[Ffmpeg-devel-irc] ffmpeg.log.20141001

[burek]

[00:01] <kepstin-laptop> x264 has all sorts of fancy algorithms to tweak the quantizer numbers based on frame type and perceptual modelling, but allows you to give a map of offsets as a hint to say "this area of the image is more or less important than this other area"
[00:21] <seasc> Hello, i'm quite confused by a video i'm trying to encode... it fails as it requires 'ass' as subtitle codec, but 'ass' was set...  see line 1 & 199 -> http://paste.fedoraproject.org/137959/15470141
[00:22] <seasc> however it is using: hdmv_pgs_subtitle which i've never heard before...
[00:23] <kepstin-laptop> seasc: I don't think you can currently transcode pgs subtitles into ass, although that's probably coming soon, there are people working on related stuff right now.
[00:23] <kepstin-laptop> the hdmv_pgs_subtitle format is a text subtitle format used on blurays
[00:24] <seasc> ty kepstin-laptop, any advice which codec i could try? dvd_subs didnt work, nor did srt
[00:25] <seasc> or just be patient? :)
[00:26] <kepstin-laptop> ouch, it is an image subtitle format, apparently :/
[00:27] <kepstin-laptop> so to convert to text subs they'd need OCR. I wonder if you can use "-c:s copy" to transfer them as-is to the output.
[00:27] <seasc> trying
[00:28] <seasc> nope
[00:28] <seasc> well i guess it must be without subtitles then, thank you kepstin
[00:29] <seasc> n/m its encoding... :D
[00:29] <seasc> it just took a few more secs to get started
[00:29] <kepstin-laptop> if all else fails, you can use mkvextract to save them to an external file, then put them back into your output with mkvmerge :/
[00:33] <seasc> hope i keep thinking of those tools if copy, i guess thats my new favorite subtitle codec, wont work.. thank you so much :)
[01:17] <voip_> guys what menas error, and how to fix it ?
[01:17] <voip_> [mpegts @ 0x2ebc820] PES packet size mismatchme=00:00:44.65 bitrate=2001.8kbits/s
[01:17] <voip_> [h264 @ 0x3631080] out of range intra chroma pred mode at 5 51trate=1999.1kbits/s
[01:17] <voip_> [h264 @ 0x3631080] error while decoding MB 5 50
[01:17] <voip_> [h264 @ 0x3631080] concealing 2204 DC, 2204 AC, 2204 MV errors in I frame
[09:32] <HideMyAss> Can anyone please help https://groups.google.com/forum/#!topic/discuss-webrtc/qgAtJ5415BU
[09:32] <dahat1> Can anyone see why the following modified make file cannot build librtmp (from rtmpdump) under mingw with the MSVC toolchain? http://pastebin.com/yi1QRjVp (I've also tried without renaming instances of .a to .lib) For note the error I get is seen at http://pastebin.com/KH9HTCqc
[12:27] <relaxed> j
[12:30] <sacarasc> :o
[18:09] <PoeticallyEvil> Hi, I'm using ffmpeg to screencast and I'd like to lower the cpu usage (right now it's at ~105% shared between 2 cores), here's the command: http://pastebin.com/r2M6NLES
[18:09] <PoeticallyEvil> Which option offer the best lower-cpu/(quality-of-file*file-size) ratio, right now ?
[18:11] <PoeticallyEvil> (lower-cpu*file-size)/quality or smth*
[18:55] <plumzl> Hi, I'm wondering when createing a webm, what's the difference between codec libvpx and libvpx-v9?
[18:55] <plumzl> Does it default to use v8?
[19:17] <Diogo> Hi this is possíble using ffmpeg generate the first chunks of hls shorter than the rest
[19:26] <Diogo> ?
[19:30] <dahat1> Can anyone see why the following modified make file cannot build librtmp (from rtmpdump) under mingw with the MSVC toolchain? http://pastebin.com/yi1QRjVp (I've also tried without renaming instances of .a to .lib) For note the error I get is seen at http://pastebin.com/KH9HTCqc
[19:59] <Lac3rat3d> i have some old avi files that are msmpeg4v3, i want to convert them to mpeg4
[19:59] <Lac3rat3d> any suggestions
[20:02] <ChocolateArmpits> Do the files playback correctly with ffplay?
[20:03] <Lac3rat3d> i don't know
[20:03] <Lac3rat3d> they play fine on my computer. but my ps3 doesn't support msmpeg4
[20:03] <Lac3rat3d> so i can't directplay them to ps3 via dlna
[20:05] <ChocolateArmpits> Are you familiar with ffmpeg ?
[20:05] <Lac3rat3d> yes
[20:05] <Lac3rat3d> i use it regularily to convert files
[20:05] <Lac3rat3d> just never with msmpeg4 source stream
[20:06] <ChocolateArmpits> Did you try converting any of the files yet?
[20:09] <Lac3rat3d> yep
[20:10] <Lac3rat3d> i tried converting to mp4/x264 and avi/mpeg4
[20:13] <Lac3rat3d> the source files are not high quality, i'm just looking for suggestions on the quickest/best settings to convert
[20:13] <Lac3rat3d> that will be compatible with plex -> dlna -> ps3
[20:14] <Lac3rat3d> directplay
[20:45] <Lac3rat3d> i figured out a good setting that seems to work
[20:45] <Lac3rat3d> ffmpeg -i input.avi -c:v mpeg4 -q:v 5 -tag:v DIVX -s 640x480 -c:a libmp3lame -q:a 5 -ac 2 -ar 44100 output.avi
[20:45] <Lac3rat3d> thanks anyways
[21:47] <Phlarp> can FFMPEG arbitrarily mix .mov and .mp4?
[21:51] <kepstin-laptop> Phlarp: I'm not sure what you mean by "mix".
[21:51] <Phlarp> I'd like to overlay a .mov with alpha channels onto an mp4.
[21:53] <Phlarp> tl;dr version: I tried to do this several different ways and the command isn't erroring, but the video gets made without the overlay.
[21:53] <kepstin-laptop> well, ffmpeg can read most .mp4 and .mov files, there's a filter that allows overlaying one video stream onto another to make a new video stream, and it can encode to a variety of codecs that can be stored in mp4 files
[21:58] <kepstin-laptop> should be as simple as ffmpeg -i background.mp4 -i overlay.mov -filter_complex overlay output.mp4
[21:58] <kepstin-laptop> add codec settings for encoding to taste.
[21:59] <kepstin-laptop> you might have to do some additional video scaling before overlaying, or possibly setting position on the overlay filter.
[22:13] <kippi_> hey
[22:15] <kippi_> I can get ffmpeg to detect slience, but I would like to output the current levels too
[22:15] <kippi_> trying this: http://pastebin.com/L5kNqjsH
[22:15] <kippi_> could someone point me in the right direction?
[22:41] <Phlarp> Kepstin-laptop: it was a scaling issue, not a format problem. Thanks for the tip!
[22:43] <max246> hello
[22:43] <max246> anyone expert in youtube live streaming?
[22:45] <BtbN> It's bad.
[22:45] <max246> lol
[22:45] <max246> so nobody? :d
[22:46] <BtbN> Someone humms a "bad" song in the background, stream ends and you get banned.
[22:46] <max246> well I dont have song
[22:46] <max246> just streaming an installation :p
[22:47] <BtbN> Then just record a video of it and upload it?
[22:47] <max246> was looking youtube because its free and I dont get ads
[22:47] <max246> it has to be live streaming
[22:47] <BtbN> Seriously, YouTube livestreaming is horrible.
[22:47] <max246> ok
[22:47] <max246> ustream?
[22:47] <max246> ads...
[22:47] <max246> what is my option? :p
[22:47] <BtbN> YouTube - ads
[22:48] <max246> so what do you suggest?
[22:48] <BtbN> No idea, i don't even know what you want to do.
[22:48] <max246> streaming a webcam
[22:48] <BtbN> live streaming is quite bandwidth intensive, so why would anyone give it away for free
[22:50] <max246> mhh ok
[22:57] <ChocolateArmpits> Youtube will be ok for webcame streams
[22:59] <ChocolateArmpits> You get a special link to which you output your encoded data upon creating a live event
[23:01] <hroi> hello
[23:01] <hroi> I have an MP4 encapsulated H 264 video recording
[23:02] <hroi> with hissing noise on the audio.
[23:02] <hroi> I now want to clean up the audio --- wouldnt the best choice be to separate the video and audio encoding on the file without converting or re-encoding the video (I don't want to touch the video quality)
[23:03] <hroi> I imaging its possible to separate audio and video -> fix audio -> recombine fixed_audio and video  (without re-encoding video)
[23:03] <hroi> right?
[23:03] <ChocolateArmpits> Absolutely
[23:04] <ChocolateArmpits> If you drop the video on Audacity it should read the audio tracks directly
[23:05] <ChocolateArmpits> When finished processing save it as a wave, you can then combine with the video back using ffmpeg
[23:06] <ChocolateArmpits> eg. ffmpeg -i [video input] -i [audio input] -codec copy -map 0:v -map 1:a [output]
[23:06] <ChocolateArmpits> key here would be to map the inputs using the "-map"
[23:06] <ChocolateArmpits> commands
[23:06] <ChocolateArmpits> If you don't map the inputs the output will only use the first input, your processed audio will go missing
[23:07] <ChocolateArmpits> The type of processing on audio depends on noise properties
[23:09] <hroi> ChocolateArmpits, thanks -- hm,.. so -codec copy part instructs it to simply combine the video without re-encoding, right?
[23:09] <ChocolateArmpits> -code copy says to not re-encode any audio or video streams
[23:10] <ChocolateArmpits> basically it copies them to the output
[23:10] <Mavrik> bear in mind that you'll probably want your audio in AAC for MP4 container, not raw wav that audacity produces :)
[23:10] <hroi> ChocolateArmpits, exactly what I need, thanks!
[23:10] <hroi> Mavrik, right -- AAC supposedly optimal for utube
[23:12] <ChocolateArmpits> hroi, '-vcodec copy' is more specific and orders to only copy the video stream
[23:12] <ChocolateArmpits> '-acodec copy' acts the same but only for audio streams
[23:12] <Mavrik> mhm
[23:13] <hroi> ChocolateArmpits, Mavrik --- great , think I can now proceed  (maybe make a script -- if I can script audio cleanup somehow)
[23:15] <ChocolateArmpits> hroi, the only command line option for that I think is using SoX coupled with ffmpeg
[23:15] <ChocolateArmpits> SoX has a function that samples a "silent area" that has noise only
[23:15] <ChocolateArmpits> And then uses that to make noise level adjustments
[23:16] <hroi> ChocolateArmpits,  right :)  just I need remember to shut up in the beginning of all my videos :)
[23:16] <ChocolateArmpits> To find the silent area automatically, below a threshold level you can use Ffmpeg filters
[23:16] <ChocolateArmpits> Then using the timecode found from the filter input it into SoX
[23:17] <ChocolateArmpits> the feature in SoX is called noisered
[23:17] <ChocolateArmpits> Ffmpeg silence detection filter is silencedetect
[23:17] <ChocolateArmpits> look in the filters documentation
[23:18] <ChocolateArmpits> I haven't tested this method but it should in theory work considering your noise pattern and level is the same throughout all of the recordings
[23:18] <ChocolateArmpits> If not you would have to adjust level originally on top
[23:25] <sanjosekid_82_> hi, I'm wondering if AV Reader and AV Sender in the libavformat RTP stack handle synchronization of incoming audio and video packets , or if there is additional synchronization mechanism needed
[23:27] <sanjosekid_82_> sorry, to clarify,  av_read_frame logic
[23:28] <siddeek> Hi, I have a question when building ffmpeg. I want to scale videos, what dependencies do I have to enable to make the 'scale' filter available? I want my build to be minimal.
[23:30] <siddeek> Can someone point me to docs that tells me what dependencies each ffmpeg components requires then? Such a thing would be helpful.
[00:00] --- Thu Oct  2 2014