[Ffmpeg-devel-irc] ffmpeg.log.20160309

[burek]

[02:26:12 CET] <mrec> the ffmpeg configure script is super complicated...
[02:37:07 CET] <ethe> mrec: indeed, there's a lot to configure.
[02:37:48 CET] <ethe> well, it's just large, which is why it looks very complicated.
[02:39:48 CET] <mrec> if you want to trace where the ffmpeg binary is enabled or disabled it's almost impossible
[02:40:14 CET] <mrec> it's very compact...
[02:43:45 CET] <J_Darnley> mrec: would you prefer some autotools?
[02:44:19 CET] <mrec> J_Darnley: I'd prefer some comments inside .. I'm digging my way through it
[02:44:28 CET] <J_Darnley> :)
[02:44:39 CET] <J_Darnley> true, it does lack those
[02:44:39 CET] <ethe> mrec: I guess the point is that it's meant to be pretty self documenting
[02:44:49 CET] <ethe> in terms of code clarity
[02:44:51 CET] <mrec> try to trace which checks "ffmpeg" is going through!
[02:44:53 CET] <mrec> good luck!
[02:46:45 CET] <ethe> why do you need to trace it, it's all output into the config.log anyways
[02:47:00 CET] <mrec> I want to modify it a little bit for my needs
[02:47:11 CET] <mrec> it's okay I'm getting forward
[02:52:12 CET] <mrec> I'm removing SDL1 and using it with SDL2 for testing (on MacOSX)
[02:52:17 CET] <mrec> that's basically what I'm doing
[02:52:27 CET] <mrec> seems to fail with SDL
[02:53:31 CET] <J_Darnley> I was sure someone was working on that.
[02:53:32 CET] <ethe> SDL -> SDL2 is a little more involved than changing the config
[02:53:36 CET] <ethe> J_Darnley: me.. lol
[02:57:06 CET] <mrec> I modified the source already and compiled it manually using gcc
[02:57:34 CET] <ethe> ffplay, or the sdl device?
[02:57:45 CET] <ethe> or both?
[02:58:15 CET] <mrec> well a little bit more.. I split ffplay removed SDL from it completely and moved SDL to a separate file
[02:58:32 CET] <mrec> finally I'll use it with cocoa
[02:58:40 CET] <mrec> it works well so far
[03:04:02 CET] <mrec> the sdl version check disabled it it wasn't just changing sdl-config to sdl2-config :-)
[03:12:26 CET] <vladmg> guys, do you know how to install ffmpeg on centos with http://pastebin.com/jvKpP0c1
[03:12:52 CET] <vladmg> on ubuntu I use "sudo add-apt-repository ppa:kirillshkrogalev/ffmpeg-next"
[03:13:17 CET] <furq> http://johnvansickle.com/ffmpeg/
[03:13:20 CET] <furq> are those builds not good enough
[03:13:23 CET] <mrec> furq: I wouldn't do that
[03:13:36 CET] <furq> ?
[03:14:01 CET] <mrec> some existing applications might be linked against an older version and break if you compile and install the latest version
[03:14:26 CET] <furq> those are static binaries
[03:14:33 CET] <mrec> --enable-shared ?
[03:15:06 CET] <mrec> if you remove it maybe ;-)
[03:15:10 CET] <furq> you seem to have me confused with somebody else
[03:15:25 CET] <mrec> oh with vladmg
[03:15:27 CET] <mrec> sorry
[03:15:42 CET] <vladmg> hahah
[03:16:09 CET] <furq> anyway if i was stuck on centos i'd probably just use those static builds
[03:16:19 CET] <furq> they don't have all the libraries from that paste but they have everything common
[03:17:32 CET] <vladmg> ok, i will test
[03:21:00 CET] <vladmg> furq: still not working.. i hate centos
[03:21:13 CET] <vladmg> thanks anyway
[04:18:22 CET] <t4nk624> Hi, does anyone know how to apply slow motion to only a specific time in an mp4? I can do it for the portion of video with this command: `ffmpeg -y -ss 00:02:10 -i KobeDunk.mp4  -filter:v "setpts=4.0*PTS" -t 10 slow-kobe-dunk.mp4` but what I would really like to do is have the whole original video and apply slow motion in that specific timeframe. Thank you very much for any help or links even if it's not a specific answer to my que
[05:13:41 CET] <PrestonL> Trying to do a live audio stream with ALSA  Stream Source -> Server says "RECEIVE_DATA", but when a client connects it just says, "WAIT_FEED"
[05:14:06 CET] <PrestonL> When I kill ffserver, a brief second of audio gets played by the client.
[05:14:10 CET] <PrestonL> Any ideas?
[05:14:46 CET] <PrestonL> using ffmpeg 3.0
[05:14:59 CET] <PrestonL> for both server and the alsa source
[05:15:12 CET] <PrestonL> does the same if I even try to stream a .ogg file to the server
[05:51:41 CET] <jookiyaya> i need a player that can do menu on bluray
[05:57:11 CET] <relaxed> jookiyaya: did you try vlc?
[05:57:48 CET] <jookiyaya> yup
[05:58:04 CET] <jookiyaya> why is it so hard to support menu
[05:59:19 CET] <relaxed> probably because it requires java
[06:00:25 CET] <jookiyaya> and?
[06:00:44 CET] <jookiyaya> if it wasn't in java. then it would have supported menu by now?
[06:01:28 CET] <jookiyaya> what about for dvd ,is it in java?
[06:02:55 CET] <relaxed> no
[06:03:35 CET] <jookiyaya> then what is it in?
[06:17:47 CET] <relaxed> jookiyaya: vlc supports bluray menus
[06:17:53 CET] <jookiyaya> it doesn't
[06:17:58 CET] <jookiyaya> i just tried
[06:18:39 CET] <jookiyaya> where does it say that it does
[06:19:03 CET] <relaxed> google, as in #videolan
[06:19:08 CET] <relaxed> ask*
[06:27:49 CET] <galex-713> Hi, whats the common container for .flac and is it possible to include chapters in it?
[06:29:39 CET] <TD-Linux> flac has its own container format by default and no, there are no chapters
[06:29:48 CET] <TD-Linux> but you can do chapters with flac in matroska
[06:30:09 CET] <TD-Linux> alternately, a separate cue file, or just separate flac files per chapter because you can play them back gaplessly
[06:36:29 CET] <galex-713> TD-Linux: so I have to make an mka?
[06:36:36 CET] <galex-713> whats a separatue cue file?
[06:58:00 CET] <TD-Linux> a way to index into an audio file. not that great of a solution
[06:58:59 CET] <galex-713> ANd whats about that? https://wiki.xiph.org/Chapter_Extension
[06:59:05 CET] <galex-713> Is that supported by VLC?
[07:03:38 CET] <galex-713> Oh wait that makes VLC crash
[07:03:50 CET] <galex-713> And mplayer freeze
[07:06:28 CET] <galex-713> Well no now I defined a chapter at 00:00:00.000 and mplayer just take a lot of time to seek
[07:06:36 CET] <galex-713> but doesnt freeze
[07:06:48 CET] <galex-713> VLC still crashes
[07:09:03 CET] <galex-713> Removed chapters, VLC works
[07:09:10 CET] <galex-713> Going to #vlc to see
[08:17:50 CET] <jookiyaya> how do i cut videos without losing x264 encoding setting info
[08:19:28 CET] <jookiyaya> and how come this command does not work?  ffmpeg64 -ss 03:00 -i 1.mkv  -to 29:00 -c copy R:\1.mkv -y
[08:21:02 CET] <pzich> jookiyaya: what doesn't work about it?
[08:21:24 CET] <jookiyaya> -ss 03:00  works but   -to 29:00  doesn't work
[08:22:15 CET] <relaxed> jookiyaya: see if it works with -ss after the input
[08:22:16 CET] <jookiyaya> if the original video is 30 min,  i want to cut  3 min of the beginning video  and last 1 min of end of the video
[08:22:17 CET] <pzich> if you're using -c copy, you'll need to copy at keyframes, and there may be other restrictions
[08:22:34 CET] <jookiyaya> -ss part works
[08:22:35 CET] <pzich> if you want frame/time accurate encoding, you'll probably need to re-encode
[08:22:56 CET] <jookiyaya> if i do:   ffmpeg64 -ss 00:00 -i 1.mkv  -to 29:00 -c copy R:\1.mkv -y
[08:22:58 CET] <jookiyaya> it works
[08:23:09 CET] <pzich> -ss 0?
[08:23:13 CET] <pzich> so, no seek?
[08:23:43 CET] <jookiyaya> no
[08:24:05 CET] <jookiyaya> if you have 30 min video,  and if you want to cut 1st min of the video and the last 1 min of the video. what do you type
[08:24:39 CET] <jookiyaya> ffmpeg64 -ss 01:00 -i 1.mkv  -to 29:00 -c copy R:\1.mkv -y   is it this?
[08:25:21 CET] <relaxed> place -ss after the input
[08:25:41 CET] <jookiyaya> no, that will screw up the video
[08:26:00 CET] <relaxed> explain
[08:26:20 CET] <jookiyaya> besides it's ss part that is working
[08:26:27 CET] <jookiyaya> -to part that is NOT working
[08:27:06 CET] <relaxed> you ask for help but assume you know everything
[08:27:19 CET] <jookiyaya> i don't
[08:27:24 CET] <relaxed> I believe with -ss after the input, -to will work
[08:27:40 CET] <jookiyaya> i just know if you put -ss after then it screws up the video
[08:27:44 CET] <jookiyaya> that part i do know
[08:28:19 CET] <relaxed> "screw up" is  vague
[08:29:01 CET] <jookiyaya> when you play back the video, it won't play for about 3 seconds
[08:29:16 CET] <jookiyaya> if you dont' believe me try it
[08:29:40 CET] <kgkg_> Hiall. A small problem: I need to add a (silent) audio stream to an IPCamera Stream. I tried both -i anullsrc -i aevalsrc=0, which I found on some guides. Both produce a "No such file or directory". I'm on Win 10 with yesterdays ffmpeg build.
[08:30:18 CET] <relaxed> jookiyaya: add -copyinkf
[08:36:47 CET] <relaxed> kgkg_: did you use "-f lavfi -i aevalsrc=0"
[08:36:49 CET] <relaxed> ?
[08:38:38 CET] <jookiyaya> relaxed okay,  what do i type exactly
[08:39:32 CET] <relaxed> jookiyaya: try, ffmpeg64 -i 1.mkv -ss 01:00 -to 29:00 -c copy -copyinkf -y R:\1.mkv
[08:39:48 CET] <jookiyaya> ok thanks
[08:41:54 CET] <jookiyaya> it worked but i screwed up the video like i said
[08:41:59 CET] <jookiyaya> it
[08:42:18 CET] <relaxed> does audio start first?
[08:42:30 CET] <jookiyaya> yes
[08:42:39 CET] <jookiyaya> but cutting part worked
[08:42:41 CET] <relaxed> add -async 1
[08:43:17 CET] <jookiyaya> can i just do this?  ffmpeg64  -ss 01:00 -i 1.mkv -to 29:00 -c copy -copyinkf -y R:\1.mkv
[08:44:00 CET] <relaxed> I don't think -to will work if -ss is before the input, which you've already found
[08:44:19 CET] <relaxed> ffmpeg64 -i 1.mkv -ss 01:00 -to 29:00 -c copy -copyinkf -async 1 -y R:\1.mkv
[08:45:43 CET] <jookiyaya> didn't work
[08:45:58 CET] <jookiyaya> cutting part worked
[08:47:23 CET] <jookiyaya> you are right:   cutting park did not work with :  ffmpeg64  -ss 01:00 -i 1.mkv -to 29:00 -c copy -copyinkf -y R:\1.mkv
[08:47:28 CET] <jookiyaya> part*
[08:48:35 CET] <relaxed> you may have to transcode, then
[08:49:56 CET] <jookiyaya> i don't understand why this is so difficult
[08:50:00 CET] <jookiyaya> [03:54] <c_14> ffmpeg -ss 1:00 -i input -t 58:00 out.mkv
[08:50:00 CET] <jookiyaya> [03:54] <c_14> or ffmpeg -i input -ss 1:00 -to 59:00 out.mkv
[08:50:00 CET] <jookiyaya> [03:55] <relaxed> maybe -copyinkf would help
[08:50:00 CET] <jookiyaya> [03:55] <jookiyaya> oh, so i have to use -to  instead of -t
[08:50:00 CET] <jookiyaya> [03:57] <c_14> you can also use ffmpeg -ss 1:00 -i input -copyts -to 59:00 out.mkv
[08:50:34 CET] <jookiyaya> what about what C_4 said?
[08:52:56 CET] <jookiyaya> is that reencoding?
[08:55:21 CET] <kgkg_> relaxed: no -f flv which I need to stream to youtube
[08:58:29 CET] <jookiyaya> relaxed  what does  -copyinkf  exactly do
[09:00:27 CET] <kgkg_> relaxed: oh, I see complex filter for the input. now it wroks! thx a bunch!
[09:05:21 CET] <jookiyaya> and is there way i don't lose the x265/x264 encoding setting info  after i cut/edit the mkv file with ffmpeg
[09:06:20 CET] <jookiyaya> x264 core 65
[09:06:21 CET] <jookiyaya> Encoding settings              : cabac=0 / ref=5 / deblock=1:-2:-2 / analyse=0x1:0x111 / me=esa / subme=7 / psy_rd=1.0:0.0 / mixed_ref=0 / me_range=16 / chroma_me=1 / trellis=0 / 8x8dct=0 / cqm=0 / deadzone=21,11 / chroma_qp_offset=-2 / threads=1 / nr=0 / decimate=1 / mbaff=0 / bframes=0 / keyint=250 / keyint_min=25 / scenecut=40 / rc=2pass / bitrate=5840 / ratetol=1.0 / qcomp=0.70 / qpmin=10
[09:06:21 CET] <jookiyaya> / qpmax=51 / qpstep=4 / cplxblur=20.0 / qblur=0.5 / ip_ratio=1.40 / aq=1:1.00
[09:06:21 CET] <jookiyaya> Default                        : Yes
[09:25:34 CET] <relaxed> jookiyaya: did you see https://trac.ffmpeg.org/wiki/Seeking ?  Look at the bottom
[09:26:05 CET] <jookiyaya> yes
[09:42:33 CET] <rossome> so im running fedora with the nvidia binary driver. i have an i4790k capable of intel's quicksync. i've got libmfx and an ffmpeg binary with quicksync enabled
[09:43:10 CET] <rossome> however i keep getting "Error initializing an internal MFX session"
[09:43:40 CET] <rossome> is there a way i can tell ffmpeg to ignoe the nvidia driver and intels so i can make this configuration work?
[09:44:05 CET] <rossome> ignor the nvidia driver and use intel's*
[09:47:25 CET] <rossome> it looks like my motherboard does not have the igpu turned on, derp
[09:50:31 CET] <relaxed> rossome: wouldn't nvenc be faster?
[09:53:01 CET] <rossome> it would and is, however the 2 stream limit is a problem
[09:53:34 CET] <rossome> i'm trying to get this to work with my emby install and i usually have 2+ streams going at once
[09:54:00 CET] <relaxed> you have a beefy cpu too
[09:54:14 CET] <rossome> wouldnt be a problem if i could afford a quadro card :'(
[09:55:09 CET] <relaxed> yeah, they are spendy
[09:55:39 CET] <rossome> here's a new rig i just build my buddy for the same purpose http://i.imgur.com/LJiLtJy.jpg
[09:56:40 CET] <relaxed> couldn't you add another nvidia card?
[09:57:11 CET] <rossome> not in my rig, and that one will eventually have an hba and a 10Gbit nic installed
[09:57:38 CET] <rossome> my rig already has the extra pcie slots used up
[09:58:24 CET] <relaxed> your cpu is enough to pick up the slack
[09:59:04 CET] <rossome> totally is, it's just be nice to make it run that much more efficiently. the cpu in the imgur pic is an i7-6900k
[09:59:47 CET] <rossome> we have no issues with transcoding as of now. i'd just like to utilize all available resources to the fullest
[10:00:12 CET] <relaxed> enable intel's igpu and use it too
[10:01:49 CET] <rossome> yeah that's what i was forgetting i think. hopefully with it enable in bios and no monitor plugged in, it's not being automatically disabled
[10:02:26 CET] <rossome> im fairly certain on the 6900 rig its enabled but not showing up in lspci
[10:02:34 CET] <rossome> enabled in bios*
[10:04:19 CET] <rossome> ill play with it on mine later tonight
[11:51:02 CET] <tuga> any one in where that can help compile ffmpeg with nvenc suport please
[11:52:03 CET] <relaxed> tuga: sure
[11:52:53 CET] <relaxed> tuga: http://johnvansickle.com/ffmpeg_nvenc.txt
[11:53:20 CET] <relaxed> note: the latest nvidia SDK is 6.0.1
[12:05:41 CET] <marco_c> Good morning everyone! I have been trying to setup a stream from my webcam using ffmpeg and ffserver, but no luck. It seems that the server is receiving the data, but when I try to connect to the stream, I don't get any data. If I use the Preroll option, I do see the start of the stream, but then it stops. I'm sure I'm missing something silly... My current setup is here: http://pastebin.com/jbtaTnSL
[12:06:37 CET] <marco_c> I am using version 3.0 on Mac (precompiled binaries). I adde the details about the versions of the libraries used on the paste
[12:32:48 CET] <rossome> tuga: you still around?
[12:33:54 CET] <tuga> yes i am where?
[12:34:10 CET] <rossome> still need halp with nvenc?
[12:34:14 CET] <tuga> yap
[12:34:24 CET] <rossome> what os you runnin?
[12:34:33 CET] <tuga> ubuntu 14.04
[12:34:40 CET] <rossome> ...
[12:36:30 CET] <rossome> all you need to do is obtain nvEncodeAPI.h and put in in /usr/include/nvenc then run ldconfig
[12:36:47 CET] <tuga> i tink i have done all
[12:36:54 CET] <tuga> i tryed all the ways
[12:36:58 CET] <rossome> https://developer.nvidia.com/video-sdk-601
[12:37:01 CET] <tuga> but seam like sdk 5.0
[12:37:03 CET] <tuga> is out date
[12:37:09 CET] <rossome> i use the newest
[12:37:33 CET] <tuga> i am tired to try and try
[12:37:58 CET] <rossome> you're obviously using --enable-nvenc correct
[12:37:59 CET] <rossome> ?
[12:38:05 CET] <tuga> yap
[12:38:21 CET] <tuga> and --enable-nvresize
[12:38:21 CET] <rossome> is it compiling and installing?
[12:38:36 CET] <tuga> no it is no compling
[12:38:42 CET] <tuga> ift give a error on hx264
[12:38:44 CET] <tuga> :(
[12:38:53 CET] <rossome> can you elaborate on the error
[12:39:03 CET] <tuga> is compli once again
[12:39:06 CET] <rossome> give me the full error message
[12:39:10 CET] <tuga> i will tell you in a few seconds
[12:39:15 CET] <rossome> k
[12:39:29 CET] <tuga> libavcodec/libavcodec.a(libx264.o): na função `X264_init': /teste/ffmpeg/libavcodec/libx264.c:820: referência indefinida a `x264_encoder_open_1 collect2: error: ld returned 1 exit status make: ** [ffmpeg_g] Erro 1
[12:39:34 CET] <tuga> amazing
[12:40:44 CET] <rossome> apt-get install libx264-dev
[12:41:14 CET] <tuga> already done this
[12:41:22 CET] <rossome> or try libavcodec-extra
[12:42:06 CET] <rossome> the problem is with your x264 library not nvenc
[12:42:22 CET] <tuga> trying the extra
[12:42:44 CET] <tuga>  --extra-cflags="-I"$HOME"/nvenc_5.0.1_sdk/Samples/common/inc"
[12:42:51 CET] <tuga> what i put on $home?
[12:43:10 CET] <rossome> dont do that
[12:43:19 CET] <rossome> you do not need to
[12:43:35 CET] <tuga> what i do?
[12:43:36 CET] <rossome> is nvEncodeAPI.h in /usr/include ?
[12:44:43 CET] <tuga> nvEncodeAPI 163 kb
[15:07:52 CET] <leonora> Hi!
[15:08:12 CET] <leonora> I'd like to create a timelapse video with ffmpeg, can some one help me with that?
[15:08:32 CET] <flux> leonora, so in practice you have a lot of images you want to convert into a video?
[15:08:41 CET] <leonora> flux: exactly
[15:18:21 CET] <relaxed> leonora: https://trac.ffmpeg.org/wiki/Create%20a%20video%20slideshow%20from%20images
[15:48:20 CET] <j03>  /join #postgis
[15:48:22 CET] <j03> oops
[16:51:50 CET] CTCP VERSION:  from PrestonL (PrestonL!~Preston at node3.thefirehorn.com) to #ffmpeg
[17:16:27 CET] <_ggg_> Hi all, I'm trying to write some color converter in C and using ffmpeg. At this moment I'm stucked as I got some distorted frame. I've posted the code and image here: http://stackoverflow.com/questions/35894522/how-to-extract-grayscale-image-from-a-video-with-ffmpeg-library Has anybody some hints? Thanks in advance
[17:26:44 CET] <kepstin> _ggg_: looks like you didn't update the ppm header to say that the image you wrote is greyscale
[17:27:39 CET] <kepstin> probably want to change the P6 to P5 - also, you want to use the 'GREY8' pixel format, not 'GREY16'
[17:47:29 CET] <_ggg_> Hello Kepstin, I've edited the code like you wrote but now there's some zoom problem. The inner part is zoomed into the picture and the rest is cropped. Any further hints? I've added a comment at http://stackoverflow.com/questions/35894522/how-to-extract-grayscale-image-from-a-video-with-ffmpeg-library/35897294#35897294 too
[18:25:56 CET] <Boscop> hi. is it possible to use the ffmpeg library to build a program that acts as a filter in a webcam stream?
[18:26:09 CET] <Boscop> (for both audio and video)
[18:26:23 CET] <Boscop> on windows
[18:26:27 CET] <Boscop> with low latency
[18:28:48 CET] <Boscop> so essentially the question boils down to, if virtual camera output is possible with ffmpeg?
[18:31:02 CET] <Boscop> DirectShow output
[18:33:11 CET] <Mavrik> There's no DS output in ffmpeg at this time.
[18:33:16 CET] <Mavrik> You could use the API and write your own tho.
[18:47:39 CET] <gvas> Hi. I'd like to downsample an audio file to ~1 sample/s. The new sample should be the moving average of the source samples. For .vox files I have tried the following switches:
[18:47:47 CET] <gvas> ffmpeg -f s16be -ar 8000 -c adpcm_ima_oki -i source.vox -af aresample=1 -map 0:a -c pcm_u8 -f data out.dat
[18:47:56 CET] <gvas> The resulting .dat file contains only 0x80 and 0x7F bytes, and the transcoding is very slow. Can somebody help me with the correct command line switches?
[19:08:33 CET] <lroe> anyone have any good ideas for load testing a live internet stream?
[19:09:30 CET] <JEEB> write emulated clients to connect to it from various VMs that you can punch into any "cloud" provider available
[19:10:25 CET] <JEEB> or if you are wanting to maximize load and not emulate an actual user, just wham the streaming server by getting all of the segments :P
[19:10:42 CET] <JEEB> (I will guess this is using one of the HTTP based streaming things)
[19:11:58 CET] <lroe> yea an HLS stream
[19:13:33 CET] <ninjudd_> is there a way to get the total duration of the video as a variable for use in a filter?
[19:14:15 CET] <ninjudd_> specifically, i want a crop filter that will pan from left to right and back, ending at the far right exactly when the video ends.
[19:16:07 CET] <durandal_1707> nope
[19:18:17 CET] <ninjudd_> ok. good to know :) I'll stop searching for it and just hardcode it
[20:56:01 CET] <SpeakerToMeat> Hello
[20:57:05 CET] <SpeakerToMeat> When I'm working with independant pictures for each frame as input, a) the right form would be (for files called 000000.j2c onwards) "-start_number 0 -i %06d.j2c" right? and b) how do I tell ffmpeg to interpret input format to be at a 24fps cadence?
[20:58:02 CET] <furq> yes, and -framerate 24 (before -i)
[20:58:23 CET] <SpeakerToMeat> Thanks
[20:58:40 CET] <SpeakerToMeat> Can I mux audio files with the images to produce a full audio/vid output in a single run?
[20:58:45 CET] <furq> sure
[20:58:49 CET] <SpeakerToMeat> Do I just add more -i ?
[20:58:57 CET] <furq> yeah it's the same way as it would work for a video
[21:00:07 CET] <SpeakerToMeat> Perfect, tyvm, I guess ffmpeg recognizes the image files as a single video track, and audio files as audio tracks... what about if I have audio in 2 mono files, should I premux them into a stereo? or can I use 2 -i and specify each to be left/right channel?
[21:00:35 CET] <furq> https://trac.ffmpeg.org/wiki/AudioChannelManipulation#a2monostereo
[21:00:42 CET] <SpeakerToMeat> Thank you
[21:01:40 CET] <SpeakerToMeat> furq: That is a supreme guide! My heartfelt thanks
[21:08:05 CET] <SpeakerToMeat> furq: Stream specifier ':a' in filtergraph description [0:a][1:a]amerge=inputs=2[aout] matches no streams
[21:08:39 CET] <SpeakerToMeat> Yet it recognizes the audio files being read: Stream #1:0: Audio: pcm_s24le ([1][0][0][0] / 0x0001), 48000 Hz, 1 channels, s32 (24 bit), 1152 kb/s
[21:09:08 CET] <furq> it should presumably be 1:a and 2:a
[21:09:22 CET] <SpeakerToMeat> Ah of course
[21:09:26 CET] <SpeakerToMeat> since streams are 1 and 2
[21:09:39 CET] <SpeakerToMeat> Bingo. thanks
[21:10:32 CET] <llogan> don't forget to add a -map for your video too
[21:11:35 CET] <SpeakerToMeat> llogan: Yes I just found out.....
[21:11:38 CET] <SpeakerToMeat> humm
[22:32:05 CET] <radu> hi! I'm processing a live stream into a mpeg1 file, however I can't tell if I'm using any GPU acceleration and if it's possible at all
[22:32:07 CET] <radu> http://pastebin.com/rdnpWipa
[22:32:22 CET] <radu> I'm not specifying any codecs
[22:33:16 CET] <J_Darnley> no, you're not
[22:33:21 CET] <radu> I run multiple processes on a server and it's starting to choke.. probably normal
[22:34:11 CET] <radu> I've read the HWAccelIntro doc
[22:34:33 CET] <radu> If I have an Intel Video Card and I use -c:v h264_qsv it would make things better fe?
[22:34:50 CET] <J_Darnley> If you want to run several, I doubt it
[22:35:04 CET] <J_Darnley> last I heard hardware decoders didn't like being shared
[22:35:15 CET] <radu> oh, didn t know that
[22:35:17 CET] <radu> :)
[22:35:57 CET] <JEEB> it's basically an ASIC
[22:36:11 CET] <JEEB> so it's limited in its performance by the drivers and the ASIC's capabilities
[22:37:13 CET] <J_Darnley> I wouldn't expect your use to be that demanding
[22:37:29 CET] <J_Darnley> What CPU is it running on?
[22:37:42 CET] <J_Darnley> What options was ffmpeg compiled with?
[22:37:46 CET] <jkqxz> Intel hardware can easily run multiple streams simultaneously, in independent processes or not (at least, it can using the vaapi/libva on Linux).
[22:38:47 CET] <radu> CentOS  Intel Core i7 (Nehalem Class Core i7), 8-core, 8GB RAM
[22:38:58 CET] <radu> it's ok atm, running ~16 processes like the above
[22:39:04 CET] <radu> but I need to add more, double that
[22:39:23 CET] <radu> I'm basically doing the jsmpeg thing: http://phoboslab.org/log/2013/09/html5-live-video-streaming-via-websockets
[22:39:29 CET] <jkqxz> Nehalem doesn't have any graphics hardware.
[22:39:30 CET] <kepstin> radu: nehalem doesn't support qsv (it doesn't have a gpu video decoder block)
[22:39:43 CET] <radu> I know, I was wondering if it'd improve with different hardware
[22:41:02 CET] <kepstin> hmm. you'd be using it for the decoder only (it doesn't support encoding mpeg1). It's hard to say, but it might free up some cpu for encoding the output stream, i guess?
[22:43:29 CET] <J_Darnley> Even if you don't use the qsv feature, a better CPU will speed things up a little.
[22:44:10 CET] <kepstin> nehalem is a pretty old core design, it doesn't even support AVX extensions :)
[22:44:47 CET] <kepstin> (although that would only affect the decoder if anything, i doubt anyone's done avx optimizations to the mpeg1 encoder ;)
[22:45:08 CET] <radu> I m not a hardware guy :) I'll look into that
[22:45:11 CET] <radu> thank you!
[22:53:04 CET] <roddyr> Greetings, first and foremost great code!
[22:53:46 CET] <roddyr> Could you please assist with a command to add a wav file (audio) to a mp.4 video file?
[22:55:12 CET] <c_14> ffmpeg -i wav -i mp4 -map 0 -map 1 -c:v copy -c:a aac out.mp4
[22:55:33 CET] <roddyr> I've crafted "ffmpeg -i video.mp4 -i audio.wav -c:v copy -c:a aac -strict experimental merged.mp4" and it successfully runs, but no audio in the mp.4 afterwards.
[22:55:44 CET] <roddyr> Thank you let me give that a try right now!
[22:56:53 CET] <kepstin> your command should have worked; if there's no '-map', it'll use all streams
[22:57:39 CET] <c_14> "all"
[22:57:44 CET] <c_14> 1 audio, 1 video and 1 subtitle
[22:58:40 CET] <kepstin> hmm, right, if the video file already had an adio track...
[22:59:03 CET] <kepstin> might want -map 0:v -map 1:a then
[23:01:22 CET] <roddyr> FANTASTIC!
[23:01:32 CET] <roddyr> ffmpeg -i wav -i mp4 -map 0 -map 1 -c:v copy -c:a aac out.mp4
[23:01:40 CET] <roddyr> did the trick, AWSOME code!
[23:01:44 CET] <roddyr> Thank you!
[23:16:46 CET] <roddyr> Greetings, after successfully adding a wav file to a mp4, on my local machine the mp4 wroks great, i.e. vid and audio. I've uploaded to youtube, now no audio.
[23:17:27 CET] <roddyr> we used: ffmpeg -i wav -i mp4 -map 0 -map 1 -c:v copy -c:a aac out.mp4
[23:18:09 CET] <roddyr> I'm researching, and so far all we have seen is "To get a single file, with both the Audio & Video Streams, you want Multiplexing set to anything but "None.""
[23:18:22 CET] <roddyr> Could this be the problem, i.e. Multiplexing?
[23:20:59 CET] <c_14> Can you upload an ffprobe of the file to a pastebin service?
[23:22:33 CET] <roddyr> Did not try, could you please recommend a known good pastebin service?
[23:22:43 CET] <c_14> sprunge.us ix.io
[23:24:19 CET] <roddyr> could you please share the ffprob syntax?
[23:24:43 CET] <c_14> ffprobe file
[23:27:32 CET] <roddyr> Thank you for your help! ffprobe file all good, syntax to upload an ffprobe of the file to sprunge.us ix.io. I'm on Linux...
[23:27:50 CET] <J_Darnley> Copy, paste!
[23:28:23 CET] <roddyr> sprunge.us ix.io copy-paste-output?
[23:28:43 CET] <furq> is there some way to prevent version.sh from running at compile time (i'm running it before configure to get around this ropey build system)
[23:28:55 CET] <furq> i tried touch .version but it still seems to run
[23:29:00 CET] <J_Darnley> Change the makefile?
[23:29:06 CET] <furq> ideally without changing the makefile
[23:30:08 CET] <c_14> roddyr: sprunge.us has a form you can use to paste into, that or pipe into curl -F 'sprunge=<-' http://sprunge.us
[23:30:28 CET] <furq> maybe i'm reading the makefile wrong but it looks like version.sh shouldn't run if .version is present
[23:35:30 CET] <roddyr> ffprobe mp4 | curl -F 'sprunge=<-' http://sprunge.us ix.io
[23:35:44 CET] <c_14> without the ix.io
[23:36:14 CET] <roddyr> Thank you: http://sprunge.us/Ojhd
[23:36:45 CET] <c_14> you probably need |& instead of |
[23:37:20 CET] <roddyr> Thank you for your help c_14!
[23:37:46 CET] <roddyr> and J_Darnley
[23:37:59 CET] <roddyr> 100% new to ffmpeg
[23:38:22 CET] <J_Darnley> 100% new to computers too by the look of things.
[23:38:30 CET] <roddyr> :-)
[00:00:00 CET] --- Thu Mar 10 2016