[Ffmpeg-devel-irc] ffmpeg.log.20120927

[burek]

[00:14] <elxa> [FFmpeg-devel] MVC Extension for ffmpeg: http://ffmpeg.org/pipermail/ffmpeg-devel/2012-June/126139.html
[00:14] <elxa> does anyone know the status of his work?
[00:22] <JEEBsv> elxa: as far as I can see there was a single response by mn and nothing after that
[00:24] <elxa> JEEBsv: I know, but since the guy is writing this as part of a master thesis there could be more information available
[00:24] <elxa> JEEBsv: maybe I should ask on the mailing list?
[00:25] <JEEBsv> I'm not sure if that brings up anything, you probably should poke the guy who said he was going to do something directly if you want to inquire about status
[00:25] <JEEBsv> although I'm going to be sure that if he did something, he did something partial. Or more realistically, nothing got done yet.
[00:26] <JEEBsv> most students and such seriously don't get stuff done if it's for a thesis etc. (anything merge'able etc., look at the early "put stuff on the GPU and stick into x264" kind of things as a glaring example)
[00:27] <JEEBsv> I am of course a pessimist/realist, but that does come from several years of being around #x264 etc. :)
[00:28] <JEEBsv> anyways, poke the guy himself for any status
[00:28] <JEEBsv> that's your best chance
[01:08] <booi> I'm using the "overlay" filter to map an image on top of a video, but in the process I'm losing the audio. How do i pipe out the audio? -filter_complex '[1:v]scale=640:360[a],[0][a]overlay=0:0[out]'
[01:10] <booi> ohh. i have to -map the audio as well...
[01:17] <jegade> hi, i like to add text to a given streaming - something like a news-ticker
[01:31] <elxa> JEEBsv: thx, I will do that
[05:02] <FelipeS>  is there any way to remove just the audio frames from a leaving the video alone. When I attempt to do this using ffmpeg -i segment3.ts -an -vcodec copy segment2_na.ts the segment's video pts's are modified
[05:20] <praveen> hi. has tx3g been implemented yet in FFMPEG
[05:20] <praveen> i cant find it so just checking
[06:09] <booi> I'm overlaying an image over a video using the 'overlay' filter, but I'm finding that the first few frames of the video don't have the overlay. Anyway to make it more consistent?
[08:59] <jesseg> Hi!  I'm trying to run two-pass encoding with both passes running nearly at the same time. However, pass2 seems unhappy to read from a named pipe.. Is there a way I can pipe the log data right from pass1 to the pass2 process? I have multiple CPUs.
[09:00] <jesseg> The pass2 process dies with "Error while reading file 'test.flv.toflv.sh.tmp-0.log': Bad address" as soon as the pass1 process writes to the pipe..
[09:03] <jesseg> (test.flv.toflv.sh.tmp-0.log is a named pipe that I created.)
[09:04] <jesseg> oh.. I see.. looking at the source code, it looks like it tries to read the whole thing instantly and close the file before even starting to recode..?
[09:39] <praveen> has tx3g subtitles for MP4 been implemented yet in FFMPEG?
[09:40] <cbreak-work> jesseg, doing two pass with both passes at the same time is stupid.
[09:40] <cbreak-work> the whole point of two pass encoding is the log file, which you can not analyze before the first pass is done.
[09:56] <ubitux> praveen: afaict, yes
[10:18] <dddh> hi
[10:20] <dddh> I had a working script 'time ffmpeg -i ${i} -threads 0 -vcodec libx264 -antialias auto -vpre slowfirstpass -vpre baseline -b 480k -r 13 -acodec aac -ab 192k -sameq -pass 1 -f rawvideo -an -y /dev/null && time ffmpeg -i ${i} -threads 0 -vcodec libx264 -antialias auto -vpre hq -vpre baseline -b 480k -r 13 -acodec aac -ab 192k -ac 2 -sameq -pass 2 ${i}.mp4'
[10:20] <dddh> to encode video for my zte-v9
[10:20] <dddh> today ffmpeg said : "File for preset 'slowfirstpass' not found"
[10:20] <dddh> something happened?
[10:21] <dddh> ffmpeg version 0.10.3
[10:54] <carnau> I'm trying to add a fade to a video, but don't know why. I know I have to use the fade filter, but I need to be applied only for [logo2](see pastebin). Is there any way to acomplish this? http://pastebin.com/qV6s3ZXE
[11:22] <praveen> trying to encode subtitles
[11:22] <praveen> input is mp4 and srt. i want the output to be mpegts
[11:23] <praveen> anyone has an ideaa on how to do this?
[11:23] <ubitux> what kind of text subtitles can you put into mpegts?
[11:24] <praveen> @ubitux, from wikipedia. what i think it is, is dvbsub
[11:24] <ubitux> carnau: look at the filter_complex examples in the documentation
[11:24] <ubitux> praveen: dvdsub are bitmaps subtitles
[11:24] <praveen> http://en.wikipedia.org/wiki/Comparison_of_container_formats
[11:24] <praveen> oh...
[11:24] <ubitux> we don't synthetise images from subtitles text yet
[11:24] <carnau> ubitux, I'll do, thanks
[11:25] <praveen> is there anyway to softsub an mp4 or an mpegts?
[11:25] <praveen> i've seen handbrake do softsubs for mp4 but cant get it to work on ffmpeg
[11:25] <ubitux> in mp4 you can put some softsub yes
[11:25] <JEEB> there's ttxt for mp4
[11:25] <ubitux> with -c:s mov_text maybe
[11:25] <JEEB> (3gpp timed text)
[11:25] <JEEB> yeah, that might be the codec name for it
[11:26] <praveen> oh...
[11:26] <ubitux> maybe we should add a ttxt or something.
[11:26] <ubitux> ttxt alias*
[11:26] <praveen> but when i want to segment it into a mpegts, the softsubs wont be retained right?
[11:30] <ubitux> unless you find a way to mux text subtitles, i guess no
[11:30] <ubitux> maybe some teletext would be possible, but i don't know
[11:33] <praveen> @ubitux, apple documentation on HLS says to use some sort of teletext. but i cant find very much information on it
[11:33] <praveen> http://www.quora.com/Are-subtitles-closed-captions-supported-for-HTTP-Live-Stream-videos-on-Apple-iOS-devices
[11:33] <ubitux> hls can be mp4 so&
[11:34] <JEEB> CEA-608 or CEA-708 captions (the US ATSC caption standards methinks) seem to be playable?
[11:34] <JEEB> basically now you need something to implement those if you want mpeg-ts :P
[11:34] <JEEB> (and some kind of subtitles)
[11:35] <ubitux> we don't have cea[67]08 support unfortunately
[11:35] <ubitux> patches welcome
[11:35] <JEEB> I'm still weirdly in love with ISDB-T captions
[11:35] <JEEB> they're completely over-the-top and crazy
[11:35] <praveen> @ubitux lol i dont even know half of what im doing. patches would be stretching it
[11:36] <JEEB> but they're the most robust thing I've seen so far for any kind of broadcast text-based subtitling
[11:36] <praveen> well it seems i've really hit a dead end then
[11:36] <ubitux> if you're output is mp4 you can put mov text in it
[11:36] <ubitux> for mpeg-ts you can mux dvdsubtitles if you have them
[11:37] <ubitux> what's limiting you is that there is nothing to synthetise dvdsubtitles with text-based subtitles at the moment
[11:37] <praveen> @ubitux, could you check my command for ttxt and mp4
[11:37] <praveen> http://pastebin.com/Kv9im5kj
[11:37] <praveen> i seem to get an error
[11:38] <ubitux> what's in bigbuck1.ttxt?
[11:38] <praveen> @ubitux, http://pastebin.com/5Q9fFVhR
[11:39] <JEEB> uhh
[11:39] <JEEB> that's the GPAC-specific format
[11:39] <ubitux> ah we don't support this :)
[11:39] <ubitux> praveen: convert it to srt, or ass
[11:39] <ubitux> and try muxing it again
[11:39] <ubitux> that xml-weird format is not supported
[11:40] <JEEB> it's something GPAC came up with when they started implementing 3gpp timed text I guess >_>
[11:42] <praveen> ah okay. at least i got that to work
[11:43] <praveen> @ubitux, so you are saying at the moment, there is nothing in FFMPEG to get me to do softsubs in mpegts
[11:44] <ubitux> right
[11:44] <ubitux> afaict.
[11:45] <praveen> @ubitux, but maybe any other programs i can try?
[11:45] <ubitux> no idea
[11:45] <praveen> @ubitux, ok. thanks for your help
[11:45] <ubitux> np
[11:57] <gionnico> Hi! I have a MKV video with H264 video stream, chapters, ac3 and DTS audio streams and subtitles. How can I convert this to a MP4 without losing those informations?
[11:58] <gionnico> I tried ffmpeg -i file.mkv -acodec copy -vcodec copy file.mp4 ; but that will loose chapters, subtitles, and also will only pick one audio track and drop the others.
[11:59] <relaxed> gionnico: mp4 only supports certain subs and maybe not chapters at all.
[12:01] <relaxed> To copy everything try ffmpeg -i file.mkv -map 0 -c copy output.mp4
[12:02] <gionnico> relaxed, indeed: Could not write header for output file #0 (incorrect codec parameters ?)
[12:02] <gionnico> and [mp4 @ 0x1e87450] track 4: could not find tag, codec not currently supported in container
[12:03] <gionnico> relaxed, and can I extract chapters from the stream?
[12:04] <gionnico> I'd like to make a DVD from the file, so I'd like to keep the audio tracks, the subtitles and possibly also chapters. so I need to extract all of them separately with ffmpeg? then use some another program to mux them to VIDEO_TS structure?
[12:07] <relaxed> DVDs use vobsubs which are bitmap images and I assume the subs in your input are text based like ass/srt.
[12:08] <relaxed> So you may need to hardsub them and I think chapters are just metadata.
[12:09] <gionnico> relaxed, doesnt it exist a container for this metadata alone?
[12:09] <gionnico> that some tool like dvdauthor may later use to split in chapters the dvd?
[12:16] <relaxed> gionnico: you can use mkvextract to get chapters
[12:22] <gionnico> relaxed, why is is ffmpeg converting DTS audio to AC3 in a dvd?
[12:22] <gionnico> ffmpeg -i INPUT.mkv -acodec copy -map 0:0 -map 0:1 -map 0:2 -map 0:3 -target pal-dvd -b:v 7600k -threads 0 dvd_completo.mpeg
[12:22] <gionnico> http://bpaste.net/show/jqWfpjyHBbe1mtQcyYO9/
[12:22] <JEEB> because by default AC3 is the video codec selected for DVDs
[12:22] <JEEB> if you want to copy audio
[12:22] <JEEB> you need to set -c:a copy
[12:23] <gionnico> ok but I think it's also too high bitrate (from a blu-ray) that is not supported by dvd
[12:24] <gionnico> so i'd need to keep the format but re encode the bitrate. can ffmpeg encode in DTS ?
[12:24] <JEEB> uhh
[12:24] <JEEB> if you are going to re-encode, you will want AC3
[12:24] <JEEB> DTS is notorious for being crappy at compression
[12:24] <gionnico> JEEB, ok. can I tell ffmpeg to only copy audio stream from ONE specific map?
[12:25] <JEEB> most probably yes, but I don't really use mapping that much
[12:25] <gionnico> JEEB, seems so with -c:a:#
[12:25] <relaxed> -map 0:a:$stream_number -c:a copy
[12:26] <gionnico> relaxed, uhm this seems the opposite approach. i'm going to try both
[12:26] <relaxed> If you only want a specific audio stream and to copy it
[12:27] <relaxed> You'll also need -map 0:v for video, etc.
[12:28] <relaxed> It's usually best to map exactly what you want when there are many streams in the input.
[12:28] <dddh> hm
[12:29] <gionnico> my blu-ray stream has 2 audio tracks. a 640 kb/s AC3 and a 1536kb/s DTS. i'll need to re-encode anyways because 640 is still too much for dvd that support only ~400kb/s
[12:29] <JEEB> uhh
[12:29] <gionnico> so i would re-encode from the DTS since it's a much higher bitrate source. am I wrong?
[12:29] <JEEB> DVD can take audio up to ~1400-1500kbps
[12:29] <JEEB> because it can take raw PCM
[12:29] <JEEB> 2ch
[12:29] <JEEB> 16bit
[12:29] <gionnico> JEEB, same for AC3?
[12:30] <JEEB> 48000kHz
[12:30] <gionnico> so i could copy the AC3 ?
[12:30] <JEEB> I'm pretty sure 640kbps AC3 is a-OK
[12:30] <gionnico> how about the DTS ?
[12:30] <JEEB> on the limit, and generally it compresses worse off so you usually don't want it
[12:30] <gionnico> if I try to copy the dts to dvd-target i get this erro: [dvd @ 0xba6450] Application provided invalid, non monotonically increasing dts to muxer in stream 2: 450 >= 450
[12:30] <gionnico> av_interleaved_write_frame(): Invalid argument
[12:31] <JEEB> yeah, the DVD muxer didn't like the decoding time stamps in it
[12:31] Action: dddh needs help with two pass ffmpeg/x264/vdpau encoding :(
[12:31] <gionnico> JEEB, ok i'll keep original AC3 then and re-encode other-language-DTS to AC3.
[12:31] <JEEB> I don't have the DVD specs tho, but looking at the fact that PCM tracks can go up to 1400kbps+ you should be fine with the AC3 track
[12:32] <relaxed> dddh: vdpau is used for hardware decoding
[12:32] <gionnico> The official allowed formats for the audio tracks on a DVD Video are: AC-3: 48 kHz sampling rate, 1 to 5.1 (6) channels, up to 448 kbit/s
[12:32] <JEEB> gionnico, is that the spec?
[12:32] <JEEB> or random voice on the internet
[12:32] <gionnico> JEEB, seems so ( https://en.wikipedia.org/wiki/DVD-Video#Audio_data )
[12:32] <JEEB> no
[12:32] <JEEB> wikipedia is not the spec
[12:32] <dddh> relaxed: yes, I know, but it is possible to both encode and decode on GPU
[12:33] <JEEB> the spec is when you have the PDFs next to you
[12:33] <JEEB> and you can actually read them
[12:33] <relaxed> dddh: not with ffmpeg/x264
[12:33] <JEEB> we've had enough cases of wikipedia being wrong
[12:33] <dddh> relaxed: oh
[12:37] <dddh> and I guess I have different version of ffmpeg that doesn't know about presets and two pass encoding?
[12:42] <dddh> it was SVN-r21005 built on Jan  9 2010 11:50:53 with icc 1110 ;(
[12:42] <JEEB> yeah... that's a couple of years old
[12:42] <JEEB> lol
[12:42] <gionnico> is this worrying? [dvd @ 0xaff450] buffer underflow i=1 bufi=2397 size=2560 [dvd @ 0xaff450] packet too large, ignoring buffer limits to mux it
[12:42] <gionnico> (flooded) ?
[12:43] <JEEB> gionnico, yeah -- that means that your buffer rules were broken and thus the vbv maxrate/bufsize set weren't kept
[12:44] <gionnico> JEEB, ah ok. and this is if I copy the 640 kb/s ac3 stream..
[12:47] <dddh> and what should I use now?
[12:48] <gionnico> I don't seem to get a decent bitrate for a dvd with this line: $ ffmpeg -i carnage.mkv -map 0:0 -map 0:2 -map 0:3 -target pal-dvd -b:v 7000k -threads 0 dvd_completo.mpeg
[12:48] <gionnico> it seem to ignore the -b:v bitrate option even though the output do tell the target bitrate was set
[12:49] <gionnico> is there another way of making use of the whole dvd space by adapting the bitrate? -ps option maybe?
[12:49] <gionnico> -ps doesnt seem to accept "too high" numbers like 4700000000 ...
[12:55] <dddh> finally found that two pass encoding is lame
[12:55] <dddh> http://ffmpeg.org/trac/ffmpeg/wiki/x264EncodingGuide
[12:59] <JEEB> it's not lame, it does what you need in a certain usage case
[13:00] <JEEB> 1pass crf for "I'd like more or less this quality level" and 2pass bitrate for "I want this set file size"
[13:03] <gionnico> can I use some presets while encoding mpeg-2 / dvd to increase quality for encoding time?
[13:04] <JEEB> unfortunately there are no built-in presets in the mpeg-2 encoder, and I think there are no ffpreset files for it either
[13:04] <gionnico> JEEB, so the best i can obtain is 2-pass.
[13:04] <gionnico> i hope 2-pass will respect more my bitrate target..
[13:04] <JEEB> yes
[13:05] <JEEB> -pass 1 is just a test pass
[13:05] <JEEB> it will write a log
[13:05] <gionnico> i tried with 1-pass and it created a 2 gb files instead of 4.7gb as expected!
[13:05] <JEEB> and then -pass 2 shall follow the set bit rate constraints better
[13:05] <gionnico> with 4500 kb/s instead of 7000 kb/s i had told.
[13:06] <JEEB> also, the mpeg-2 encoder does have specific settings, but I'm not sure which other than motion estimation and such you can "Safely" switch for DVDs
[13:07] <JEEB> http://x264dev.multimedia.cx/archives/102 <- if you ctrl+f for MPEG-2 you will find the settings Jason used for an "everything up to eleven" test, but i'm pretty sure that quite a few settings from there break DVD specs
[13:07] <JEEB> like 8 bframes, and a 250 frame GOP length
[13:07] <JEEB> and yes, the settings aren't exactly explained :D
[13:08] <gionnico> JEEB, anyways -target pal-dvd do set quite a number of flags already and i think they respect the spec
[13:08] <JEEB> yes, those mostly should respect the spec
[13:08] <JEEB> but I'm not sure what they set stuff like motion estimation etc. to
[13:09] <JEEB> it'd be fun to see which settings could be upped to eleven
[14:18] <carnau> I'm trying to execute this(extracted from the docs):"ffmpeg -i input.avi -i whatever.png -filter_complex 'overlay=10:main_h-overlay_h-10' output.avi", but it says: "Cannot find a matching stream for unlabeled input pad 1 on filter Parsed_overlay_1"
[14:19] <carnau> how can I label my inputs?
[14:21] <relaxed> carnau: refer to the example in the man page
[14:21] <relaxed> look at the -filter_complex section
[14:22] <carnau> hmmm, I don't have man page, let's see in the --help section
[14:23] <relaxed> http://ffmpeg.org/ffmpeg.html
[14:25] <relaxed> Instead of the "-h" output it would be great if ffmpeg displayed the man page, but maybe that's too bloated.
[14:25] <carnau> relaxed, Yes, the command I posted is from the online doc
[14:26] <carnau> it's inside http://ffmpeg.org/ffmpeg.html#toc-overlay-1
[14:26] <relaxed> "look at the -filter_complex section"
[14:29] <carnau> reading...
[14:44] <carnau> wow, so hard. finally got it. thanks relaxed
[15:00] <carnau> So, can I fade out only one input stream? I'm setting "[2:v]fade=out:155:45" inside -filter_complex. but it returns: "Too many inputs specified for the "fade" filter."
[15:03] <ubitux> carnau: what's your full cmd line?
[15:07] <ubitux> you don't need a filter complex in this particular case
[15:08] <carnau> ffmpeg -y -sameq -vsync 2 -i 'video1280x720_30seg.wmv' -i 'animation_188x224.flv' -i banner.png -filter_complex "[2:v]fade=out:155:45" custom.mp4
[15:08] <carnau> I aso want to overlay [1:v] and [2:v]
[15:12] <carnau> ubitux, i can fade out [0:v], which is the main video, but not the other inputs.
[15:13] <ubitux> are you trying to make a fade between the two videos?
[15:13] <ubitux> -sameq is most likely not doing what you expect to btw
[15:14] <carnau> no, I want to add one animation and one banner that only appears the first 5 seconds of the main video
[15:15] <ubitux> and where is the fade?
[15:16] <carnau> -filter_complex "[2:v]fade=out:155:45"
[15:17] <carnau> but it not works, it just only works if I set fade to the main video input ([0:v])
[15:18] <ubitux> no i mean, where do you want it?
[15:18] <ubitux> overlay video1280x720_30seg.wmv + banner.png at the beginning, fade, and then animation_188x224.flv ?
[15:20] <carnau> think in a music tv program. They show a video, the first 10 seconds they shoy the title / artist of the song played, and then this information fade out.
[15:21] <carnau> like this: http://www.youtube.com/watch?v=XkUwuyi1IQ4
[15:22] <carnau> just a banner info that is showed at the beginning, but not in all the video, just the first 5 seconds
[15:24] <ubitux> that's not answering my question but well.
[15:24] <ubitux> you'll likely need an overlay filter, a concat and a fade; unfortunately i don't know if you can fade between two videos
[15:27] <ubitux> basically, fade in the video, concat the result with the two videos, and maybe use the sendcmd event filter to inject a banner with overlay at some point
[15:27] <carnau> Wanto to play video1280x720_30seg.wmv + banner.png + animation_188x224.flv at the beginning, and fade only banner.png and animation_188x224.flv after 5 seconds.
[15:28] <ubitux> wait
[15:28] <ubitux> < carnau> Wanto to play video1280x720_30seg.wmv + banner.png + animation_188x224.flv at the beginning
[15:28] <ubitux> what does that mean?
[15:28] <carnau> yes, I want to show the 3 inputs at the beginning
[15:28] <ubitux> are you queuing the two video one after the other, and put the banner on both?
[15:28] <ubitux> how?
[15:29] <ubitux> overlay one on the other?, play them one after the other?, play the banner?
[15:30] <carnau> I overlay banner and animation_188x224.flv over video1280x720_30seg.wmv
[15:30] <ubitux> oh, ok.
[15:30] <ubitux> are you able to do that already?
[15:31] <carnau> yes
[15:31] <ubitux> can i see the cmd line?
[15:31] <carnau> sure, wait
[15:33] <carnau> ffmpeg -y -sameq -vsync 2 -i 'video1280x720_30seg.wmv' -vf "movie='animation_188x224.flv' [logo1]; movie=banner.png [logo2]; [in][logo1] overlay=W-w-3:3 [in+logo1]; [in+logo1][logo2] overlay=W-w-3:H-h-3 [out]" custom.mp4
[15:33] <carnau> but then, someone said me that with this I can't fade only [logo1] and [logo2], that I need to use -filter_complex
[15:34] <carnau> and here I am, facing how to do it.
[15:36] <ubitux> carnau: one last question before i try some experiment; what's the duration of video1280x720_30seg.wmv and animation_188x224.flv?
[15:36] <ubitux> and what is the size of banner?
[15:37] <carnau> video1280x720_30seg.wmv about 30 sec, animation_188x224.flv about 16 sec, and banner is 200x42 pixels
[15:38] <ubitux> thx
[15:38] <carnau> np, to you!
[15:43] <ubitux> so now you want logo1 & logo2 to disappear, right?
[15:43] <carnau> yes
[15:46] <krs> Which tool for calculating ssim ?
[15:46] <ubitux> carnau: it will go to black, but try adding ",fade=out:25" just after movie='animation_188x224.flv'
[15:47] <ubitux> not sure how you could make it fade properly to transparency
[15:47] <carnau> hmmm, let's try
[15:47] <carnau> I think it has to provide an alpha channel, not sure anyway
[15:49] <ubitux> mmh
[15:49] <ubitux> carnau: try to add: ",format=rgba,fade=out:20:25:alpha=1" then :)
[15:50] <ubitux> and do the same after movie=banner.png
[15:50] <ubitux> btw, you can try this with ffplay
[15:50] <ubitux> seems it doesn't work for the banner though
[15:50] <ubitux> since it's not animated
[15:51] <ubitux> saste: would be nice to have a pts version of fade, so it would work with image
[15:51] <ubitux> or maybe there is a trick?
[15:52] <ubitux> carnau: anyway, i got something close to what you are looking for with ./ffplay base.flv -vf "movie=anim.flv,format=rgba,fade=out:20:25:alpha=1 [logo1]; movie=banner.png [logo2]; [in][logo1] overlay=W-w-3:3 [in+logo1]; [in+logo1][logo2] overlay=W-w-3:H-h-3"
[15:52] <carnau> ubitux, oh yes, it works for the animation :D
[15:52] <ubitux> right
[15:52] <ubitux> now it might be tricky for the pick
[15:52] <ubitux> pic*
[15:58] <carnau> never used ffplay
[15:59] <ubitux> well it's just easier to test but do as you please :p
[16:01] <carnau> last chance is transform the png in a video
[16:01] <ubitux> there might be another trick
[16:01] <carnau> with an alpha channel, and process it like the animation.
[16:04] <carnau> ubitux, anyway, thanks for the help, very appreciated ;-)
[16:04] <ubitux> np :p
[16:04] <ubitux> i'm not able to find a simple solution for the banner :(
[16:05] <ubitux> trying some setpts and stuff, but it doesn't help :(
[16:05] <ubitux> ah well
[16:05] <ubitux> whatever :)
[16:55] <carnau> have to go, while I don't find another solution, I solved it transforming the png into a movie.
[16:55] <carnau> Thank you!
[17:00] <krs> Hello, can ffmpeg be used to calculate SSIM?  or is there another tool for that?
[17:04] <ubitux> no filter does that atm afaict
[17:04] <ubitux> but it shouldn't be hard to write one
[19:19] <breik0> Hello there! I'm trying to compile ffmpeg and I got this error ERROR: libfaac not found
[19:19] <breik0> I have downloaded and compiled libfaac but I still get the same error
[19:19] <diroots> hi there, I just got the 0.11.2 version of ffmpeg, and there are no more preset files in /usr/share/ffmpeg,... -vpre hq does not work, neither -preset hq,... how can I get the default available presets?
[19:43] <relaxed> diroots: look at `x264 --fullhelp|less` and then use ffmpeg ... -preset $preset ...
[19:43] <relaxed> ffmpeg now uses libx264's internal presets
[19:44] <FelipeS> Is there any way to re-encode a video and set iframes at a specific periodic interval?
[19:45] <relaxed> diroots: you probably want -preset veryslow
[19:45] <FelipeS> using AVC
[19:46] <relaxed> FelipeS: I believe that's what -g does.
[19:48] <FelipeS> relaxed, thanks!
[19:50] <tmatth> is there any reason why rtpproto doesn't proxy a "reuse" option (to be used as a UDP option)?
[19:57] <DelphiWorld> hi everyone
[19:57] <DelphiWorld> burek: :P
[19:57] <DelphiWorld> anyone know how to encode a CDA track into mp3 or anything?
[20:01] <relaxed> DelphiWorld: I would use ripperx with flac output, then ffmpeg to encode mp3s.
[20:02] <DelphiWorld> relaxed: i have the *.cda files, how can i do with them :-P
[20:02] <DelphiWorld> reactormonk: i guess putting the cda files is not enough
[20:03] <relaxed> for i in *cda; do ffmpeg -i "$i" -c:a libmp3lame -q:a 0 "${i%.*}"; done
[20:03] <relaxed> oops, that should be "${i%.*}".mp3
[20:05] <DelphiWorld> relaxed: lol. i'm confused
[20:06] <relaxed> It's a bash for loop that encodes each *.cda file to mp3.
[20:06] <relaxed> it should be:  for i in *cda; do ffmpeg -i "$i" -c:a libmp3lame -q:a 0 "${i%.*}".mp3; done
[20:08] <DelphiWorld> relaxed: Track11.cda: Invalid data found when processing input
[20:08] <DelphiWorld> relaxed: but yes, wait i'm still updating it to the latest git
[20:23] <DelphiWorld> relaxed: it's easy to rip with mPlayer?
[23:46] <bat-beaver> http://dpaste.com/807024/ am i right in thinking this is missing an mp3 encoder, don't seem to see any liblame?
[23:49] <llogan> bat-beaver: yes, you are correct
[23:49] <bat-beaver> when i do ffmpeg -F it has MPEG3 D E, is that a video codec though?
[23:49] <bat-beaver> er -formats
[23:50] <llogan> the output should include A or V to indicate if it is video or audio
[23:50] <llogan> however, you're using an ancient version of ffmpeg and i can't remember its exact behavior
[23:50] <bat-beaver> yeah i was worried about that
[23:50] <bat-beaver> normally i build my own
[23:50] <bat-beaver> but this is just a sandbox vm
[23:51] <bat-beaver> guess there might be a debian repo for newer ffmpeg around
[23:51] <llogan> you can adapt this: https://ffmpeg.org/trac/ffmpeg/wiki/UbuntuCompilationGuideHardy
[23:52] <llogan> or use a static build
[23:52] <llogan> or simply pipe from ffmpeg to lame: ffmpeg -i input -f wav - | lame -V4 - output.mp3
[00:00] --- Fri Sep 28 2012