[Ffmpeg-devel-irc] ffmpeg.log.20150311

[burek]

[05:34:02 CET] <soahccc> Hey guys. I try to transcode a HLS stream to rtmp and push it to an rtmp server. Curiously the rtmp server understands what ffmpeg does there but even though I can watch the HLS stream and the remote RTMP stream, I cannot open the local RTMP stream. VLC complains about "undefined codec". What could I possibly screwed up there with the codec? This is my ffmpeg command: https://gist.github.com/2called-chaos/ba6c901c803a8e66daf0
[05:35:29 CET] <soahccc> I mean vcodec is copy and I can open the original footage. The output must be valid as well since the remote RTMP can process it.
[05:39:39 CET] <soahccc> And other flash based players also work. I must have done something VLC unfriendly :/
[07:03:57 CET] <ajsd9900> Is anyone available to answer a simple ffmpeg question?
[07:05:04 CET] <relaxed> ask it
[07:06:43 CET] <ajsd9900> I'm using 'ffmpeg -ss 00:09:35 -i in.mp4 -to 0:09:49 -c copy out.mp4' and the out file lasts until the end of the video rather than at that timestamp. Is there any way around this besides moving -ss after -i?
[07:07:37 CET] <relaxed> it works with -ss after the -i ?
[07:14:37 CET] <ajsd9900> yeah
[07:15:14 CET] <relaxed> see if -to before -i works
[07:15:50 CET] <ajsd9900> with the -ss before, the -to is the time after it. e.i. 9 minutes 49 seconds added to 9:35
[07:15:52 CET] <ajsd9900> it doesnt
[07:17:46 CET] <relaxed> -t works before -i, so I would consider that a bug
[07:20:00 CET] <ajsd9900> "Option to (record or transcode stop time) cannot be applied to input file <->. -- you are trying to apply an input option to an output file or vice versa. Move this option before the file it belongs to." "Error parsing  options for input file  <->" "Error opening input files: Invalid argument"
[07:21:38 CET] <ajsd9900> are you sure it works? could you try it out?
[07:24:39 CET] <relaxed> -t works but -to does not
[07:25:01 CET] <relaxed> which, imo, makes no sense
[07:26:13 CET] <ajsd9900> it completed successfully, but it still goes until the end of the video
[07:26:32 CET] <relaxed> which version?
[07:27:26 CET] <ajsd9900> N-70606-g2e5605f (latest windows static)
[07:28:31 CET] <relaxed> -t doesn't work?
[07:28:49 CET] <ajsd9900> -t works, but it adds the time instead of ending at that time
[07:29:07 CET] <relaxed> I understand the difference
[07:30:48 CET] <ajsd9900> would you still say its a bug?
[07:30:55 CET] <relaxed> hmm, -ss before the -i and -to after it works here
[07:31:35 CET] <relaxed> using latest git here
[07:40:06 CET] <ajsd9900> http://pastebin.com/iW2mrum5
[07:54:18 CET] <ajsd9900> relaxed http://pastebin.com/iW2mrum5
[07:56:10 CET] <relaxed> I tested with matroska. Do you have a mkv video to see if it works?
[07:56:56 CET] <ajsd9900> nope
[07:57:42 CET] <relaxed> if it works with -ss after the -i then just use that
[08:00:08 CET] <ajsd9900> i found that having -ss before -i solves this problem http://stackoverflow.com/questions/23672179/why-ffmpeg-split-avi-movie-freezes-when-played
[08:00:28 CET] <ajsd9900> do you know of another way to solve it?
[08:01:35 CET] <ajsd9900> (that is more simple than the answer linked)
[08:01:52 CET] <relaxed> adding -copyinkf might fix it
[08:03:47 CET] <ajsd9900> which does? I can't find any documentation
[08:05:02 CET] <ajsd9900> seems to have fixed it though
[08:07:28 CET] <relaxed> it's covered in the man page: -copyinkf  When doing stream copy, copy also non-key frames found at the beginning.
[08:08:22 CET] <relaxed> -ss before the -i seeks to keyframes, after the -i it decodes to a certain point
[08:14:09 CET] <ajsd9900> i found that the first second is grey and distorted in VLC, but it still plays. however windows media player still freezes. i got a similar result when -ss was before -i but it wasnt as bad
[08:14:58 CET] <ajsd9900> that's odd. i uploaded it to see what it would look like on youtube and the upload failed
[08:15:21 CET] <sagar_> Hi everyone,
[08:15:34 CET] <sagar_> I am trying to concat two video
[08:16:44 CET] <sagar_> but failing due to some SAR issue , I don't want to land up in such situation again so for that before concatinating video how would I check that two videos are compatible or not
[08:17:05 CET] <sagar_> or should I encode them first and concat them ?
[08:18:46 CET] <ajsd9900> relaxed would you agree with me that i just do the extra math with -ss before -i to avoid these issues?
[08:18:56 CET] <ajsd9900> or is this something that needs to be fixed?
[08:19:11 CET] <ajsd9900> on ffmpeg's part
[08:59:18 CET] <relaxed> ajsd9900: I would use -ss after the -i along with -copyinkf because it provides more accurate seeking.
[09:02:50 CET] <pzich> or both
[09:14:56 CET] <relaxed> pzich: huh?
[09:20:27 CET] <pzich> e.g. ffmpeg -ss 00:15:30 -i file.mp4 -ss 00:00:30 ...
[09:21:04 CET] <relaxed> two -ss ?
[09:21:04 CET] <pzich> https://trac.ffmpeg.org/wiki/Seeking#Combinedseeking
[09:21:24 CET] <relaxed> who said anything about that?
[09:22:31 CET] <pzich> hmm, I guess I should time these at some point, I guess the fast isn't relevant anymore?
[09:42:43 CET] <kryo_> when trying to read an FLV i get this: [concat @ 0x196d1c0] Line 1: unknown keyword 'FLV??'
[09:46:09 CET] <relaxed> command?
[09:50:37 CET] <kryo_> ffmpeg -re -f concat -i Part01.flv -vcodec copy -acodec libmp3lame -ar 44100 -f flv rtmp://
[09:50:41 CET] <kryo_> relaxed ^
[09:53:09 CET] <relaxed> concat one input?
[09:53:29 CET] <kryo_> the actual command has * instead of Part01.flv
[09:53:53 CET] <kryo_> here is the header from Part01.flv if that's any help... :P http://puu.sh/gvBQO/28b69edc0b.png
[09:57:10 CET] <kryo_> i'm assuming it's a problem with the file?
[10:02:59 CET] <kryo_> huh... i was under the impression that concat would accept a list of files
[10:03:47 CET] <kryo_> guess it needs to be a text file
[14:19:00 CET] <amol> hello
[14:20:22 CET] <amol> can anyone help me to find out video of DTMB format. I want to download a video file which has subtitles in it.
[14:21:18 CET] <durandal_1707> what is DTMB?
[14:22:25 CET] <amol> it is a Digital Terrestrial Multimedia Broadcast used in china.
[14:23:12 CET] <amol> in short it's a type of television broadcast.
[14:35:50 CET] <Bobbyt> hello, can you please tell me if I can split a video into two, one containing odd frames and the other the even ones?
[14:39:36 CET] <saste> Bobbyt, yes, use select
[14:39:50 CET] <saste> i mean the select filter
[14:43:37 CET] <Bobbyt> can you explain me what select="not(mod(n-1\,2))" does?
[14:43:56 CET] <Mavrik> select - select frames
[14:44:06 CET] <Bobbyt> I mean who's n?
[14:44:06 CET] <Mavrik> "n" is frame number
[14:44:42 CET] <Mavrik> mod(n-1,2) converts 1,2,3,4,5,6,7,8,9 into 0,1,0,1,0,1,0...
[14:44:59 CET] <Mavrik> so basically "select every second frame"
[14:45:17 CET] <Bobbyt> ok
[14:45:45 CET] <Bobbyt> and afterwards how can I create a video file using only the 0 - type frames
[14:46:21 CET] <Mavrik> just remove the not
[14:48:30 CET] <Bobbyt> I think I am starting to get it, thanks :)
[14:59:10 CET] <saste> Bobbyt, yo can create two outputs with select
[14:59:24 CET] <saste> so you can store both odd and even frames at the same time
[15:45:57 CET] <schorsch> Hello everyone! is there any option for slow-mo ?
[16:29:38 CET] <vilain> Hi guys.
[16:30:35 CET] <vilain> Do i need to encode a video stream to apply the cgop flag?
[21:14:43 CET] <newbie999> Does ffmpeg 2.5 ABI compatable with 2.6 ?
[21:18:35 CET] <pgmer6809> Does ffmpeg accept donations thru paypal? If not why not?
[21:19:10 CET] <pgmer6809> Q2. Does ffmpeg support Sony Atrac family of codecs? convert Atrac to mp3?
[21:26:07 CET] <kepstin-laptop> newbie999: it looks like programs compiled against 2.5 should continue to work with 2.6 (the opposite is not necessarily true)
[21:28:20 CET] <Jonas__> okay, my issue: streaming my encoding to an rtmp url instead of a file causes significant framedrop
[21:28:53 CET] <Jonas__> saving to file and then streaming *that* in realtime(-re) to the same rtmp url causes no framedrop
[21:29:30 CET] <Jonas__> possible solution, encode to pipe:1 then use another ffmpeg to grab input from pipe:1 and just copy it all to the rtmp url?
[22:30:44 CET] <techtopia> hey guys
[22:30:55 CET] <techtopia> i dropped my keyboard in the middle of an encode
[22:31:10 CET] <techtopia> and now ffmpeg is going crazy showing hex super fast
[22:31:18 CET] <techtopia> instead of the encoding lines
[22:31:25 CET] <techtopia> is there anyway to fix it
[22:31:31 CET] <techtopia> or abort and start over?
[22:31:59 CET] <c_14> h
[22:32:13 CET] Last message repeated 1 time(s).
[22:32:13 CET] <c_14> Wait, no that was the wrong one.
[22:32:15 CET] <c_14> I think it's x?
[22:32:17 CET] <c_14> let me check
[22:32:35 CET] <c_14> But no, don't abort the encoding.
[22:32:51 CET] <c_14> Ah no, it was 'h'
[22:33:03 CET] <c_14> Press 'h' a couple of times to toggle through the output levels.
[22:33:25 CET] <techtopia> doesn't seem to be working
[22:33:36 CET] <techtopia> but as long as it's encoding still i can just wait for it to finnish
[22:33:44 CET] <c_14> It's still encoding.
[22:34:11 CET] <c_14> Pressing 'h' does nothing?
[22:34:32 CET] <techtopia> nope just red text showing hex code
[22:34:43 CET] <c_14> What was it showing before?
[22:35:03 CET] <c_14> Ah, it's currently showing the red hex?
[22:35:09 CET] <techtopia> like frame count, fps encoding rate and encoded time
[22:35:11 CET] <techtopia> yes
[22:35:20 CET] <c_14> Maybe you accidentally disabled stdin reading somehow.
[22:39:51 CET] <kaniu> common.mak seems to contain a bug:
[22:40:07 CET] <kaniu> common.mak:160: *** missing separator.  Stop.
[22:41:24 CET] <kaniu> unfortunately, I am not familiar enough with makefiles to even understand what the error message could mean
[22:41:55 CET] <c_14> What version are you trying to build, with what configure, on what system?
[22:43:34 CET] <kaniu> commit fd4e17fcc2b6f693f86b5620d1a0d1fa91239085
[22:43:45 CET] <kaniu> ./configure --toolchain=msvc
[22:44:22 CET] <kaniu> windows 10, cl version 19
[22:50:49 CET] <c_14> Hmm, might be an incompatibility in the build system with the version of msvc you're running. I'd suggest opening a bugreport with as much specific information as possible.
[23:17:21 CET] <kaniu> c_14, thing is, the makefile from the 2.5.4 tarball works just fine
[23:17:36 CET] <kaniu> with the same os/compiler
[23:17:49 CET] <kaniu> but I'll take a look
[00:00:00 CET] --- Thu Mar 12 2015