[Ffmpeg-devel-irc] ffmpeg.log.20120302

burek burek021 at gmail.com
Sat Mar 3 02:05:01 CET 2012


[04:32] <herdingcat> hi all, I use ffmpeg to record camera video stream and specify bitrate as 64k but seems the final output file still with a different bitrate
[07:00] <bparker> I'm trying to use swscale (it is working, but a little too slow) and I get this message:
[07:00] <bparker> [swscaler @ 0x1bc74a0] Warning: data is not aligned! This can lead to a speedloss
[07:00] <bparker> any idea how I can fix this issue?
[07:01] <bparker> libswscale I mean
[07:36] <codinho> guys, how can I do the video resize with ffmpeg
[07:36] <codinho> -s WxH does't help
[07:36] <codinho> ffmpeg -i 2.ts -vcodec copy -acodec copy -s 480x270 22.ts
[07:37] <codinho> source file is 1280x720
[07:37] <codinho> after that command output file is 1280x720 too
[07:53] <grepper> codinho: if you are copying the video it will be the same size, no ?
[07:54] <grepper> I think you need to actually encode the output to change the size ...
[07:56] <codinho> grepper, does ffmpeg has h264 encoder?
[08:47] <Haferstroh> hello, my command for screencasting is this:
[08:47] <Haferstroh> ffmpeg -f x11grab -r 25 -s 1280x800 -i :0.0 -vcodec mszh -sameq  -f alsa -i hw:0 -acodec libvorbis bildschirmvideo2.mkv
[08:48] <Haferstroh> when I delete the option: -sameq, then the videoquality is not so good than with this option enabled, although mszh is a lossless codec
[08:49] <Haferstroh> why is this ?? and what means  -sameq ??
[09:13] <cbsrobot> Haferstroh: -sameq means same quantisation - or something similar
[09:15] <Haferstroh> when I doesnt use this option, I always got fragments in the Video and q=31 what is this ?
[09:18] <cbsrobot> q=quantisation value
[09:18] <Haferstroh> this is my output:
[09:18] <cbsrobot> the less the better the output
[09:18] <Haferstroh> frame=  109 fps= 15 q=31.0 Lsize=    1317kB time=00:00:07.05 bitrate=1529.9kbits/s dup=0 drop=6    video:1228kB audio:81kB global headers:4kB muxing overhead 0.302786%
[09:19] <Haferstroh> with this command:
[09:19] <Haferstroh>  ffmpeg -f x11grab -r 15 -s 1280x800 -i :0.0 -vcodec mszh -f alsa -i hw:0 -acodec libvorbis bildschirmvideo2.mkv
[09:20] <Haferstroh> and with -sameq  then q=0.0
[09:20] <sacarasc> You should have your output options AFTER your input options.
[09:20] <cbsrobot> try -qcomp 15
[09:21] <Haferstroh> my output options are -vcodec mszh, right? what is your advise where to put it? I am a noob to this kind of stuff :-)
[09:21] <sacarasc> After all of your -i things.
[09:23] <Haferstroh> okay I give it a try, and where should I put -qcomp 15 ?  there? -vcodec mszh -qcomp 15 ?
[09:23] <cbsrobot> yes
[09:25] <Haferstroh> now when I do this command :
[09:25] <Haferstroh> ffmpeg -f x11grab -r 15 -s 1280x800 -i :0.0 -f alsa -i hw:0 -acodec libvorbis -vcodec mszh -qcomp 15 bildschirmvideo2.mkv
[09:25] <Haferstroh> it says: unknown encoder mszh
[09:27] <sacarasc> What does `ffmpeg -codecs|grep mszh` say?
[09:28] <Haferstroh> this :D V D  mszh            LCL (LossLess Codec Library) MSZH
[09:29] <cbsrobot> Haferstroh: the way it's supposed to work is: ffmpeg [options for first input] -i XXX [options for 2nd input] -i XXX [options for first output] file.mpeg [options for second output] file2.mpeg
[09:29] <sacarasc> Haferstroh: It doesn't have any E on it, so you can't encode using it.
[09:31] <Haferstroh> my first thought was, to make a lossless screencast and then encode it in the second step
[09:32] <sacarasc> -vcodec libx264 -crf 0 -preset ultrafast
[09:33] <Tjoppen> didn't you try what I pasted yesterday?
[09:33] <Tjoppen> also, -pix_fmt yuv444p (it defaults to yuv420p IIRC)
[09:33] <Haferstroh> I tried it, but it didnt worked
[09:34] <Haferstroh> when I change mszh to libx264  lossless_ultrafast is an unknown option  :File for preset 'lossless_ultrafast' not found
[09:34] <cbsrobot> use -vprofile ultrafast
[09:34] <sacarasc> Who said lossless_ultrafast?
[09:34] <Haferstroh> Tjoppen
[09:35] <sacarasc> Not that I can see...
[09:35] <Tjoppen> worksforme
[09:35] <Tjoppen> but: I discovered a while back that the preset might not be named correctly, confusing ffmpeg
[09:35] <cbsrobot> the -profile option could be used for audio or video .... !
[09:36] <cbsrobot> presets and profile for x264 will directly be passed to libx264
[09:38] <cbsrobot> so use ffmpeg -f x11grab -r 15 -s 1280x800 -i :0.0 -f alsa -i hw:0 -c:a libfaac -b:a 128k -c:v libx264 -crf 0 -vprofile high -preset ultrafast bildschirmvideo2.mkv
[09:39] <Haferstroh> thanks I will try it
[09:39] <Haferstroh> mmh says "unknown encoder libfaac"
[09:39] <cbsrobot> hehehe
[09:39] <cbsrobot> ok
[09:40] <cbsrobot> libvorbis it is then
[09:40] <Haferstroh> ok I will change it
[09:40] <cbsrobot> ffmpeg -f x11grab -r 15 -s 1280x800 -i :0.0 -f alsa -i hw:0 -c:a libvorbis -b:a 128k -c:v libx264 -crf 0 -vprofile high -preset ultrafast bildschirmvideo2.mkv
[09:40] <cbsrobot> or recompile with libfaac :-)
[09:41] <Haferstroh> x264 [error]: high profile doesn't support lossless
[09:41] <Haferstroh> [libx264 @ 0x161bcc0] Error setting profile high.
[09:41] <cbsrobot> ah
[09:41] <Haferstroh> I will delete -vprofile high maybe?
[09:41] <cbsrobot> drop the -vprofile hight
[09:41] <cbsrobot> yeah
[09:43] <Haferstroh> Thank you, this worked
[09:43] <cbsrobot> np
[09:45] <Haferstroh> But I didnt get it what this command is about, more presicely what does command does
[09:45] <Haferstroh> +this
[09:46] <sacarasc> It encodes using x264 a lossless stream.
[09:46] <sacarasc> You might also want -threads 0 in your output options to make it even quicker.
[09:46] <Haferstroh> ffmeg is clear, -f =  force ; x11grab  okay than -r = ratio or somethin = 15 fps, -s = screensize = 1280x800. -I = input this is the screen at :0.0
[09:47] <Haferstroh> okay the same with the alsa-server
[09:47] <Haferstroh> and then?
[09:47] <sacarasc> -f is format, not force.
[09:47] <sacarasc> -r is frame rate
[09:47] <Haferstroh> why is now the -c:a = -codec:audio  before -c:v = -codec:video ??
[09:48] <Haferstroh> okay -r = framerate and -f = format
[09:48] <cbsrobot> order doesn't matter
[09:48] <cbsrobot> Haferstroh: try to understand: ffmpeg [options for first input] -i XXX [options for 2nd input] -i XXX [options for first output] file.mpeg [options for second output] file2.mpeg
[09:49] <Haferstroh> and if I only want one file? then I delete the second options with file2.mpg
[09:49] <Haferstroh> ?
[09:49] <cbsrobot> the order of the options [...] normally does not matter - ok maybe a  bit but don't push it :-)
[09:49] <cbsrobot> yes
[09:49] <cbsrobot> it's just an example
[09:49] <cbsrobot> how to encode to two different files
[09:49] <Haferstroh> okay thx
[09:50] <cbsrobot> let's say you want to encode vidoe to one file and audio to another
[09:50] <Haferstroh> okay got it. and the difference between -c:a and -acodec ? is there any or is this the same ?
[09:51] <Haferstroh> I might remeber in the documentation that there is it mentioned, it is the same, but I am not sure
[09:54] <Haferstroh> sorry for my confused writing ^^
[09:55] <cbsrobot> its the same
[09:55] <cbsrobot> it's confusing but you get used to it
[09:57] <Haferstroh> What makes it hard that there are plenty of options and I dont know when to put which where *g*, e.g. this -threads 0 - thing, in first try I would put it behind the -preset ultrafast. Is this okay
[09:57] <Haferstroh> ?
[09:58] <cbsrobot> yeah
[09:58] <cbsrobot> if its before the -i it will be used for decoding
[09:58] <cbsrobot> after both -i xx it's used for encoding
[09:59] <Haferstroh> My thing is, I want to make a screencast of a virtual machine, which got one cpu used. where should I put it in this case ? for decoding or for encoding? and are there options for -threads ?
[10:00] <cbsrobot> one cpu ?
[10:00] <Haferstroh> I got 2 cpu and one of them is taken by the vm
[10:00] <cbsrobot> hehe
[10:01] <cbsrobot> then I guess you wont use this particular option
[10:01] <Haferstroh> and I thought I read something about this -threads -thing, that it has something to do with the amount of cpus used
[10:01] <cbsrobot> lol
[10:01] <Haferstroh> what is so funny about this ? :-)
[10:02] <cbsrobot> well if you know what threads does ....
[10:03] <Haferstroh> not exactly :-)
[10:03] <cbsrobot> that's what I thought: http://en.wikipedia.org/wiki/Multithreading_%28computer_architecture%29
[10:04] <cbsrobot> in short you need multiple cpus for it to work
[10:05] <Haferstroh> okay in this case this option would be redundant
[10:06] <cbsrobot> nah - not redundant - rather useless
[10:07] <Haferstroh> anyways  :-) I thank you for your help. Its really appreciated
[10:07] <cbsrobot> np - just help somebody else today and we're even :-)
[10:08] <Haferstroh> I will remember it
[10:08] <Haferstroh> bye
[10:10] Action: cbsrobot will change my quit message to read: me left because ffmpeg options are a f**king mess
[10:56] <Radex> hi, any ideas how to run variable video bitrate on x264?
[10:59] <sacarasc> -crf
[11:00] <Radex> thx will check
[11:35] <chrisb_> hi guys
[11:35] <chrisb_> not sure if on correct channel
[11:35] <chrisb_> have an aac question ...
[11:36] <zap0> does it related to ffmpeg?
[11:37] <chrisb_> im basically lookin to insert an aac packet into an mp4 stream, and i want the packet to be silence of a duration and timestamp
[11:38] <chrisb_> No its not specifically ffmpeg, but i know alot of the guys in here have excellent a/v knowledge
[11:38] <chrisb_> maybe someone would know a more specific channel
[12:24] <kriegerod> how should app dispose AVFrame filled by avcodec_decode_audio4()? afaiu it allocates buffer for samples, and this buffer should be freed. For video, we likely have avpicture_free(), and what's for audio? Should this disposing act be done if app pushes frame to filtergraph using av_asrc_buffer_add_buffer()?
[13:16] <elmobaer> Hello this morning someone here helped me to make screencasts. now I want to convert this files to .wmv. Does anyone knows how to do this properly ? If I do this, the outputfile just has sound and no video, but a white area where the video should be in windows media player
[13:18] <kriegerod> elmobaer: show what you have tried
[13:18] <elmobaer> just give me a second
[13:19] <elmobaer> my command for screencast is this :ffmpeg -f x11grab -r 25 -s 1280x800 -sameq  -i :0.0 -f alsa -i hw:0  bildschirm01.mp4
[13:19] <elmobaer> and to convert it to .wmv I use:
[13:19] <elmobaer> ffmpeg -i bildschirm01.mp4 -sameq -c:v wmv2 -c:a wmav2 bildschirm01.wmv
[13:20] <JEEB> sameq is not what you think it is
[13:20] <elmobaer> In the tutorials they write : -sameq = trying to keep quality
[13:20] <JEEB> that's untrue
[13:20] <kriegerod> try dropping -sameq option
[13:20] <elmobaer> if I use -sameq than q stays 0.0 if I delete it than q = 31.0
[13:21] <JEEB> the real meaning of -sameq is "copy quants from source" and that is not what you want in most cases
[13:21] <JEEB> you will have to use another rate control mode
[13:21] <JEEB> either bitrate or whatever equivalent of crf libavcodec might or might not have
[13:22] <JEEB> I wish I got a dollar for every fellow who gets -sameq wrong :P
[13:22] <elmobaer> okay than how should the command look like, to get a raw input like:
[13:22] <elmobaer> ffmpeg -f x11grab -r 25 -s 1280x800  -i :0.0 -f alsa -i hw:0  bildschirm01.mp4
[13:22] <kriegerod> try just without -sameq
[13:23] <elmobaer> I did and q =31.0 and quality = crappy
[13:23] <kriegerod> try with sth like -b:v 300k instead of -sameq
[13:23] <kriegerod> which will set specific video bitrate
[13:23] <elmobaer> okay than -b =  bitrate ?
[13:23] <kriegerod> y
[13:23] <elmobaer> I will try it
[13:23] <elmobaer> one sec
[13:24] <elmobaer> q stays = 31.0
[13:25] <elmobaer> I did this: ffmpeg -f x11grab -r 25 -s 1280x800 -b:v 300k  -i :0.0 -f alsa -i hw:0  bildschirm01.mp4
[13:25] <kriegerod> try much bitter bitrate
[13:25] <kriegerod> like 3m, 10m
[13:26] <elmobaer> bitter  = better ? what bitrate do you suggest ?
[13:26] <elmobaer> okay
[13:27] <elmobaer> q stays at 31.0
[13:27] <undercash> hi
[13:27] <undercash> i m streaming to rtmp is it possible to output the remaining time of the file?
[13:28] <kriegerod> afaik no
[13:29] <undercash> no?
[13:29] <undercash> ok
[13:31] <elmobaer> there is no difference in which bitrate I use. It alwasy stays with q=31.0
[13:33] <kriegerod> why dont you set output codecs on grab? ffmpeg -f x11grab -r 25 -s 1280x800 -i :0.0 -f alsa -i hw:0 -c:v wmv2 -c:a wmav2 bildschirm01.wmv
[13:33] <elmobaer> let me try it, one moment
[13:34] <kriegerod> well add -b:v also
[13:34] <kriegerod> somewhere in middle
[13:34] <kriegerod> near -c:v
[13:34] <elmobaer> bevore or after -i ?
[13:34] <kriegerod> near -c:v
[13:34] <elmobaer> okay
[13:35] <elmobaer> makes no difference
[13:35] <elmobaer> q is still at 31.0
[13:38] <kriegerod> and what about actual picture?
[13:39] <kriegerod> did you get good picture with your initial grab command?
[13:40] <Haferstroh> with the -sameq option it is okay, without it and q =31 itis crappy
[14:42] <undercash> hi again
[14:42] <undercash> i m testing a little bash script inspired from a doc on ffmpeg site
[14:42] <undercash> http://pastebin.com/RXxwvRdt
[14:43] <undercash> i would like to know if anyone know how to add a delay between the 2 video
[14:43] <undercash> works best with their site
[14:46] <Dynamic_Fail> I used to be able to stream from my webcam via: ffmpeg -f video4linux2 -i /dev/video0 -f mpegts udp://10.1.1.1:50010
[14:46] <Dynamic_Fail> I reinstalled ubuntu\
[14:46] <Dynamic_Fail> reinstalled ffmpeg and now I get this:
[14:46] <Dynamic_Fail> [video4linux2 @ 0x99302c0]Wrong size (0x0)
[14:46] <Dynamic_Fail> /dev/video0: Error while opening file
[14:46] <Dynamic_Fail> Any ideas?
[15:12] <herdingcat> hi all, I use ffmpeg to record camera video stream and specify bitrate as 64k but seems the final output file still with a different bitrate
[15:48] <Haferstroh> Hello its me again.
[15:51] <Haferstroh> I use this command to make a screencast and then I want to convert it to .wmv :
[15:51] <Haferstroh> ffmpeg -f x11grab -r 25 -s 1280x800 -i :0.0 -f alsa -i hw:0 -c:a libvo_aacenc -b:a 128k -c:v libx264 -crf 0  -preset ultrafast bildschirmvideo2.mp4
[15:51] <Haferstroh> I use this command to convert it :
[15:52] <Haferstroh> ffmep -i bildschirmvideo2.mp4 -c:v wmv2 -c:a wmav2 bildschirmvideo.wmv
[15:52] <Haferstroh> but when I open it in windows media player, then it just shows a white screen. The audio is there but nur video
[15:53] <Haferstroh> ffmep = ffmpeg
[15:53] <Haferstroh> nur video = no video
[16:13] <cbsrobot> Haferstroh: why wmv ? does somebody force you to do that ?
[16:22] <von> Hello
[16:23] <von> I just want to ask if I can build ffmpeg on my home mac and then just copy the binary to my work mac or would I have to recompile it?
[16:29] <sacarasc> If you have all the same libraries in the same places, you should be able to just copy it.
[16:31] <von> Thanks
[16:32] <sacarasc> And you put the binary in the same place.
[16:32] <von> Oh
[16:33] <von> You mean exactly the same location?
[16:33] <sacarasc> I think so, it might work otherwise, though.
[16:33] <DynamicFail> I used to be able to stream from my webcam via: ffmpeg -f video4linux2 -i /dev/video0 -f mpegts udp://10.1.1.1:50010
[16:33] <DynamicFail> On a new version of ubuntu I get this now: [video4linux2 @ 0x99302c0]Wrong size (0x0)
[16:33] <DynamicFail> /dev/video0: Error while opening file
[16:33] <DynamicFail> Anyone know why?
[16:34] <sacarasc> Can you open the stream in something else?
[16:34] <DynamicFail> ffmpeg won't stream it
[16:34] <DynamicFail> but I can open my input (my webcam) in vlc
[16:34] <sacarasc> And it's /dev/video0?
[16:34] <DynamicFail> yeah
[16:35] <sacarasc> Try adding -s wxh in the input options.
[16:36] <herdingcat> can anyone tell me how to calculate the video filename1.avi size? here's ffprobe output. Thank. http://fpaste.org/sO5E/
[16:36] <DynamicFail> ffmpeg -s wxh -f video4linux2 -i /dev/video0 -f mpegts udp://10.1.1.1:50010
[16:36] <DynamicFail> ?
[16:37] <DynamicFail> Incorrect frame size
[16:37] <sacarasc> Substituting w and h for the width and height of the video.
[16:37] <sacarasc> herdingcat: (total bitrate * length)/8
[16:38] <herdingcat> sacarasc, but I didn't get the accurate size.
[16:38] <sacarasc> Then add in some overhead, which is variable.
[16:38] <herdingcat> sacarasc, the file size is 13MB
[16:38] <herdingcat> sacarasc, could you tell me how did 13 MB come out?
[16:39] <sacarasc> (1221*83.48)/8/1024
[16:39] <sacarasc> + some overhead.
[16:39] <DynamicFail> Unknown input or output format: udp://10.1.1.1:50010
[16:39] <DynamicFail> Although I did see my camera turn on for an instant
[16:40] <herdingcat> sacarasc, oh, I get it. Thanks mate
[16:51] <necktie> hi. i'm trying to convert h264 mkv files to mp4 so they play on my ps3. i'm using "ffmpeg -i input.mkv -vcodec copy -acodec ac3 -ac 6 -ab 384k output.mp4". most work but some don't, does anyone know why?
[17:10] <rainmaker1> Is it ok to have droped frames while transcoding?
[17:10] <rainmaker1> I mean, is there any limit/ratio?
[17:29] <Haferstroh> cbsrobot: Hello, I have been away from keyboard. Yes I am forced to convert this to .wmv because the person I want to send it to just has windows xp and is not used to other things apart from windows media
[17:44] <necktie> hi. i'm trying to convert h264 mkv files to mp4 so they play on my ps3. i'm using "ffmpeg -i input.mkv -vcodec copy -acodec ac3 -ac 6 -ab 384k output.mp4". most work but some don't, does anyone know why?
[17:48] <cbsrobot> Haferstroh: I suggest putting it into a avi file with divx codec
[17:49] <cbsrobot> normally this works for most windows users
[17:49] <j3j5> hi there, I'm having problems while converting a .mp4 file to gif, the thing is that I'm getting the gif file, and it works fine, but the colors are really crappy. Here is the command and the output http://ffmpeg.pastebin.com/VVTPxKrh . My guess is that it is related to the pixel format that GIF supports (only rgb24) but I'd love to hear some tips
[17:49] <Haferstroh> cbsrobot, I tried the libxvid codec, but the terminal tells me, there is no
[17:50] <Haferstroh> and ffmpeg -codecs | grep divx gives no output
[17:55] <iive> necktie: you must first explain what do you mean by "doesn't work"
[18:26] <ePirat> can i specify the location of yasm somehow?
[18:46] <ePirat> anyone around?
[19:25] <necktie> iive: it says file corrupted on my ps3
[19:25] <necktie> plays fine on computer tho
[19:31] <iive> necktie: well, ps3 have constraints, not all mkv files are encoded within them.
[19:33] <j3j5> hey, I had to go and I don't know if somebody has answered already
[19:34] <j3j5> but I'm converting from mp4 to gif and the result has some color problems. the command and the output are here http://ffmpeg.pastebin.com/VVTPxKrh
[19:35] <j3j5> I have the feeling it's related with the pixel format for gif but I'm not sure and I'd love to hear some other ideas
[19:45] <j3j5> here you can see and image where my problem with the colors can be seen --> http://wstaw.org/m/2012/03/02/gif_vs_video.png
[19:45] <j3j5> any idea?
[19:52] <iive> j3j5: looks like 4 bit color and no dithering.
[19:54] <j3j5> iive: uhmm...ok, any idea how to get it better? I've been checking on the docs
[19:54] <j3j5> but I don't find anything to choose the amount of bits for color
[19:54] <iive> me too, and I don't see the options for swscale that control that.
[19:55] <iive> the dither is more important.
[19:56] <necktie> iive: they all have the same video specs in mediainfo
[19:57] <iive> i'm not familiar with mediainfo, does it report h264 profile and level?
[19:57] <necktie> yes
[19:57] <necktie> all are 4.1
[20:02] <necktie> iive: i think the problem has to do with the audio. i get this error http://pastebin.com/d6Ki771p
[20:04] <iive> necktie: you can remake one without audio and you'll be sure
[20:04] <j3j5> I cannot find any option related to dithering or color bit at all
[20:04] <j3j5> I think I'm going to try extracting first the frames and then using imagemagick
[20:07] <necktie> iive: yes, that works. also all movies work if i just do -ss 60. but when i do the full movie, that's when i get the problem
[20:19] <DynamicFail> Unknown input or output format: udp://10.1.1.1:50010
[20:19] <DynamicFail> ffmpeg -s 640x480 -f video4linux2 -i /dev/video0 -f mpegts -f udp://10.1.1.1:5001
[20:19] <DynamicFail> any thoughts
[20:26] <DynamicFail> Could someone help me simply stream my webcam?
[20:43] <Haferstroh> Hello, what is the common problem, if the voice of a screencast is faster than the video ?
[20:44] <Haferstroh> so it is not synchronised ?
[20:44] <Haferstroh> DynamicFail, do you wanna make a simple screencast ?
[20:45] <iive> necktie: this is even stranger. is -ss of ffmpeg command?
[20:45] <iive> It is most probably a sync issue. try different sync methods.
[20:46] <Haferstroh> what sync methods are available ?
[20:46] <Haferstroh> my ffmpeg command atm is:
[20:46] <iive> not sure... there were one or two.
[20:46] <Haferstroh> ffmpeg -f x11grab -r 25 -s 1280x800 -i :0.0 -f alsa -i hw:0 -c:a libvo_aacenc -b:a 256k -c:v libx264 -crf 0  -preset ultrafast bildschirmvideo2.mp4
[20:47] <Haferstroh> this is my screencast command
[20:47] <iive> Haferstroh: btw, I was talking to Haferstroh  :)
[20:48] <iive> yff...
[20:48] <iive> Haferstroh: btw, I was talking to necktie
[20:48] <Haferstroh> lol ^^
[20:49] <Haferstroh> anyways, maybe the answer for him suits me too ^^
[20:49] <relaxed> necktie: that's not the error. pastebin the complete command and all output.
[20:49] <Furao> I'm using ffmpeg to convert a AAC+ RAW file into a .m4a using "ffmpeg -i raw.aac -absf aac_adtstoasc -acodec copy output.m4a" but when I run "file output.m4a" it says: "ISO Media, MPEG v4 system, iTunes AAC-LC" but my RAW content is an AACv2 SBR+PS and should be "ISO Media, MPEG v4 system, version 2" is there a way to force ffmpeg to write appropriate header?
[20:50] <iive> there are vsync, async and isync, they take numbers
[20:51] <iive> Haferstroh: try another codec, like mp2 or ac3, or even pcm.
[20:51] <relaxed> Furao: maybe, look at -profile in "ffmpeg -h"
[20:51] <Furao> profile? thanks
[20:51] <microchip_> use mp4box :p
[20:52] <Furao> mp4box?
[20:52] <microchip_> yes
[20:52] <microchip_> it's a muxer for mp4/m4a
[20:53] <Furao> http://www.videohelp.com/tools/mp4box ?
[20:53] <microchip_> yes
[20:55] <Furao> provided by gpac
[20:55] <relaxed> yes
[20:59] <Furao> I will take a look at it, in mean time I'm looking at -profile and there is no flag that match v2 specs
[21:02] <Haferstroh> for conversion of a .mkv file into an avi file, what options are best suiting ? the .mkv-file contains a video encoded in libx264 and a audio encoded in libvorbis
[21:02] <DynamicFail> Haferstroh, no... I just need to figure out how to stream video (using my webcam for now but eventually will be a rtp or rtsp video stream) across my network with as low latency as possible
[21:02] <DynamicFail> I know it may require me to use a different client to view the video as well
[21:03] <Haferstroh> DynamicFail, sorry in this case I cant help you
[21:07] <DynamicFail> anyone know how to use the zero latency option with h264 and how I would change this:
[21:07] <Furao> well I was able to switch the file to v2 using an hex editor
[21:07] <DynamicFail> ffmpeg -s 640x480 -f video4linux2 -i /dev/video0 -f mpegts udp://10.1.1.1:50010
[21:08] <Furao> it cannot be more dirty
[21:12] <necktie> iive: sorry, i'm new to ffmpeg. how do you mean?
[21:14] <iive> try -async 0/1/2/3 or/and -vsync 0/1/2/3
[21:15] <Furao> microchip_ relaxed any hint on how to convert a .raw to a .m4a using MP4Box without a re-encoding?
[21:15] <rainmaker1> iive: do you know why do we need to sync audio and/or video?
[21:15] <Haferstroh> what does the numbers behind -vsync mean?
[21:16] <rainmaker1> iive: I mean, why when we have perfect input ffmpeg produces out of sync output?
[21:16] <iive> i think they mean different algorithms
[21:16] <necktie> iive: alright. what exactly does that do?
[21:16] <microchip_> Furao: mp4box -add yourfile.aac -out output.m4a
[21:16] <iive> not sure...
[21:16] <Furao> just that? I'm I was looking at all the -h sub-section
[21:17] <iive> necktie: what i suspect is that these files may have big starting a-v offset, so the player tries to play it faster than realtime and not been able to drop enough frames.
[21:17] <microchip_> Furao: wait, i think you need to add one more flag. wait a second
[21:17] <Furao> microchip_: yes it just show general help
[21:17] <Furao> -new ?
[21:17] <Furao> no
[21:18] <Haferstroh> is anything about async and vsync written in the ffmpeg documentation ?
[21:18] <microchip_> Furao: mp4box -add yourfile.aac#audio:sbr output.m4a
[21:19] <relaxed> Haferstroh: yes, read the man page.
[21:19] <necktie> iive: i get this error now http://pastebin.com/
[21:19] <Haferstroh> I havent found anything about synchronisation in there yet
[21:19] <relaxed> necktie: that's the home page.
[21:20] <necktie> relaxed: oh lol
[21:20] <necktie> iive: http://pastebin.com/g9Eg5czE
[21:21] <iive> necktie: you took my writing too literal . try 0 then 1...
[21:21] <iive> not sure for the values.
[21:21] <Furao> hmm still ISO Media, MPEG v4 system, iTunes AAC-LC
[21:21] <microchip_> Furao: mp4box -info yourfile.m4a
[21:22] <necktie> iive: i tried 0 and now i get this http://pastebin.com/xrtuPh9V
[21:23] <Furao> microchip_: http://pastebin.com/SrfXzt6r
[21:23] <microchip_> Furao: don't trust on what 'file' says
[21:23] <microchip_> Furao: looks correct
[21:23] <Furao> -info says LC
[21:24] <microchip_> Furao: SBR+PS are a superset of LC
[21:24] <iive> necktie: no idea...
[21:24] <Furao> hmm
[21:24] <Furao> the thing is ffmpeg was able to convert .raw to m4a that worked everywhere except winamp
[21:24] <Furao> where it play in slow motion
[21:25] <Furao> and the only m4a sbr+ps that worked in winamp I found where v2
[21:29] <DynamicFail> why don't more people use ffmpeg to stream
[21:29] <DynamicFail> everything is file conversion
[21:29] <DynamicFail> Is there a better tool to use for video streaming?
[21:30] <microchip_> DynamicFail: you can stream with ffserver
[21:31] <rainmaker1> DynamicFail: maybe because more people like vlc which has a GUI, but for sure you can stream with ffmpeg
[21:32] <DynamicFail> VLC seems to be more "reliable" with the streaming... but the few times I have gotten ffmpeg to stream my latency is only 20% of what vlc has.... so i would really like to get it to work with ffmpeg
[21:33] <DynamicFail> but there is next to no examples on streaming
[21:33] <rainmaker1> DynamicFail: it's not that hard
[21:33] <rainmaker1> maybe if you tell what fo you want to stream someone here will be able to help you
[21:35] <DynamicFail> My Goal: http://pastebin.ca/2123589
[21:35] <Haferstroh> My new command it this:
[21:35] <Haferstroh> ffmpeg -f x11grab -r 15 -s 1280x800 -i :0.0 -f alsa -i hw:0 -c:a libvorbis -b:a 128k -async 1  -c:v libx264 -vsync 1 -preset ultrafast bildschirmvideo2.mkv
[21:36] <DynamicFail> Nutshell version: I need to do low latency video over a lossy network (latency wayyyy over quality) from linux box and view on windows machine via some client device.
[21:36] <Haferstroh> I put -async & -vsync behind and I think it works now
[21:36] <DynamicFail> My current solution is using vlc but has way too much lag
[21:36] <Haferstroh> but my main problem is still to convert a screencast made with this command to a windows-media suitable format
[21:37] <DynamicFail> any thoughts rainmaker1 ?
[21:38] <bitpurity> Hello. Has anybody run into vp8dsp_init_arm.c:(.text.unlikely+0x310): undefined reference to `ff_vp8_luma_dc_wht_dc_armv6'?
[21:38] <bitpurity> Trying to compile ffmpeg for android with neon optimizations
[21:39] <rainmaker1> DynamicFail: I am not an expert (far from that) but wouldnt be the first step to try to stream from file?
[21:40] <rainmaker1> DynamicFail: this way, if it works as expected you can then focus on input device settings and ffmpeg parameters
[21:41] <Furao> microchip_: thanks, even if it's the same AAC-LC winamp can now play it correctly& while the output of ffmpeg was in slow motion
[21:42] <microchip_> no problem :)
[21:42] <Furao> now, I need to know which players MP4Box output will be broken :)
[21:55] <bitpurity> Actually I'm getting this error while compiling ffmpeg for android #error "unsatisfied condition: defined __ARM_PCS_VFP"
[22:09] <rainmaker1> Is there any limit/ratio of droped frames while transcoding?
[22:10] <exussum> Hi, Ive got a DVD(5) that says it has 991 1.1GB files on it - obviously that is wrong but how can I rip it using ffmpeg
[22:31] <TACPILOT> does ffmpeg just do basic deinterlacing or does it do more advanced calculations providing results like hi-end deinterlacers ?
[22:41] <iive> yadif is more like mid-range one.
[22:50] <TACPILOT> u like the results from yadif better than ffmpeg ?
[22:52] <TACPILOT> oh I C yadif is what ffmpef uses
[22:53] <rainmaker1> yadif is filter thats you use in ffmpeg. I think that ffmpeg does not use any filter by default. So if your source is interlaced and you are looking video on non-interlaced screen you should be using some filter
[22:54] <TACPILOT> thats good to know TY
[23:43] <TACPILOT> the results of yadif look pretty good. Are there any others out there that can do better, or perhaps a chain of filters that will produce better results ??
[23:44] <JEEB> not in ffmpeg, but avisynth has a filter called QTGMC
[23:44] <TACPILOT> is better than yadif ?
[23:44] <JEEB> it can get better results, sure. but it can also get lolslow
[23:44] <JEEB> http://forum.doom9.org/showthread.php?t=156028
[23:45] <TACPILOT> just need for post productions so speed not a big deal
[23:45] <JEEB> then I recommend QTGMC :)
[23:45] <TACPILOT> cool TYTY
[23:46] <JEEB> that is, if you have "true interlaced" content :)
[23:48] <JEEB> if you happen to have telecined content, you just use an inverse telecine filter for that
[23:51] <TACPILOT> content is interlaced from camera
[23:51] <ePirat> hm when trying to compile an app. which needs ffmpeg i get "av_new_stream is deprecated" is there a list somewhere of the old/new or a changelog and is it hard to fix?
[00:00] --- Sat Mar  3 2012


More information about the Ffmpeg-devel-irc mailing list