[Ffmpeg-devel-irc] ffmpeg.log.20140723

[burek]

[00:59] <relaxed> daala-person: ffmpeg -h encoder=libx264
[01:11] <daala-person> relaxed: Thanks.
[06:24] <slowguy> hi
[06:30] <slowguy> i need to make an application which can take a live video input and dump to a file..with a button to tell the app to skip the advertisements
[06:30] <slowguy> so that when the button is pressed then video is not dumped to file
[06:31] <slowguy> can someone give me pointers..how can i do that  either by an existing program or a method to make a new program
[08:19] <relaxed> llogan: my build on the download page now points to burek's builds
[12:06] <jchillerup> Hi, I'm trying to stream my microphone over RTP, but I want to reduce the internal buffer to a minimum (i.e. I want near-live streaming of the microphone input). Right now I use this command to do what I want, but there is over a second's worth of buffer: ffmpeg -f alsa -i hw:0,0 -acodec libmp3lame -ab 256k -ac 1 -f rtp rtp://234.5.5.5:1234
[12:15] <jchillerup> Oh, turned out to be on the receiver side
[13:27] <luc4> Hello! I know DVB-T is typically an mpeg2 ts with h264 in some resolution like 1080i or similar. Can I transcode a test video to such a format with ffmpeg to test with my video player?
[13:28] <luc4> In particular I would like to know what is the proper way to create the interlaced video. The conversion to ts and h264 is pretty simple.
[13:41] <kippi> I am using, -vf "select='gt(scene, 0)',showinfo" -f null - is there away to only show when theres movement
[14:36] <Enverex> When using concat, if you're going from one format to another, say 3 AAC files concatting to a single AAC file, does it reencode it?
[14:36] <sacarasc> Depends what options you use.
[14:36] <Enverex> e.g. ffmpeg -i "concat:wotw1.m4a|wotw2.m4a|wotw3.m4a" -c copy "wotwfykk.m4a"
[14:36] <sacarasc> That shouldn't, because you used -c copy.
[14:38] <Enverex> The reason I asked is because I saw this, which made me think it was reencoding - "size=  270535kB time=06:28:56.65 bitrate=  95.0kbits/s, video:0kB audio:266606kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 1.473666%" - that said, it only took a few seconds so it couldn't have been reencoding it, it would have taken ages to do 6 and a half hours :P
[14:40] <Enverex> Cheers sacarasc
[16:42] <nypenick> hey ffmpeg gyus
[16:42] <nypenick> I'm using the ffmpeg lib for decoding audio
[16:43] <nypenick> I have the problem that av_seek_frame seems to be imprecise
[16:44] <nypenick> for example I want to seek to timestamp 23000 but after seeking the timestamp is at 23400
[16:45] <nypenick> has anyone an idea what I'm doing wrong?
[17:22] <sfan5> nypenick: precise seeking is probably jut not possible, you may need to find the earliest seekable point before where you want to seek and decode everything until you get where you want to end up
[17:22] <sfan5> what I just said may be or may not be helpful
[17:23] <sfan5> I know almost nothing about using libav*
[17:32] <nypenick> sfan5: so av_seek_frame doesn't do that automatically?
[17:32] <sfan5> looks like it doesn't
[17:33] <nypenick> but how can I find the earliest seekable point?
[18:16] <joselo> Hi there
[18:17] <joselo> I'm new using ffmpeg and i´m having some trouble when trying to save to a file a live stream
[18:18] <joselo> ok im on it
[18:19] <joselo> http://pastebin.com/Nxb5b1E6
[18:20] <joselo> there it is
[18:21] <joselo> the problem is that the save process stops always at 1minute 4 second
[18:21] <joselo> so I always end up with a mp4 file with that duration
[18:23] <c_14> Does the same thing happen if you get rid of the -t option?
[18:28] <joselo> yeah the same result
[18:28] <joselo> also tried with -to
[18:30] <c_14> Is that ' live=1' part of the url?
[18:30] <joselo> no
[18:31] <joselo> also tried with out that param
[18:31] <joselo> just with "rtsp://192.168.0.102:554/live.sdp"
[18:31] <c_14> Have you tried using it with -codec copy?
[18:32] <joselo> no, I read a bit the docs but I wasn't able to get it to work
[18:33] <c_14> Eh, try -c copy and use something like mkv as output just for testing purposes. I'm guessing mp4 doesn't like amr_nb as an audio format.
[18:36] <joselo> ok ill try that out
[18:39] <Mavrik> hmm, do we have any comparisons of quality of libswscale resample algorithms?
[19:50] <jdolan> is there a convenient way to compare av_gettime() against a packet's pts?
[20:33] <zenderz> hi all
[20:35] <zenderz> anyone had any experience with generating smooth streaming format content. I see the format has recently been added
[20:39] <zenderz> looking at using something like: #ffmpeg -i testinput.mpg -c copy -f smoothstreaming output
[20:39] <zenderz> but there is nothing in the docs on how to use -f smoothstreaming
[20:40] <c_14> Check the docs for ismv, I'm pretty sure that's what the format's called.
[20:41] <c_14> https://ffmpeg.org/ffmpeg-formats.html#mov_002c-mp4_002c-ismv
[20:41] <zenderz> ya ismv is the file format. but smoothstreaming also generates the manifest files
[20:42] <zenderz> i think -f smoothstreaming is only newly added
[20:43] <c_14> Ah, ye.
[20:43] <c_14> https://ffmpeg.org/ffmpeg-formats.html#smoothstreaming
[20:44] <c_14> No examples yet though...
[20:44] <Katharsis> i'm using this for recording my desktop:
[20:44] <Katharsis> ffmpeg -f x11grab -s 1280x768 -r 25 -qscale 1 -i :0.0 file.mp4
[20:44] <Katharsis> i don't know why there is no sound
[20:44] <Katharsis> any ideas?
[20:44] <c_14> Because you never added sound?
[20:45] <c_14> You'll need to add an alsa or pulse device.
[20:45] <c_14> https://ffmpeg.org/ffmpeg-devices.html#alsa-1
[20:45] <c_14> https://ffmpeg.org/ffmpeg-devices.html#pulse
[20:46] <Katharsis> you mean -f alsa ?
[20:46] <Katharsis> what's the diff between them?
[20:46] <c_14> Pulse and alsa?
[20:47] <c_14> Pulse is an abstraction layer on top of alsa that lets you do fun things sometimes.
[20:49] <Katharsis> encoder 'aac' is experimental and might produce bad results.
[20:49] <Katharsis> Add '-strict experimental' if you want to use it.
[20:49] <Katharsis> i got this exception
[20:51] <c_14> If you want to use aac, the internal encoder is usually fine with decent bitrates, just add -strict experimental. You can also use one of the alternate encoders or a different codec for audio.
[20:51] <zenderz> c_14. any idea (short of reading the source code) where i can get some examples on smooth streaming use?
[21:00] <Katharsis> c_14: ok, i used `ffmpeg -f alsa -i default -f x11grab -s 1280x768 -r 25 -qscale 1 -i :0.0 -strict experimental filename.mp4` and it works
[21:00] <Katharsis> am i used the best quality (sound / video) option in here?
[21:01] <c_14> zenderz: Well, ffmpeg -i file -strict -2 -b:v bitrate -f smoothstreaming out; seems to create a directory called out with a Manifest file, and a bunch of fragments inside of some QualityLevels folders, but I can't tell you what to do with them since I don't understand the format. You could always wait here to see if somebody understands the format better or you could send an email to the ffmpeg-user ml to ask
[21:01] <c_14> around there.
[21:02] <c_14> Katharsis: That'll default to whatever settings are default for mp4, which usually aren't that great. You might want to pick codecs and bitrates by hand.
[21:02] <c_14> Maybe look at: https://trac.ffmpeg.org/wiki/Encode/H.264
[21:06] <zenderz> thanks c_14. just FYI. the manifest points to the different file chunks which can be of diferent rates for adaptive streaming (much like hls). It seems to be generating everything correctly for me but im getting some errors which i dont understand. ill wait around for a bit. Thanks again for you help
[21:07] <iRonNCSU> what exactly is the "start" value for an input video? why shouldn't it always be 0? Input #0, mov,mp4,m4a,3gp,3g2,mj2, from 'camVid60s.mp4':
[21:07] <iRonNCSU>   Metadata:
[21:08] <iRonNCSU>     major_brand     : isom
[21:08] <iRonNCSU>     minor_version   : 512
[21:08] <iRonNCSU>     compatible_brands: isomiso2avc1mp41
[21:08] <iRonNCSU>     encoder         : Lavf55.33.100
[21:08] <iRonNCSU>   Duration: 00:01:00.07, start: 0.072562, bitrate: 422 kb/s
[21:08] <iRonNCSU>     Stream #0:0(und): Video: h264 (Main) (avc1 / 0x31637661), yuv420p, 320x240, 289 kb/s, 29.97 fps, 29.97 tbr, 11988 tbn, 59.94 tbc (default)
[21:08] <c_14> Next time pastebin please.
[21:08] <iRonNCSU> oh sorry
[21:09] <iRonNCSU> http://pastebin.com/uwUcgKyT
[21:10] <iRonNCSU> basically, i'm trying to get 2 videos, ostensibly of the same framerate, to the exact same length, but the offset on this "start" on one of the videos always throws off the lengths, even if I use the same -t and -ss values
[21:10] <c_14> Are you reencoding when you're cutting?
[21:12] <iRonNCSU> yes
[21:15] <c_14> I'm not entirely sure, but it might have something to do with certain encoders adding buffers of nothing to the beginning of the stream.
[21:15] <c_14> Might want to wait for someone who knows more though.
[21:16] <iRonNCSU> in the pastbin for example, that was the file (originally an FLV) convereted to mpeg4, it was supposed to be exactly 1 minute long, but has that "start" value. the orignal flv doesn't have this value.
[21:23] <Katharsis> c_14: thx for your tip and time
[21:27] <hilacha> hi all. i'm developing a alcatel voip call recoder based in g723.1 codec inside h.323. i've identified (by the headers and the bytes count by packet) that it's 6.3kbit/s. i've extracted it from the RTP stream (like i do with the g729 codec) but i can't convert it with ffmpeg nor asterisk. To convert the stream to wav, there is need to specify additional options to ffmpeg to convert it?
[21:32] <iRonNCSU> so i've narrowed my issue down to the audio track. Even using async 1 it doesn't match up in durations. i MUST reencode the audio to make the times match (which is what adds that start: value), but the problem is that the audio is alredy 32kbps, so reencoding it sounds horrible
[22:04] <llogan> relaxed: i squawked about that a few days ago, but it should be fixed now. i would have fixed it myself but i was still travelling.
[22:20] <nexusIL> Hi I'm recording from a stream, But for different times I run the ffmpeg command I get diffrent order of stream for the channel. I'm trying to understand where this is coming from.
[22:26] <Tox1city> Hey Guys, I'm having some trouble compiling FFMpeg, I keep getting the ERROR: x265 not found
[22:26] <llogan> Tox1city: did you install x265 first?
[22:26] <Tox1city> Yes
[22:26] <llogan> does the tail of config.log give you any clues?
[22:31] <nexusIL> any ideas?
[22:35] <llogan> nexusIL: you can use a pastebin site to show the first ~50 and last ~50 lines of config.log
[22:36] <llogan> also, how did you install x265?
[22:37] <Tox1city> Thanks
[22:37] <llogan> nexusIL: sorry, that was for Tox1city
[22:37] <nexusIL> K
[22:37] <Tox1city> It seems I was missing x265.pc in my PKGconfig
[22:38] <llogan> i've seen that once before when the user attempted to install specific components of x265 and omitted x265.pc for some reason
[22:40] <llogan> i mean he was attempting to manually install specific files.
[22:41] <Tox1city> Okay
[22:42] <Tox1city> To be honest, I'm not an advanced linux user but the x265 command did tell me x265 was installed
[22:42] <Tox1city> It seems to be compiling well now
[22:44] <llogan> nexusIL: you should have asked on superuser, not stackoverflow since SO is for programming questions only
[22:50] <nexusIL> oh I see, I didn't know where to ask that
[22:54] <nexusIL> Anyway I think it has something to do with it not recognized as a program
[23:04] <nexusIL> how can I add demux capabilties to ffmpeg? I want to demux mp2
[23:07] <llogan> demuxing usually means you want to take something out of a container, but mp2 in ffmpegese refers to the mp2 audio decoder or encoder
[23:11] <nexusIL> Ok, So I'm looking on adding decoding ability to ffmpeg for mp2, I have encoding as `ffmpeg -formats` says
[23:13] <llogan> ffmpeg -decoders | grep mp2
[23:14] <llogan> ffmpeg can natively decode mp2 audio without any special steps needed
[23:15] <nexusIL> ah nice, so what's the formats refers to?
[23:15] <JEEB> demuxers/muxers
[23:15] <JEEB> lavf stuff
[23:16] <nexusIL> ok tnx
[00:00] --- Thu Jul 24 2014