[Ffmpeg-devel-irc] ffmpeg.log.20140715

[burek]

[00:37] <Michael653> hi there. Can a timeout be specified for ffmpeg -i http://something ? I seem to get "Unrecognized option 'timeout'", although I see a timeout here: https://ffmpeg.org/ffmpeg-protocols.html#http
[00:38] <spaam> commandline ?
[00:41] <Michael653> nevermind. Used an old version... :( I've downloaded ffmpeg today, but on another computer.
[00:41] <Michael653> thanks!
[06:59] <waressearcher2> how to join 2 videos but also make transition effect when one video dissapeared or becomes transparent and other video start to appear so in other worlds to overlay two videos for about 1 or 2 seconds with transition effect ?
[06:59] <waressearcher2> s/worlds/words/
[07:00] <waressearcher2> I don't even know how that effect called so I can find it in google, is it transition effect ?
[07:27] <waressearcher2> anyone ?
[07:30] <sacarasc> I am not sure that ffmpeg can do that. It's more of an video editor's job than an encoder.
[07:32] <waressearcher2> what about filter ?
[07:33] <waressearcher2> there should be a filter like that
[07:38] <waressearcher2> http://stackoverflow.com/questions/7565962/ffmpeg-fade-effects-between-frames from that page: "Please tell a solution for Centos server because i am using FFmpeg on this server only", really ? is it matter where do you use your ffmpeg ? is there difference between say ffmpeg on "Centos server" and on "Ubuntu" ?
[07:40] <sacarasc> By default, Centos ships an absolutely ancient one.
[07:40] <sacarasc> And Ubuntu doesn't ship ffmpeg at all.
[07:57] <waressearcher2> "-s 320x240 -aspect 4:3" why do I need to use -aspect there ?
[09:21] <AreaScout> morning, i am trying to feed ffmpeg library from memory via avio_alloc_context, the callback function to read from memory does feed ffmpeg with chunks, each chunk has a header+payload, my question is, is this the right way to do it ? or do you have to feed it in some special way i.e first the header, then the payload, because i only get a constant green screen, if i feed the file directly with
[09:21] <AreaScout> avformat_open_input it works ??
[10:10] <termos> http://blog.zencoder.com/2011/12/08/announcing-the-clouds-most-efficient-http-live-streaming/ is there something like this available in ffmpeg? Optimizing the mpegts segments file size
[14:02] <DannyZB> if anyone knows : how to automatically black-border a video to a certain ratio ( pure ffmpeg , no size calculations outside of it )
[15:19] <t4nk724> hi
[15:19] <Mavrik> ih.
[15:19] <t4nk724> how can i gram my desktop with audio and save it to mp4 format (best qualirt)?
[15:19] <t4nk724> quality*
[15:19] <t4nk724> i have problem with audi delay
[15:20] <t4nk724> ffmpeg -f alsa -ac 2 -i hw:0,0 -f x11grab -r 30 -s 1366x768 -i :0.0 -vcodec mpeg2video -sameq $1
[15:20] <t4nk724> i'm using it
[15:21] <t4nk724> and?
[15:22] <Mavrik> you really shouldn't use sameq.
[15:23] <Mavrik> there's a whole guide on ffmpeg wiki which is up to date, what you're using seems very very obsolete :)
[15:24] <t4nk724> Mavrik: i dont have a time for guides
[15:24] <t4nk724> i need best solution
[15:24] <Mavrik> Oh.
[15:24] <Mavrik> So you want us to do your solution for you.
[15:24] <t4nk724> no, I hope you know the best of it
[15:24] <Mavrik> Maybe someone else will :)
[15:24] <t4nk724> for screencast recording
[15:25] <t4nk724> it should be simple, i need best quality with audio synch
[15:26] <t4nk724> and?
[15:27] <t4nk724> i tested all from them:
[15:28] <t4nk724> http://wklej.to/hClVn/text
[15:28] <t4nk724> and nothing is good enought
[15:28] <t4nk724> ;(
[15:28] <t4nk724> where can I find some help if not here?
[15:30] <t4nk724> is there any commercial support for ffmpeg?
[15:30] <t4nk724> i need to talk with expert
[15:32] <t4nk724> hello?
[16:05] <CodingFabian> hi guys - are there 64bit static nightly builds from master available? http://ffmpeg.gusari.org/static/ builds from 2.2. I assume? Maybe related is there a timeline for 2.3?
[16:07] <c_14> I'm pretty sure gusari builds from git master HEAD
[16:08] <CodingFabian> okay, does it say somewhere how it is compiled?
[16:09] <CodingFabian> because i am getting different results using self compiled and nightly build
[16:09] <c_14> If you download the binary run ./ffmpeg -version and it'll print the configuration if that's what you wanted.
[16:11] <CodingFabian> hmm okay, thanks for the input I need to figure then out why i am getting different results on pre-compiled and my own compile :)
[16:12] <CodingFabian> thanks!
[16:14] <AreaScout> hi, i am trying to feed ffmpeg library from memory via avio_alloc_context, the callback function to read from memory does feed ffmpeg with chunks, each chunk has a header+payload, my question is, is this the right way to do it ? or do you have to feed it in some special way i.e first the header, then the payload, because i only get a constant green screen, if i feed the file directly with
[16:14] <AreaScout> avformat_open_input it works
[16:17] <Mavrik> hmm
[16:17] <Mavrik> AreaScout, anything to pass in goes through demuxer anyway
[16:17] <Mavrik> so if you set proper callbacks it should "just work"
[16:24] <AreaScout> Mavrik, so it should also work if i just pass first the half chunk and then the other, or should i pass the chunks exactly ? because i have tried really a lot now and it does not work ? on the first package i get 'unkown RoQ chunk (0000)' but the package seems to be ok, so is there an example to look at ?
[16:24] <Mavrik> AreaScout, that really depends on demuxer
[16:24] <Mavrik> I had some issues where e.g. MPEG2-TS demuxer wanted to have data in full 188B chunks
[16:24] <Mavrik> some others don't
[16:25] <AreaScout> it's id RoQ codec from DOOM3
[16:25] <Mavrik> since you didn't say which format you have it's hard to say... rule of thumb is to always pass full chunks
[16:26] <AreaScout> yep but full chunks doesn't work either ..., is it possible that av_probe_input_format does make problems ?
[16:30] <Mavrik> yes. :)
[16:30] <Mavrik> but since I don't know the format... it's hard to say
[16:30] <Mavrik> usually stepping through the code using gdb can give you some ideas what goes wrong
[16:32] <AreaScout> Mavrik, ok good idea, format is id RoQ
[16:33] <AreaScout> http://samples.mplayerhq.hu/game-formats/idroq/
[16:35] <Mavrik> AreaScout, anyway, a little hint, compile ffmpeg with debug flags when developing with libav libraries
[16:35] <Mavrik> which will let you step through code from breakpoints
[16:35] <Mavrik> makes life easier when something wierd is going on
[16:37] <AreaScout> :) k, thx cool ^^
[16:50] <liam-kelly> Is it possible to decode a mxg stream and rencode it in another format?
[16:52] <c_14> should be
[16:52] <c_14> If with mgx you are in fact referring to mxPEG
[16:52] <liam-kelly> yes mxPEG
[16:53] <liam-kelly> I am able to decode and view the stream with ffplay but not convert it with ffmpeg
[16:53] <liam-kelly> I get an error about "Unable to find a suitable output format..."
[16:55] <liam-kelly> This is what I am running ffmpeg "http://camera_ip/cgi-bin/faststream.jpg?stream=MxPEG&fps=25" -f mxg -f libx264 test.mxg
[16:57] <c_14> What format/codec are you trying to convert to?\
[16:57] <c_14> -\
[16:59] <liam-kelly> libx264
[17:00] <liam-kelly> I also compiled ffmpeg with the option to include libx264, so it should be supported
[17:00] <c_14> And you want it as a raw h264 stream or in what format do you want it?
[17:01] <c_14> If you want raw h264, use ffmpeg -i mxpegthingy -c:v libx264 [encoder settings] -f h264 out.h264
[17:05] <liam-kelly> I tried that and now I get "Invalid data found while processing input"
[17:05] <c_14> If FFmpeg can't detect the input format you'll need ffmpeg -f mxg -i mxpegthingy [etc]
[17:07] <liam-kelly> It works now! Thank you so much!
[17:08] <liam-kelly> Just a random question are there any good books on ffmpeg?
[17:12] <c_14> None I know of.
[18:28] <Steve57> I've had a rather odd "problem" with ffmpeg for the past several months (latest git updates, running on linux).
[18:29] <Steve57> It works perfectly, but shows no progress while reencoding, and the "q" key does nothing.
[18:30] <Steve57> It works fine, it just doesn't show it. Starts with its usual display, then stops updating. This is on console, btw.
[18:36] <Steve57> Doesn't seem to matter what options I use, or what I'm encoding to/from. No big deal, really, I just can't figure out what's causing it.
[19:18] <c_14> Steve57: Did/does the progress show with older versions? It might be a regression after some change. If it did/does work with an older version could you try bisecting to find the commit that breaks it?
[19:20] <Steve57> c_14: Yeah, it worked fine with a revision from last year. I have no idea at what point it crept in.
[19:20] <Steve57> But OK, I can try plowing through it.
[19:22] <c_14> you can try just building minimal versions of FFmpeg and doing short test encodes. Not perfect, but with a problem like you seem to be having it might be specific to your setup so I don't know of a better way to test.
[19:23] <Steve57> That's probably true! It almost has to be something really simple and rather unique.
[19:23] <Steve57> Anyway, thanks, I'll see what I can find out. :)
[19:51] <stephenmac7> Hello, I'm having a bit of trouble streaming an rtsp stream to HLS.
[19:51] <stephenmac7> ffmpeg does not error, but I cannot play the output.
[19:52] <stephenmac7> The problem goes away if I re-encode the h264 stream
[20:24] <stephenmac7> Nevermind, -bsf:v dump_extra seems to work
[22:07] <i_s> when piping image frames to ffmpeg for an overlay, is there a way to get the current pts? (have variable frame rate)
[23:24] <Mista-D> c_14: Thanks a lot with the concat and split filters the other day.
[23:25] <Wintershade> hey guys
[23:25] <Wintershade> a quick question
[23:25] <Wintershade> is it still normal to get the "Warning: data is not aligned! This can lead to a speedloss" thing? or am I doing something wrong again?
[23:51] <Plorkyeran> if you're using the api, it means you're not using av_malloc where you should be
[23:51] <Plorkyeran> (or your own aligned memory allocator)
[23:57] <sobaah> hi all, I am having a bit of an issue with a large MKV file (9GB), where I am using the "-attach FONTFILE.ttf" to the file and the main issue is this:    ffmpeg seems to write the entire file to memory first but I don't have the 9GB it would need.  Is there any way to around this?
[00:00] --- Wed Jul 16 2014