[Ffmpeg-devel-irc] ffmpeg.log.20121226

[burek]

[00:19] <Zeeflo> do I need to react to this message: [mpeg4 @ 0x46288e0] Invalid and inefficient vfw-avi packed B frames detected ?
[00:21] <JEEBsv> not in any other way than try not to create such files yourself :)
[00:22] <JEEBsv> the decoding itself should go just fine
[01:26] <kewo> how do i permanently change tint/hue of a video
[01:29] <beastd> kewo: permanent as in the video itself. probably using ffmpeg's hue filter or similar and re-encode. but beware is not a light step.
[01:30] <kewo> "light step"?
[01:30] <beastd> i mean you probably lose quality and will need to find good params for encoding.
[02:48] <mapreduce> Is calling av_free_packet(pkt) then av_free(pkt) a problem?
[02:48] <mapreduce> same pointer
[02:53] <ubitux> nope
[02:53] <ubitux> the pkt can be on the stack
[03:57] <mapreduce> ubitux: Thanks.
[03:58] <brx_> whats the shortest possible duration for a h264 video?
[03:59] <brx_> for example if i record just a second sometimes the video seems to trim back to the closest keyframe (and the sound continues whilst the video freezes, for half a second +/-)
[04:01] <mapreduce> Looking in mux.c I see av_interleaved_write_frame, which contains an infinite loop (for (;;)).  I seem to be stuck in that every so often, particularly when my packet's pts or dts ends up as 0 but sometimes when that doesn't happen too.
[04:01] <mapreduce> It looks like it will do the same thing each time after the 2nd iteration, so I don't know why that's written as an infinite loop.
[04:58] <brx_> can i trim the audio of a h264 mp4 to the closest video keyframe?
[04:58] <brx_> (im sometimes getting a half second stall in the video at the end of the file)
[05:16] <jmhodges> hey, i'm trying to make a gif from a m4v but i'm getting oddly colored results on os x.
[05:17] <jmhodges> with -pix_fmt rgb24, the message "Incompatible pixel format 'rgb24' for codec 'gif', auto-selecting format 'pal8'"
[05:18] <jmhodges> i tried adding a -f gif, but that ruins the output%03d.gif format of the output (where i'm using gifsicle after to clean it up)
[05:19] <jmhodges> its bluish.
[05:21] <jmhodges> well, actually, with -f gif, the color is very red. no idea
[05:28] <ubitux> try to add format=rgb8,format=rgb24 at the end of your filtergraph
[05:28] <ubitux> you might get better colors
[05:28] <ubitux> also note that the gif encoder doesn't compress very much currently, so your output file will get bigger quite fast
[05:32] <jmhodges> ubitux: yep, using ImageMagick's convert to clean that up
[05:32] <jmhodges> i'll try that
[05:33] <jmhodges> hm, no go. same blue. https://gist.github.com/15a764ef0fad5782949e
[05:34] <jmhodges> apologies.
[05:35] <jmhodges> https://gist.github.com/39eb4034fbb3c4126b2e
[05:35] <jmhodges> ubitux: ^
[05:35] <ubitux> move -s 355x270 into your filtergraph
[05:35] <ubitux> scale=355:270,format=...
[05:36] <ubitux> (otherwise it is added at the end, after the formats)
[05:37] <ubitux> (iirc, maybe)
[05:37] Action: jmhodges nods
[05:37] <jmhodges> it worked. what's weird is that if you make the window smaller so that it squishes, the blue goes away
[05:38] <jmhodges> er, it did scale properly i mean, but the blue is still there
[05:38] <ubitux> can you show a before/after?
[05:39] <jmhodges> yeah
[05:39] <ubitux> remove the pix_fmt too
[05:39] <ubitux> -pix_fmt option
[05:40] <jmhodges> ahh
[05:42] <jmhodges> ubitux: original http://cl.ly/image/0x3Q3I3i3F36 first frame of output: http://cl.ly/image/1W393y0y2d3y
[05:43] <ubitux> yeah, that was expected
[05:43] <jmhodges> ubitux: oh?
[05:43] <ubitux> i'm not sure you'll get better results
[05:43] <jmhodges> ah :/
[05:43] <ubitux> you can try removing the format=rgb8, but it will likely get uglier
[05:44] <jmhodges> is this because of just how gif works, or how ffmpeg implements the gif encoding?
[05:44] <jmhodges> i think it's the latter?
[05:44] <jmhodges> something something system palette?
[05:44] <ubitux> i don't think our encoder is making a custom palette
[05:44] <ubitux> so it might pick the closer color in a large range of colors
[05:45] <jmhodges> ah, hm
[05:45] <ubitux> you need to check
[05:45] <ubitux> well anyway, our gif encoder should be improved in many ways
[05:45] <jmhodges> oh, check what?
[05:45] <ubitux> patches are very welcome
[05:45] <ubitux> check the code :p
[05:45] <jmhodges> haha ahhh
[05:45] <jmhodges> no worries
[05:46] Action: ubitux would be very happy to see improvements on the gif encoder
[05:48] Action: jmhodges nods
[05:48] <jmhodges> would be nice
[05:49] <ubitux> it could be pretty useful to make previews for videos
[05:49] <jmhodges> i rigged it up by extracting out jpegs and using convert to turn them into one gif
[05:49] <jmhodges> it would!
[05:49] <ubitux> especially given our scene detection filter
[05:49] <jmhodges> oh, hm
[05:53] <FedUs> I used the "extract frames at given intervals and then process them with ImageMagick/gifsicle" workaround a few times, maybe it could be connected to the scene detection; code/filters for gif did not raise much interest in a few years, but I could be horribly wrong
[05:54] <FedUs> s/to/with/
[05:54] <Aziroshin> ubitux: Thanks for the heads up, I'll look into that. :)
[05:55] <Aziroshin> The CPU is fine, though.
[05:59] <ubitux> http://lucy.pkh.me/bbb-scenes.gif
[05:59] <ubitux> this is generated with: ./ffmpeg -i ~/samples/big_buck_bunny_1080p_h264.mov -vf 'select=gt(scene\,.4),scale=-1:120,format=rgb8,format=rgb24' -vsync vfr -frames:v 10 -r 1 -y bbb-scenes.gif
[05:59] <ubitux> so 10 scene frames
[06:01] <ubitux> jmhodges ^
[06:01] <jmhodges> hunh, yeah
[06:02] <ubitux> but the gif is quite huge
[06:02] <ubitux> (almost 300K)
[06:07] <jmhodges> yeah
[06:15] <Keshl> Kay, so, I'm doing   ffmpeg -r 60 -i (pngstream) -vcodec mpeg4 -qscale 2 BLAH.avi -- It works fine, but there's a point in the video where the bitrate gets too high that, even from a ramdisk, the system can't handle it. How's I fix? D:
[06:21] <p4plus2> Keshl: I think you can do something like "-maxrate 400k"  (or whatever rate you need) to cap it
[06:21] <Keshl> Shiny, oÉo.
[06:22] <Keshl> And can I do a two-pass thing somehow so it doens't suddenly get fuzzy at that point, oÉo? At least not /as/ fuzzy?
[06:23] <p4plus2> Probably but I am not sure how that setup would be done -- I would test to see if the change in quality is even noticeable before worrying about that
[06:23] <Keshl> oÉo.
[06:24] <p4plus2> set maxrate to as close to a high sane value as you can and work from there I suppose
[06:37] <Keshl> It goes over maxrate. D:
[06:37] <Keshl> ...Unless 400k means 400,000, not 400 kb.
[06:39] <p4plus2> no it should mean 400kb
[06:42] <Keshl> Whelp, it goes over 40000kb when I set it to 40000k D:
[06:42] <Keshl> And still when I do 400k D:
[06:42] <Keshl> And and and D: Why does the software not acc -- .....
[06:42] <Keshl> ......
[06:42] <Keshl> One moment. >w>
[06:47] <Keshl> Kay, got it. Somehow I was using libx264 with -crf and didn't notice.
[06:48] <Keshl> Switched to mpeg4 like I shoulda been using, all good now. >w>
[09:52] <edgy> Hi, ffmpeg -i shows me 720x576 [SAR 16:15 DAR 4:3], I guess this is a bug because 16:15 is PAR not SAR, no?
[09:57] <JEEB> edgy, both ways of calling it are valid
[09:57] <JEEB> sample aspect ratio and pixel aspect ratio
[09:58] <JEEB> video format specs love to use the former as far as I know
[09:58] <edgy> JEEB: according to https://en.wikipedia.org/wiki/Pixel_aspect_ratio SAR is something different
[09:58] <edgy> SAR is storage aspect ratio
[09:59] <edgy> SAR is the resolution
[09:59] <JEEB> uhh
[09:59] <JEEB> I've never seen that being used like that
[10:00] <JEEB> from the H.264 specification f.ex.
[10:00] <JEEB> "The sample aspect ratio (SAR)"
[10:01] <JEEB> the "sample aspect ratio" wording is much more used for the aspect ratio of a single sample (usually pixel)
[10:03] <edgy> yes seems contradicting terminology ;)
[10:04] <edgy> JEEB: but since PAR is agreed upon, I guess it's much better to use and I remember I have seen ffmpeg says PAR before
[10:04] <edgy> though I may confused avcodec with ffmpeg not sure
[10:07] <JEEB> x264 uses --sar, and as far as I know ffmpeg/avcodec has so far always used SAR. Also the SAR mentioned on that wikipedia page is not in any way or form agreed upon
[10:07] <JEEB> also it limits itself to pixels
[10:07] <JEEB> (because of its name)
[10:08] <JEEB> in many cases in unofficial texts the wording "pixel aspect ratio" is used, and this is not incorrect
[10:09] <JEEB> but when you are showing the data of a video stream in most cases that data is called SAR, and let's just say that everyone agrees that the data is called that way in most if not all cases
[10:09] <JEEB> also, this is the /first/ time I have /ever/ seen the word "storage aspect ratio"
[10:12] <JEEB> basically I'm not against people calling what is signalled by the SAR as PAR, and so forth in their guides/whatever. Those are two terms that are generally agreed upon to mostly mean the same thing within digital video
[10:12] <JEEB> That storage aspect ratio tho is something I'm not even sure who uses it :P
[10:12] <JEEB> not to mention that it falls over official specification terms with many MPEG-based formats which makes it even more evil
[10:16] <edgy> JEEB: thanks for the clarification
[10:21] <JEEB> storage aspect ratio probably just is something some wikipedia editor made up to not use what most people use to note that part of the aspect ratio equation (resolution)
[10:21] <JEEB> and was quite unfortunately labeled
[10:21] <JEEB> lol
[10:36] <edgy> ;)
[14:14] <knoch> hello, I have some video clips with wrong timestamps and duration ( mplayer shows an incorrect duration but ffmpeg is correct ), how can I fix this ?
[14:16] <an3k> raw streams?
[14:21] <an3k> knoch?
[14:21] <knoch> mpeg2video
[14:21] <knoch> in TS container
[14:22] <knoch> on one video, ffmpeg is also wrong about the duration
[14:22] <Zandman26> Hi all, I'm trying to split my microphone audio from my PCM audio and Desktop caputre using map command but I only manage to get errors:/ my commands and terminal output can be found here: http://pastebin.com/VMp5Mg12
[14:31] <burek> knoch, try remuxing your videos
[14:48] <knoch> burek: je teste ça
[15:00] <knoch> burek: remuxing fixed the timestamps, it also fixed the duration but only for VLC and ffmpeg, not for mplayer
[15:01] <burek> well, im happy it works for ffmpeg at least :)
[15:02] <Macey> does the mpegtsenc deal with PCR on audio only?
[15:03] <Macey> i'm wanting to mp2 up a file and put it in a mpegts container and the PCR on the audio
[15:03] <Macey> it informs me that the ecoder don't produce pts
[15:04] <Macey> [mpegts @ 006f2d20] Encoder did not produce proper pts, making some up.
[15:13] <Macey> from the source, it looks like its a huge hack... can't find the mp2 encoder though
[15:31] <an3k> knoch: sry, didn't noticed you replied. with what tool have you remuxed the videos?
[15:50] <an3k> can ffmpeg read avs files and use DXVA / CUDA for indexing a H.264 stream packed into a BDAV m2ts container and for converting it to H.264?
[16:10] <Zandman26> Is it even possible with ffmpeg to use stream mapping together with x11grab? Can't find any examples of it:/
[16:12] <ubitux> yes
[16:12] <ubitux> just like any input you can map it
[16:14] <ubitux> Zandman26: x11grab is producing only one stream, and you try to map two of them
[16:14] <ubitux> x11grab = video only, it's not producing an audio stream
[16:20] <Zandman26> Ok. My goal is to capture video of the screen together with game sound in one file and a second file for my microphone audio, So how do I map that?
[16:20] <an3k> ubitux: nice. then i can compare my DualCPU Xeon Server with my GTX580 PC ... let's see what's faster :p
[16:21] <an3k> 8 Xeon Cores vs. one GTX580 ... anyone interested in the results?
[16:24] <ubitux> Zandman26: you need an input audio stream, look at http://ffmpeg.org/ffmpeg-devices.html
[16:35] <Zandman26> Ok, thx. Will try to figure it out
[17:23] <Zandman26> Still having no luck getting ffmpeg to split channels my see : http://pastebin.com/sx1bfKPK
[17:27] <ubitux> move the map after declaring all your inputs
[17:27] <ubitux> -map is an output option
[17:27] <ubitux> not an input option
[17:28] <Zandman26> Ok thought it needed to be declared early to map correctly.
[17:32] <siamba> hi
[17:33] <siamba> I want to encode movie for android
[17:33] <siamba> "saw"
[17:34] <siamba> 8 years old, but in fine quality(1920x1072)
[17:35] <siamba> my android 2.3.7 has 800x480 screen resolution
[17:35] <sacarasc> Resolution and quality aren't the same thing.
[17:37] <siamba> sacarasc: sorry, I mean it is h264 BDRemux
[17:38] <siamba> and 8.4G
[17:39] Action: siamba has 1.0.1 ffmpeg
[17:42] <siamba> oh, it's a "Saw [Director's Cut] (2004) BDRip 1080p-Rip"
[17:43] <ubitux> the film sux, don't waste your time
[17:44] <ubitux> btw, what's the question?
[17:44] <sacarasc> Encoding for Android.
[17:44] <sacarasc> I forget how to set baseline...
[17:44] <siamba> using cyanogenmod's android 2.3.7
[17:45] <siamba> trying with "ffmpeg -i ~/saw.mkv -s 320x240 -vcodec libx264 -acodec aac -strict experimental -ac 2 -r 15 -ab 44100 -aspect 16:9 ~/saw.mp4"
[17:45] <ubitux> -profile:v baseline ?
[17:45] <siamba> have fps=160
[17:46] <sacarasc> That's how fast you're encoding.
[17:46] <ubitux> -ab 44100
[17:46] <ubitux> wat?
[17:46] <ubitux> i think you want -ar
[17:47] <siamba> ;(
[17:48] <siamba> should I add -profile:v baseline or replace something with it?
[17:49] <siamba> sacarasc: i7 3930k has ~600% utilization encoding
[17:50] <sacarasc> ffmpeg -i input.mkv -vf scale=480:360 -profile:v baseline -c:v libx264 -b:v 500k -c:a aac -strict experimental -ac 2 -ar 44100 output.mp4
[17:50] <sacarasc> That is going by the http://developer.android.com/guide/appendix/media-formats.html
[17:52] <siamba> sacarasc: thank you
[17:52] <sacarasc> Oh...
[17:52] <siamba> ubitux: I've heard a discussion about this movie today at work
[17:53] <sacarasc> ffmpeg -i input.mkv -vf scale=480:360 -profile:v baseline -c:v libx264 -b:v 500k -c:a aac -strict experimental -b:a 128k -ac 2 -ar 44100 output.mp4
[17:53] <sacarasc> Missed the audio bitrate.
[17:56] <siamba> I've heard there was CUDA or OpenCL ffmpeg fork, is that true?
[17:58] <siamba> fps is ~190 now
[18:00] <siamba> <_<
[18:00] Action: siamba has dual GPU desktop
[18:01] <siamba> GTX 580 + 680
[18:02] <ubitux> won't help you for h264 encoding
[18:02] <siamba> ubitux: just interested if it is going to be implemented or not
[18:03] <ubitux> i don't think so
[18:03] <ubitux> in your current cmd line, maybe having a gpu scaling could help, but i'm not sure it could be faster
[18:04] <ubitux> we have an opencl filter though, if the project has some gpu optim
[18:04] <ubitux> wait no
[18:04] <siamba> time is already 01:19:43.82
[18:04] <ubitux> we don't& what am i smoking.
[18:04] <siamba> not so long to wait
[18:04] <ubitux> opencv*.
[18:04] <siamba> oh, opencv
[18:05] <ubitux> not sure if that supports scaling, i doubt it
[18:07] <siamba> encoded, "time ffmpeg ..." : real    12m51.300s, user    93m18.390s
[18:07] <siamba> not bad for "frame=147856 fps=192 q=32582.0 Lsize=  474930kB time=01:42:46.83 bitrate= 630.9kbits/s"
[18:12] <siamba> I guess it could be at least couple times faster with nVidia
[18:29] <ubitux> i doubt it
[19:13] <brx_> sometimes when I trim an mp4 the video is trimmed shorter than the audio, I think it freezes on the last keyframe, so the sound plays for half a second or so longer than the video
[19:14] <JEEB> brx_, there's a setting that ends the encoding when the shortest track finishes
[19:15] <JEEB> http://ffmpeg.org/ffmpeg.html ctrl+F and start writing "shorte"
[19:15] <brx_> roger that, thanks man
[19:15] <brx_> JEEB, i love you and i want to have your babies
[19:15] <JEEB> haha
[19:25] <brx_> JEEB, it doesnt quite work as expected
[19:25] <brx_> still a pause in the vid at the end
[19:26] <brx_> ffmpeg -y -ss 00:00:0.0 -t 00:00:1.0 -i 2013-01-01.mp4 -shortest -acodec copy -vcodec copy out.mp4
[19:52] <burek> brx_, it migh be that video and audio frames are not exactly aligned
[19:52] <burek> so when you cut the output at 1 second
[19:52] <burek> that non-alignment happens and someone will remain longer than the other
[19:53] <brx_> burek, so is there anyway to extract a second from an mp4 and guarentee the audio and vid will align?
[19:53] <brx_> im working on a simple security app type thing
[19:53] <burek> for example, you might have 15 video frames and 13 audio frames in original media, bud if you cut it at half, you'll most probably have 7 video and 6 audio frames, which might not align perfectly
[19:53] <brx_> it will record 10 seconds then allow the user to extract a second from it
[19:53] <brx_> ahh yes i see
[19:53] <burek> i guess you'll either have to re-encode
[19:53] <burek> or play around with -t
[19:54] <burek> to hit some spot
[19:54] <burek> where frames do align perfectly again
[19:54] <brx_> ok
[19:54] <burek> the rule of a thumb (for me) is if you work with very short sequences, like couple of seconds, always do re-encoding
[19:55] <burek> <brx_> burek, so is there anyway to extract a second from an mp4 and guarentee the audio and vid will align?
[19:55] <burek> just decode a/v and nothing else
[19:55] <burek> i.e. save it as uncompressed
[19:55] <burek> it will be aligned perfectly
[19:56] <brx_> thats the thing, in android i cant do that
[19:56] <brx_> there is a MediaRecoder class, theres no raw/uncompressed option
[19:56] <burek> why not do it server-side
[19:56] <brx_> its a mobile app
[19:56] <burek> and let android just access the result
[19:56] <burek> exactly
[19:56] <brx_> i want it all to be localized
[19:57] <burek> then you have a problem :)
[19:57] <brx_> :<
[19:57] <burek> android and arm devices were generally designed to consume less power in order to be able to be deployed without cooling fans
[19:57] <burek> and used in mobile devices, etc
[19:58] <burek> as media players
[19:58] <burek> not as encoders... so..
[19:58] <brx_> but surely dumping the raw/uncompressed video is less intensive than encoding an mp4?
[19:58] <brx_> which is what im being forced to do
[19:58] <burek> just decoding is far less cpu intensive than encoding
[19:59] <brx_> im confused
[19:59] <brx_> im capturing a vid with the camera, it is producing an mp4
[19:59] <burek> when you decode, you just follow the stream of bytes and transform it using known algorigthm into output buffer, but with encoding, you need to predict a lot of stuff to be sure you are always encoding the best possible way
[20:00] <burek> why not using something like
[20:00] <brx_> exactly, so its encoding an mp4, which would be more intensive than just dumping the uncompressed sreams
[20:00] <burek> ffmpeg -y -ss 00:00:0.0 -i 2013-01-01.mp4 -shortest -c:v libx264 -crf 0 -c:a copy -t 00:00:1.0 out.mp4
[20:01] <burek> or even -c:a pcm_s16le -ar 44100 -ac 2
[20:01] <burek> crf 0 will make it lossless
[20:01] <burek> but still re-encode
[20:01] <brx_> ill try that now
[20:02] <brx_> theres still apause at the end with that command
[20:02] <burek> -c:a copy or the other
[20:03] <brx_> actually thats a tad better
[20:03] <burek> try the other one
[20:03] <brx_> k trying now
[20:03] <burek> ffmpeg -y -ss 00:00:0.0 -i 2013-01-01.mp4 -shortest -c:v libx264 -crf 0 -c:a pcm_s16le -ar 44100 -ac 2 -t 00:00:1.0 out.mp4
[20:03] <brx_> i got: Could not write header for output file #0 (incorrect codec parameters ?)
[20:03] <brx_> the file is 0 bytes
[20:04] <brx_> trying the "or even" one now...
[20:05] <brx_> k
[20:07] <brx_> http://pastebin.com/ehpa7Hkt
[20:08] <burek> [mp4 @ 0x1c424e0] track 1: could not find tag, codec not currently supported in container
[20:08] <burek> ffmpeg -y -ss 00:00:0.0 -i 2013-01-01.mp4 -shortest -c:v libx264 -crf 0 -c:a aac -strict experimental -ar 8000 -ac 1 -t 00:00:1.0 out.mp4
[20:10] <brx_> http://pastebin.com/CqfttzzH
[20:11] <burek> ffmpeg -y -ss 00:00:0.0 -i 2013-01-01.mp4 -shortest -c:v libx264 -crf 0 -c:a aac -strict experimental -ar 8000 -ac 1 -ab 12k -t 00:00:1.0 out.mp4
[20:11] <brx_> that seems to have worked
[20:12] <brx_> theres a very slight glitch at the end of the vid, however the source vid has a full half second pause in vid at the end
[20:12] <brx_> ill try a few more sample source vids
[20:12] <burek> hmmm
[20:12] <burek> ffmpeg -y -i 2013-01-01.mp4 -ss 00:00:0.0 -shortest -c:v libx264 -crf 0 -c:a aac -strict experimental -ar 8000 -ac 1 -ab 12k -t 00:00:1.0 out.mp4
[20:12] <burek> put -ss after -i
[20:13] <burek> btw
[20:13] <burek> why -ss if it's 0
[20:14] <brx_> i think thats just about done it
[20:14] <brx_> will that -strict experiment ever give any vizarre results?
[20:14] <brx_> bizarre*
[20:15] <burek> -strict is just to remind you that aac is still experim. codec
[20:15] <burek> so that you dont complain too much :)
[20:15] <brx_> ok cool :)
[20:15] <brx_> im gonna experiment on a few more files seewhat the results are like
[20:15] <brx_> thanks for all the help I really appreciate it
[20:16] <burek> :beer: :)
[20:16] <brx_> while(xmas){if(beer){happy++}else{goToBeerShop()}}
[20:18] <burek> try{ assert(beer); } exception { doBrandy(); }
[22:44] <Keshl> [16:44] [Whois] Akira^^_ has been idle for 7 days, 7 hours, 7 minutes, and 4 seconds. Epic OÉO
[22:44] <Keshl> Well, 3 seconds early of being super-epic.
[22:44] <Keshl> That is all. Carry on. oÉo.
[23:23] <Penyulap> ffmpeg -i %04d.png -vcodec mjpeg -sameq "output.avi"
[23:23] <Penyulap> can someone tell me how to change this into something that outputs an OGG video rather than the avi video ? it is getting input from a series of images, it works now, but I want to change the output type
[23:24] <ubitux> select a video codec for ogg, then replace avi with ogg
[23:29] <saste> Penyulap, and don't use -sameq which doesn't work anymore
[23:29] <Penyulap> oh
[23:29] <saste> well it never properly worked
[23:30] <Penyulap> Unknown encoder 'ogg'
[23:30] <saste> container is not the same as encoder
[23:31] <Penyulap> it worked enough so that i could make avi's  (sorry, I know nothing about ffmpeg except I can't live without it :D )
[23:31] <saste> just specify out.ogg and ffmpeg will pick a suitable encoder
[23:31] <Penyulap> can you suggest a command i can try ?
[23:31] <Penyulap> I'm an artist, I don't know much about this kind of thing
[23:31] <Penyulap> but I know about CGI :D
[23:32] <saste> ffmpeg -i INPUT output.ogg
[23:32] <saste> -q:v to set the quality
[23:32] <saste> -q:v 3 may be good, the lower the better (and higher the size)
[23:32] <Penyulap> oh it's working, i cut down more stuff so it was more like that but with the whole input part left
[23:33] <saste> check ffmpeg-codecs and the libtheora docs for more details
[23:34] <Penyulap> can i add in a mp3 file so i have sound too, how can i do that ?
[23:35] <Penyulap> it's ok it worked already
[23:35] <saste> Penyulap, ffmpeg -i INVIDEO -i INAUDIO out.ogg should work
[23:35] <Penyulap> I am kindof getting an idea of it, though I try to leave room in my head for all the things i have to learn
[23:35] <Penyulap> ah
[23:35] <Penyulap> it like gets the idea of what is what by itself, smart :)
[23:37] <llogan> saste: docs for libtheora shows usage of -global_quality:v. is this preferred over qscale:v for this encoer?
[23:37] <saste> llogan, ffmpeg-codecs should be tool agnostic
[23:39] <saste> -q:v is an ffmpeg shortcut for -flags +qscale -global_quality X
[23:39] <llogan> i see
[23:43] <Penyulap> how can i tell it to make it Higher def (it worked but it is so blurry and jerky too actually)
[23:44] <Penyulap> ok
[23:44] <Penyulap> pb?
[23:45] <Penyulap> oh pb !
[23:45] <llogan> it was a shortcut to summon the fflogger bot
[23:47] <Penyulap> Input Stream #0.0 frame size changed to 960x540, rgb24
[23:47] <Penyulap> Input Stream #0.0 frame size changed to 960x540, bgrabitrate= 408.9kbits/s
[23:47] <llogan> that means nothing without context.
[23:47] <Penyulap> oh
[23:47] <Penyulap>     encoder         : Lavf52.64.2
[23:47] <Penyulap>     Stream #0.0: Video: libtheora, yuv420p, 960x540, q=2-31, 200 kb/s, 25 tbn, 25 tbc
[23:47] <Penyulap>     Stream #0.1: Audio: flac, 44100 Hz, mono, s16, 64 kb/s
[23:48] <llogan> show your ffmpeg command and the complete console output that you get after you enter your ffmpeg command. use pastebin.com, and don't paste it into the channel
[23:48] <Penyulap> ok
[00:00] --- Thu Dec 27 2012