[Ffmpeg-devel-irc] ffmpeg.log.20150413

[burek]

[01:40:13 CEST] <elgatov> hello guys, i've ripped a couple of blurays and they have quite a big size, i was wondering if removing some audio streams would help, and if i should remove DTS bearing in mind i only have stereo
[01:41:25 CEST] <elgatov> i've tried using mediainfo to check the size of each stream but i cant find it
[01:42:09 CEST] <c_14> It'll "help", but the biggest part is the video stream.
[01:45:45 CEST] <elgatov> c_14, so they aren't that big, thanks
[01:46:15 CEST] <elgatov> and for the other part of DTS in stereo, is it multiplexed by the drivers already?
[01:46:26 CEST] <elgatov> im quite a noob with multimedia
[01:47:03 CEST] <c_14> Most video players should downmix automatically.
[01:47:27 CEST] <elgatov> ah it's called downmix, great
[01:47:29 CEST] <elgatov> thank you
[01:50:59 CEST] <chungy> DTS is lossless iirc, at least re-encoding that to a lossy format can help some
[01:51:28 CEST] <c_14> Only DTS-HD MA is lossless
[01:53:08 CEST] <chungy> i should get a blu-ray drive at some point, i have Star Trek TOS on the format and it looks great on the TV :P
[02:03:28 CEST] <c_14> elgatov: a 1:28:13 DTS HD-MA track is about 2.5GiB
[02:03:33 CEST] <c_14> So you can get a feel for the size.
[05:19:53 CEST] <ac_slater> hey all. With `libavformat`, how can I "prefill" some stream info so av_get_stream_info() wont fail? I know all of the encoding parameters, rates, etc.
[07:32:40 CEST] <Zerowalker> Is libopus limited to 256kbps?
[07:44:01 CEST] <BtbN> Propably depends on the profile, or whatever it's called
[07:44:06 CEST] <BtbN> but why would you want more?
[07:44:27 CEST] <BtbN> opus is even better than aac, which is already "perfect" at around 128~196kbps
[07:52:11 CEST] <Zerowalker> BtbN, profile? well i use it for later transcode with youtube, and as space isn't much of an enough i rather use 320kbps
[07:52:26 CEST] <Zerowalker> much of an issue*
[07:52:50 CEST] <BtbN> Well, if you don't care for space, don't transcode at all?
[07:53:30 CEST] <Zerowalker> There is a limit to everything, have thought of using wave of flac, but that's a bit overkill
[07:53:51 CEST] <Zerowalker> i just want to be able to fully use opus, i don't see why ffmpeg would limit it
[07:54:43 CEST] <BtbN> opus is extremely efficient though
[07:54:52 CEST] <BtbN> i'd expect it to be close to lossless at 256kbit/s
[07:55:13 CEST] <Zerowalker> indeed that's why i use it, and yeah but it just feels better to have that extra "overhead" if you will
[07:55:43 CEST] <Zerowalker> perhaps i need to contact an ffmpeg developer about it, is 256k an universal limit for a:b?
[07:55:44 CEST] <BtbN> max bitrate for opus seems to be 512kbps
[07:55:58 CEST] <Zerowalker> yes, but i can't get it to work with ffmpeg
[07:57:46 CEST] <Zerowalker> "The bit rate 320000 bps is unsupported. Please choose a value between 500 and 256000"
[08:01:12 CEST] <BtbN> The bitrate isn't directly given. It's computed from some other values. So it depends on the settings which bitrate is max.
[08:02:10 CEST] <BtbN> But why don't you just use vbr with max quality setting?
[08:02:52 CEST] <Zerowalker> BtbN, does Opus even have that, was sure it only had bitrate indicators, not quality
[08:03:05 CEST] <Zerowalker> what do you mean the bitrate isn't directly given though
[08:03:09 CEST] <BtbN> It has a vbr option, that accepts the values on/off/constrained
[08:03:34 CEST] <BtbN> The quality/compression is given via compression_level, from 0 to 10
[08:03:36 CEST] <Zerowalker> that simply defines, "variable bitrate" "constant bitrate" "contrained variable bitrate"
[08:03:44 CEST] <Zerowalker> yeah and default for that is 10
[08:04:05 CEST] <BtbN> Yes, and generaly vbr is better for quality than cbr.
[08:04:31 CEST] <Zerowalker> isn't vbr default as well?
[08:04:48 CEST] <BtbN> yes.
[08:05:00 CEST] <BtbN> The maximum bitrate seems to be controled by the cutoff option.
[08:05:03 CEST] <Zerowalker> so higher bitrate = higher playground
[08:05:43 CEST] <BtbN> http://git.videolan.org/?p=ffmpeg.git;a=blob;f=libavcodec/libopusenc.c;h=785460972d0eac1ec2f6555dadd01ee436098bfa;hb=HEAD#l228
[08:05:44 CEST] <Zerowalker> but still the bitrate limit decides how much of an average bitrate it has to limit itself to (if it need that much)
[08:06:31 CEST] <Zerowalker> if (avctx->bit_rate < 500 || avctx->bit_rate > 256000 * avctx->channels)
[08:07:15 CEST] <BtbN> Well, won't be there for fun.
[08:07:20 CEST] <Zerowalker> well will make a ticket anyhow, as it's a defect, even if many aren't bothered by it you are prevented from utilizing opus fully
[08:07:23 CEST] <BtbN> Maybe 256kbps is the max per channel?
[08:07:37 CEST] <Zerowalker> then stereo = 512k
[08:07:44 CEST] <Zerowalker> hmm
[08:07:51 CEST] <Zerowalker> wait that might be it
[08:08:44 CEST] <Zerowalker> BtbN, yeah that was it. I was blind, didn't notice that the file i had been testing on was Mono -_-
[08:08:49 CEST] <Zerowalker> thank god i didn't make the ticket;P
[12:46:31 CEST] <chungy> Is there any real reason -to is not available as an input option?
[13:24:25 CEST] <justinX> "-to time_stop       record or transcode stop time" ?
[13:25:52 CEST] <justinX> chungy: yeah that looks like it is supposed to be per input file to me....
[13:29:31 CEST] <c_14> chungy: nobody implemented it is the most likely reason
[13:37:09 CEST] <chungy> heh yeah. Just seemed silly that -t can be used for input but not -to
[15:19:52 CEST] <kos2dk1> Looking someone for custom FFmpeg Modification, will pay. Contact me with PM
[15:26:40 CEST] <justinX> kos2dk1: you *should* probably tell something about *what* custom Modifications you want to make
[16:34:02 CEST] <d3fault_> how come when i join a video file recorded with raspivid together with an audio file, they get all out of sync randomly and inconsistently (sometimes it catches back up for a bit)?
[16:57:40 CEST] <Russ_> I have two videos 1.mp4 and 2.mp4 how to add them using ffmpeg
[16:58:59 CEST] <kepstin-laptop> by "add", you mean concatenate so one plays after the other?
[17:00:00 CEST] <kepstin-laptop> Russ_: https://www.ffmpeg.org/faq.html#How-can-I-concatenate-video-files_003f
[17:00:29 CEST] <kepstin-laptop> if your videos are the same codec, resolution, etc., you might be able to use the concat demuxer; otherwise you'll want to use the concat filter and re-encode.
[17:00:35 CEST] <Russ_> kepstin-laptop_ I did that and even file generated but that file imean video is not playing
[17:00:46 CEST] <Russ_> ok
[17:00:51 CEST] <Russ_> let me try again :)
[17:06:34 CEST] <Russ_> it's just showing first video but not second
[17:08:44 CEST] <Russ_> k
[17:10:30 CEST] <loki_> can somebody explain me one thing... if i build ffmpeg by my own, can i use optimization or all optimizations will be done automaticly?
[17:10:57 CEST] <loki_> for example --optflags or something
[17:11:34 CEST] <Russ_> it's done thanks guys
[17:12:39 CEST] <pron> mm if i have .ts file , wich starts with garbage till the first keyframe ( if i understand correctly ) whats the correct way to remove that grabage ?
[19:43:48 CEST] <QuakePhil9000> Can someone please help me with this error? http://pastebin.com/Q7XTZxL1
[19:45:57 CEST] <QuakePhil9000> If I convert them to pngs, it works just fine
[19:48:17 CEST] <filo1234> hi guys
[19:49:31 CEST] <filo1234> I have a video.mkv that contains two audio trace, Italian and English, so how can I convert it only with an audio trace?
[19:56:40 CEST] <filo1234> I have to use -vlang code?
[19:57:37 CEST] <kepstin-laptop> filo1234: you'd use the -map option to select tracks, e.g. "-map 0:v:0 -map 0:a:0" to select the first audio and first video track, then use "-c copy" to avoid reencoding.
[19:58:57 CEST] <filo1234> kepstin-laptop: ok but how can I see where first trace is mapped?
[19:59:16 CEST] <filo1234> kepstin-laptop: I'll paste you some ffmpeg output
[19:59:38 CEST] <kepstin-laptop> filo1234: the output of ffmpeg says which tracks are which if you just run "ffmpeg -i inputfile"
[19:59:46 CEST] <kepstin-laptop> or "ffprobe inputfile"
[20:02:29 CEST] <capradmar> Hi! I wonder why there is a libavformat dependency with Jackmp in the ffmpeg latest version (2.6.1)
[20:03:04 CEST] <filo1234> kepstin-laptop: http://paste.ubuntu.com/10817443/
[20:04:37 CEST] <kepstin-laptop> filo1234: ok, then you'd want to use either "-map 0:0 -map 0:1" for italian audio or "-map 0:0 -map 0:2" for english audio.
[20:04:51 CEST] <kepstin-laptop> the stuff after the Stream # is the stream id that map can use
[20:06:59 CEST] <filo1234> ok kepstin-laptop ok but, for my knowledge, how do you know that map 0:1 is Italian and map 0:2 is English?
[20:07:15 CEST] <kepstin-laptop> because it says "ita" and "eng" beside them, which are the language codes
[20:07:32 CEST] <filo1234> uhhuh wait
[20:07:53 CEST] <kepstin-laptop> on the last 3 lines in your paste
[20:08:00 CEST] <filo1234> OMG sorry
[20:08:13 CEST] <filo1234> I see that know... :|
[20:08:43 CEST] <filo1234> kepstin-laptop: thanks a lot you get my eyes opened
[20:12:23 CEST] <QuakePhil9000> I guess for now I can use something gifsicle to hold ffmpeg's hand
[20:13:08 CEST] <filo1234> kepstin-laptop: thans again
[20:13:09 CEST] <filo1234> bye
[21:30:54 CEST] Action: DragonsLord is back (gone 22:20:24)
[21:31:12 CEST] <DragonsLord> Good evening to everybody! :)
[21:39:36 CEST] <mtcjayne> How do I diagnose a segmentation fault which occurs when combining two filters, but not when using one or the other?
[21:40:10 CEST] <mtcjayne> http://pastebin.com/x6A8nnAJ
[21:40:43 CEST] <mtcjayne> I would like to get the first scale + the fade combined, but this causes a segmentation fault.
[21:44:23 CEST] <mtcjayne> An updated build was available. I am no longer seeing the behavior.
[22:09:41 CEST] <Airwave> Hi. I'm trying to decode a g2m file using ffmpeg. It's shown as "Stream #0:2(nob): Video: none (G2M5 / 0x354D3247), 2560x1440, 229 kb/s, 1k tbr, 1k tbn, 1k tbc". From what I see, ffmpeg should have support for G2M5, but it still says "No decoder for stream #0:2, filtering impossible".
[22:49:02 CEST] <heiner> I can convert a video file to an animated GIF using: ffmpeg -i f.mp4 -t 10 -r 10 image.gif
[22:49:33 CEST] <heiner> Is there also an easy way to create a time lapse GIF instead of a GIF containing all frames?
[22:51:00 CEST] <heiner> I search for a way to use e.g.  one frame every 20 seconds and create an animated GIF displaying each frame for one second
[22:52:33 CEST] <heiner> Currently I create many image files (one per frame) and stitch them together to an GIF using ImageMagick's "convert"
[22:55:59 CEST] <heiner> Should I use a different tool for this purpose?
[23:01:24 CEST] <gallardo> may someone please recommend a fast encoding lossless video codec
[23:01:34 CEST] <gallardo> I am trying dirac but it is so slow
[23:17:26 CEST] <loki_> can somebody explain me one thing, why can ffmpeg cosume less cpu when for example i run mplayer? are there some runtime kernel changes or what?
[23:18:01 CEST] <loki_> i mean less cpu and equal job done in fps. how it is possible?
[23:23:09 CEST] <klaxa> it depends on what you do
[23:23:25 CEST] <klaxa> codec copy will consume less cpu than having to decode frames
[23:23:39 CEST] <klaxa> and displaying decoded frames also takes some cpu time
[23:49:36 CEST] <c_14> gallardo: lossless H.264, ffv1 or huffyuv maybe.
[00:00:00 CEST] --- Tue Apr 14 2015