 CEST] <morissette> from ffmpeg -codecs:  DEV.L. h263                 H.263 / H.263-1996, H.263+ / H.263-1998 / H.263 version 2
[12:37:33 CEST] <furq> morissette: http://s000.tinyupload.com/index.php?file_id=82058814924573083559
[12:38:10 CEST] <morissette> hmmm
[12:38:11 CEST] <morissette> .ts
[12:38:26 CEST] <furq> it's not actually .ts, i remuxed it with vlc 2.0.5
[12:38:38 CEST] <furq> vlc raw_video.svr --sout file/ts://raw_video.ts
[12:38:38 CEST] <morissette> [mharris at mori Downloads]$ file raw_video.ts
[12:38:38 CEST] <morissette> raw_video.ts: MPEG transport stream data
[12:38:44 CEST] <furq> well yeah the remuxed file is ts
[12:38:47 CEST] <furq> i meant the svr isn't
[12:39:48 CEST] <morissette> missing audio
[12:39:54 CEST] <morissette> but video is better than nothing
[12:40:08 CEST] <morissette> plus we keep the original for chain of evidence bs
[12:40:35 CEST] <furq> http://sprunge.us/gKFD
[12:40:44 CEST] <furq> i am reasonably certain that ffmpeg doesn't know how to demux it
[12:41:02 CEST] <morissette> fair enough well I really appreciate it
[12:41:16 CEST] <morissette> this will be a much more automate-able solution
[12:41:29 CEST] <morissette> vs hacked together windows / freeware / conversion
[12:42:11 CEST] <furq> idk you could probably get a raise on the back of being forced to install TURBOSOFT VIDEO CONVERTER ULTRA 7
[12:42:39 CEST] <furq> trial version, obviously
[12:42:44 CEST] <morissette> haha no raises for me
[12:43:13 CEST] <morissette> already got like an eighth of a million
[12:43:31 CEST] <morissette> maybe a raise if i can get the audio and timestamp working in the new few hours
[12:43:49 CEST] <morissette> i have this svr and a wmv file
[12:44:04 CEST] <morissette> the wmv plays with timestamp and audio
[12:44:26 CEST] <morissette> which i imagine they want both when viewing online when not a empty boring bait car vid
[12:44:50 CEST] <morissette> you think the timestamp is another vid stream?
[14:08:31 CEST] <andrey_utkin> Could please anybody look at this - concat-ing H.264 MP4 files encoded with "-x264opts stitchable=1" (option to ensure concat-ability) gives broken video: https://gist.github.com/andrey-utkin/1cc892661dc3216de7d4c6ff5f21ef9d
[14:45:30 CEST] <andrey_utkin> fflogger !pb andrey_utkin
[14:54:31 CEST] <otila> with latest x265 (r11428) and latest ffmpeg git, crf option does not have any effect, even though I get "x265 [info]: Rate Control / qCompress            : CRF-15.0 / 0.60"
[14:54:59 CEST] <otila> I get same file output with crf 15 and 25, except that x265 info line shows different crf
[16:05:03 CEST] <DHE> When converting surround sound to stereo, I find the volume to be rather low. Besides inserting a volume filter, is there something that can help? I'm using the swr_convert() API. Everything is working fine otherwise
[16:05:29 CEST] <DHE> google suggests messing with mixing coefficients, is that really the best option?
[16:06:14 CEST] <andrey_utkin> Could please anybody look at this - concat-ing H.264 MP4 files encoded with "-x264opts stitchable=1" (option to ensure concat-ability) gives broken video: https://gist.github.com/andrey-utkin/1cc892661dc3216de7d4c6ff5f21ef9d
[16:11:31 CEST] <durandal_1707> DHE: for headphones?
[16:12:11 CEST] <durandal_1707> there is dynaudnorm filter
[16:34:07 CEST] <DHE> durandal_1707: just stereo in general. I'm comparing to a commercial product that does similar and the audio is significantly louder
[16:34:55 CEST] <kepstin> well, if it's louder, there's a good chance it might clip :/
[16:35:04 CEST] <durandal_1707> yea it depends how your matrix is
[16:35:31 CEST] <kepstin> this is one of those 'there's no single definitive good way to do it' things.
[16:35:56 CEST] <DHE> I have no control over the commercial thing. ffmpeg I'm just using swr_* functions based on the examples and whatever defaults for my channel layout conversion request
[16:39:48 CEST] <kepstin> well, if you can figure out what the commercial thing is doing, you could certainly replicate it by using custom mixing coefficients
[16:50:01 CEST] <DHE> kepstin: it's a piece of hardware, I can't unfortunately
[16:50:25 CEST] <PacketPusha> When I am defining a file output mask for frames such as frame_%d.jpg how can I make it apply leading zeros?
[16:50:44 CEST] <PacketPusha> Right now I get frame_1.jpg ..... up to frame_10.jpg
[16:51:13 CEST] <PacketPusha> But then some software pulls frame_1.jpg, frame_10.jpg, frame_11.jpg in that order.
[16:52:15 CEST] <furq> frame_%03d.jpg
[16:52:21 CEST] <PacketPusha> Nice, thanks.
[16:52:24 CEST] <furq> man sprintf
[16:52:31 CEST] <PacketPusha> Ok, will do.
[17:26:32 CEST] <Prelude2004c> hello everyone good day
[17:27:23 CEST] <Prelude2004c> can someone help me with something... sooo strnage
[17:27:25 CEST] <Prelude2004c> http://pastebin.com/350NtDyc
[17:27:30 CEST] <Prelude2004c> the audio keeps getting out of sync
[17:27:44 CEST] <Prelude2004c> not right away but over time it does that.. can anyone assist me on how to try and keep the audio within sync
[17:45:11 CEST] <Ximon> Hi, what's the correct way of escaping special characters in filenames when using the tee pseudo-mixer?
[19:00:52 CEST] <maziar> what is this error :"library.mak:30: recipe for target 'libavcodec/x86/imdct36.o' failed"
[19:02:00 CEST] <BtbN> Not the error.
[19:02:21 CEST] <maziar> BtbN look at this http://pastebin.com/nMwsHYk1
[19:02:40 CEST] <BtbN> yes, that looks like the error.
[19:03:57 CEST] <maziar> i want to compile ffmpeg on debian jessie, and i follow this tutorial : https://trac.ffmpeg.org/wiki/CompilationGuide/Ubuntu ,,, but why i get this error , and how to fix it ?
[19:07:15 CEST] <maziar> i want to compile ffmpeg on debian jessie, and i follow this tutorial : https://trac.ffmpeg.org/wiki/CompilationGuide/Ubuntu ,,, but why i get this error: http://pastebin.com/nMwsHYk1 , and how to fix it ?
[19:10:24 CEST] <llogan> maziar: never seen that one before. how do i duplicate the issue? did you modify the steps in the guide at all?
[19:10:39 CEST] <maziar> yes
[19:11:05 CEST] <maziar> llogan can you help me
[19:11:18 CEST] <llogan> maybe
[19:11:31 CEST] <maziar> llogan please kindly help me
[19:11:37 CEST] <llogan> i'm trying to
[19:52:23 CEST] <maziar> i need help about this issue
[19:53:05 CEST] <llogan> maziar: i was wating for you to answer my questions
[19:53:34 CEST] <maziar> llogan sorry, i don't see your question,
[19:54:25 CEST] <llogan> "<llogan>... how do i duplicate the issue? did you modify the steps in the guide at all?"
[19:55:38 CEST] <maziar> i did every step on that guid, but a question what you mean by :"how do i duplicate the issue? " i don't understand it
[19:59:56 CEST] <llogan> what do i need to do to get the same error?
[20:09:19 CEST] <maziar> llogan compile it, step by step on debian jessie
[20:14:35 CEST] <maziar> llogan i will try it again
[20:36:13 CEST] <blackfreighter> Hi, people. Newby here, trying to install ffmpeg in Fedora 23. RPM Repo seems empty, already ran dnf install outputs "Cannot download ffmpeg-2.8.5-1.fc23.x86_64.rpm: All mirrors were tried". Downloaded the tarball instead, but I'm not quite sure how I should go about the installation.
[20:36:52 CEST] <llogan> blackfreighter: just download a compiled binary here http://johnvansickle.com/ffmpeg/
[20:38:38 CEST] <blackfreighter> llogan: thank you! Will this install ffmpeg or just let me use it's capabilities?
[20:38:49 CEST] <blackfreighter> its*
[20:39:13 CEST] <llogan> you don't need to install it. just run it like any other executable. or point your scripts to it
[20:48:27 CEST] <llogan> maziar: i can't duplicate the issue
[20:49:16 CEST] <llogan> compiles fine, other than the x265 cmake command no longer finding a compiled yasm
[20:49:17 CEST] <maziar> llogan let me finish it
[20:49:58 CEST] <llogan> this is on Debian 8.4.0 x86_64
[21:17:19 CEST] <mike_> evening all
[21:19:26 CEST] <Guest66826> i have been looking for a way to stream raw data from pc 1  to pc 2 , where all the work will be done, via udp, i need to do this to minimise cpu utilisation to the minimum , how can i do that
[21:23:00 CEST] <Guest66826> iv been using ffmpeg -f dshow -i INPUT -codec:v  libx264  -f mpegts udp://local_ip_pc_2   , its working but i need something like this  ffmpeg -f dshow -i INPUT -c:v copy udp://local_ip_pc_2 ,   so the question is how can i feed this to another ffmpeg process, decode then encode to desired stream as final output?
[21:25:46 CEST] <Guest66826> should i worry about the raw format of the device, to decode it properly from pc2? wish means if i have different devices i should use different decoders? like for this one its a webcam and outputs rawvideo bgr24 , and ffmpeg says encoder Lavc57.34.100
[21:26:33 CEST] <maziar> llogan can i change "PATH=" PATH="/bin:$PATH" cmake -G "Unix Makefiles" -DCMAKE_INSTALL_PREFIX="$HOME/ffmpeg_build" -DENABLE_SHARED:bool=off ../../source " insist to :"PATH="$HOME/bin:$PATH" cmake -G "Unix Makefiles" -DCMAKE_INSTALL_PREFIX="$HOME/ffmpeg_build" -DENABLE_SHARED:bool=off ../../source"
[21:30:33 CEST] <maziar> llogan ping
[21:30:37 CEST] <relaxed> maziar: what are you trying to do?
[21:30:51 CEST] <relaxed> different bin and install path?
[21:31:21 CEST] <maziar> relaxed i want to install ffmpeg. path make it : /root/bin/  i want /bin
[21:33:04 CEST] <maziar> relaxed is it posible ?
[21:33:34 CEST] <relaxed> "cmake -LA ../../source" will show you all the options
[21:33:57 CEST] <relaxed> cmake -LA ../../source | grep INSTALL
[21:34:48 CEST] <llogan> maziar: install yasm from repo, then you can remove "PATH=" PATH="/bin:$PATH". that was intended for users who want to compile the newer yasm instead
[21:35:30 CEST] <relaxed> maziar: do you see the right option now?
[21:39:57 CEST] <maziar> llogan ok, let me tesrt
[21:42:36 CEST] <Guest66826> llogan ping
[21:44:13 CEST] <llogan> Guest66826: huh?
[21:55:41 CEST] <Guest66826> can you help with feeding rawvideo to ffmpeg?
[22:19:06 CEST] <maziar> llogan ping
[22:19:16 CEST] <maziar> llogan this is last output
[22:19:20 CEST] <maziar> llogan this is last output : LD	ffserver_g
[22:19:21 CEST] <maziar> CP	ffserver
[22:19:23 CEST] <maziar> STRIP	ffserver
[22:22:35 CEST] <pfelt> could someone explain to me what pts=PTS-STARTPTS does?  i'm assuming that PTS is encoded in the video stream (in my case an mpegts) and that STARTPTS is the current wall time on the machine?
[22:22:58 CEST] <pfelt> Guest11324: what is it you need?
[22:24:39 CEST] <kepstin> pfelt: in the setpts filter, 'STARTPTS' is the pts value of the first frame that the filter sees
[22:25:07 CEST] <kepstin> so that setpts=pts=PTS-STARTPTS sets the first frame's pts to be '0', and then keeps the relative frame timing for later frames.
[22:25:36 CEST] <kepstin> so, e.g. frames with pts 15, 17, 20 turn into 0, 2, 5
[22:28:19 CEST] <pfelt> are those actual timestamps, or just relative to each other?  (eg: not time 2, or 5, but really display frame pts=3 right after frame pts=2 and before frame pts=4)
[22:28:46 CEST] <Mavrik> PTS (presentation timestamps) are actually timestamps in the stream timebase :)
[22:28:50 CEST] <Mavrik> E.g. they're not in seconds
[22:28:55 CEST] <Mavrik> And they don't necessarily start at 0
[22:29:36 CEST] <Guest66826> hi pfelt
[22:30:43 CEST] <pfelt> Mavrik: but those timestamps don't really have anything to do with "current time".  how could they given that my mpegts was recorded hours ago and playback now works.  (am i following that right?)
[22:30:44 CEST] <Guest66826> i want to stream from pc 1 over udp without any encoding, max bitrate , to pc 2 , and there do the encoding
[22:31:33 CEST] <Mavrik> pfelt, no they don't have anything to do with "current time" - essentially they tell the player at what time to display a frame (or audio sample)
[22:31:48 CEST] <pfelt> Guest11324: "ffmpeg -i INPUT -vcodec rawvideo -f mpegts udp://ip:port"  work for you?
[22:31:48 CEST] <Mavrik> STARTPTS isn't wall time
[22:31:52 CEST] <maziar> what is this error for : Unknown encoder 'libfaac'
[22:31:53 CEST] <maziar> "
[22:31:53 CEST] <Mavrik> it's the PTS of the first frame (start)
[22:32:00 CEST] <Mavrik> in most videos it should be 0
[22:32:07 CEST] <Mavrik> But it's not always if they're streams or cut somewhere.
[22:32:23 CEST] <pfelt> Mavrik: ok.  yeah.
[22:32:28 CEST] <pfelt> i think i'm following you now
[22:32:33 CEST] <pfelt> (you and kepstin)
[22:33:02 CEST] <Mavrik> So in short, they're just relative to each other
[22:33:14 CEST] <pfelt> and the setpts filter does its thing on *every* frame basically
[22:33:14 CEST] <Mavrik> er.
[22:33:18 CEST] <Mavrik> That came out wrong I guess.
[22:33:47 CEST] <Mavrik> pfelt, yeah, what your "setpts=pts=PTS-STARTPTS" does it just makes sure that first frame of your video will have PTS of 0
[22:33:53 CEST] <Mavrik> and all others will then follow
[22:34:03 CEST] <Mavrik> instead of starting at some other number
[22:34:19 CEST] <Guest66826> pfelt, thats what i use at pc 1 , but how do i recieve as input from pc 2, shouldint it be ffmpeg -u udp://ip:port ...
[22:34:33 CEST] <Guest66826> sorry -i instead of -u
[22:34:41 CEST] <pfelt> Guest11324: yeah.  that should be right
[22:34:59 CEST] <pfelt> don't forget, firewalls can get in the way of stuff like this
[22:35:02 CEST] <Guest66826> without any decoding as its a raw video right?
[22:35:15 CEST] <Guest66826> its a dummy switch , local network
[22:35:37 CEST] <pfelt> Guest66826: machines can run iptables or windows firewall
[22:35:53 CEST] <maziar> pfelt do you know this error : Unknown encoder 'libfaac' ",,,, and how to fix ?
[22:35:58 CEST] <pfelt> Guest66826: but yeah.  you should be able to read it in fine with a -i
[22:36:14 CEST] <pfelt> maziar: sounds like you don't have libfaac enabled in your ffmpeg compile
[22:36:35 CEST] <maziar> pfelt, how should i do it ?
[22:36:42 CEST] <Mavrik> Don't.
[22:36:47 CEST] <Mavrik> Use one of the non-shitty AAC encoders.
[22:36:54 CEST] <pfelt> hehe
[22:36:57 CEST] <Mavrik> ffmpeg -codecs, see which one you have.
[22:37:36 CEST] <Guest66826> yes all disabled, however its saying failed to open or configure filtergraph, and ffprobe unsupported codec ...
[22:37:47 CEST] <pfelt> one last quetion relating to setpts.  it runs the formula PTS-STARTPTS on the frame's pts right? (eg: it doesn't just set up a counter then increment the counter after the first frame)
[22:38:15 CEST] <maziar> pfelt please kindly help me
[22:38:20 CEST] <Mavrik> pfelt, yes.
[22:38:24 CEST] <pfelt> Guest66826: you'd need to post the command and output somewhere in a pastebin for someone to help
[22:38:41 CEST] <Guest66826> but if i encode using libx264 it works, wish means no firewall/network issue, thats why i was thinking about mabe it needs to be decoded first before playing
[22:38:59 CEST] <pfelt> Mavrik: ah.  kk.  that's my problem then.  (i'm making modifications to the movie filter and it's erroring.  i knew it was something i was missing in my head)
[22:39:55 CEST] <pfelt> maziar: you'll need to either read up on rebuilding ffmpeg for your platform, or do as Mavrik suggested and change your command to use a different codec
[22:40:44 CEST] <maziar> pfelt what is the best encoder for audio?
[22:41:04 CEST] <pfelt> that's a pretty subjective question.  it depends on what you are doing and what you need
[22:41:22 CEST] <pfelt> currently i'm using "-an" in all my encoding and it works perfectly
[22:41:49 CEST] <pfelt> provides 100% lossy compression.     but then again, i *need* to strip audio out
[22:43:15 CEST] <maziar> pfelt i have some movie, and i want best quality, for it which is better ?
[22:44:00 CEST] <Guest66826> pfelt, command from pc1 "pastbin.com/raw/Ee0buKh6"    command at pc 2 "pastbin.com/raw/Gkwp7b4N"
[22:46:04 CEST] <pfelt> maziar: sorry, friend.  i can't answer that question for you.  you'll need to look at what codecs you have in your toolbox and research which will be the best for your application
[22:46:58 CEST] <maziar> pfelt actually i want best command for convert MP4 to m3u8, do you have suggestion ?
[22:47:20 CEST] <pfelt> Guest66826: hmm.  not sure on that myself.  does it help if you change the -f to matroska instead of mpegts?
[22:47:28 CEST] <pfelt> maziar: i don't  :(
[22:49:08 CEST] <Guest66826> pfelt, its giving some erros, il search and try to fix them using matroska and see what happens, thank you
[22:50:48 CEST] <pfelt> :)  sorry i couldn't me more help
[22:51:25 CEST] <maziar> pfelt i also get this error : Unknown encoder 'libvo_aacenc'
[22:51:48 CEST] <pfelt> i know there are some issues with mpegts in ffmpeg not being quite standard unless you specify a bunch of other parameters (enet frames sizes aren't 1316).  not sure if that's your problem tho
[22:52:06 CEST] <pfelt> maziar: i believe it was suggested to use -codecs to see what you have available to you
[22:53:21 CEST] <Guest66826> pfelt, thats a good start  :)
[22:53:57 CEST] <maziar> pfelt , would you please tell me by sample command ?
[22:54:20 CEST] <pfelt> ffmpeg -codecs
[22:54:28 CEST] <pfelt> maziar: ^^
[22:56:38 CEST] <maziar> pfelt http://pastebin.com/nwp5jWgh
[23:01:25 CEST] <kepstin> maziar: now you scroll down to the line that says 'aac' and look over to the list of encoders on the right.
[23:01:48 CEST] <kepstin> in this case, you have libfdk_aac, which is probably a good choice.
[23:02:14 CEST] <maziar> kepstin   libavutil      55. 22.100 / 55. 22.100
[23:02:15 CEST] <maziar>   libavcodec     57. 35.100 / 57. 35.100
[23:02:16 CEST] <maziar>   libavformat    57. 34.103 / 57. 34.103
[23:02:17 CEST] <maziar>   libavdevice    57.  0.101 / 57.  0.101
[23:02:18 CEST] <maziar>   libavfilter     6. 44.100 /  6. 44.100
[23:02:19 CEST] <maziar>   libswscale      4.  1.100 /  4.  1.100
[23:02:20 CEST] <maziar>   libswresample   2.  0.101 /  2.  0.101
[23:02:21 CEST] <maziar>   libpostproc    54.  0.100 / 54.  0.100
[23:02:55 CEST] <maziar> kepstin i just want to create m3u8 from a MP4, how should i do, i'm really tired to do it
[23:04:38 CEST] <kepstin> you mean you want to make segments for an HLS stream? m3u8 is just a playlist :)
[23:05:31 CEST] <Mavrik> *facepalm*
[23:07:17 CEST] <kepstin> maziar: any idea about what you want to use for segment lengths, bitrate, etc?
[23:07:42 CEST] <maziar> segment is 10 second and bitrate is 480kbps
[23:08:39 CEST] <kepstin> assuming you've already set up the rest of the command (it would be helpful if you show us your existing command line), it should be as simple as adding '-c:a copy' (if the input has aac audio), or '-c:a libfdk_aac' to fix your error.
[23:29:15 CEST] <maziar> kepstin what is this error : [NULL @ 0x4b8c980] Unable to find a suitable output format for '32'
[23:29:15 CEST] <maziar> 32: Invalid argument
[23:29:51 CEST] <maziar> kepstin what is this error : [NULL @ 0x4b8c980] Unable to find a suitable output format for '32'  32: Invalid argument
[23:30:10 CEST] <kepstin> that error almost certaily means that you made a typo on the command line.
[23:30:29 CEST] <kepstin> so please follow the instructions that the helpful bot has given you.
[23:35:19 CEST] <maziar> i want to create m3u8 from mp4 for test ffpmeg a sample , a test, how can i do it
[23:36:37 CEST] <llogan> repeating what you want to do while ignoring requests for aditional information is the best way to get help. also, make sure to provide only the errors with no context.
[23:46:33 CEST] <pfelt> ok.   another time question.  what is the timebase?  (ref: https://ffmpeg.org/ffmpeg-filters.html#setpts_002c-asetpts)
[23:49:50 CEST] <kepstin> pfelt: timebase is a fraction that you multiply the pts by to get the time in seconds.
[23:50:22 CEST] <kepstin> so, e.g. ntsc video might have a timebase of 1001/30000, then frame pts is 0, 1, 2, 3
[23:52:27 CEST] <kepstin> typically the timebase will be the inverse of the framerate, so the pts values are nice, but that's not always the case.
[23:55:32 CEST] <kepstin> I think mkv files use a timebase of 1/1000 and the pts values are just milliseconds.
[00:00:00 CEST] --- Wed Apr 20 2016


More information about the Ffmpeg-devel-irc mailing list