[Ffmpeg-devel-irc] ffmpeg.log.20130105

[burek]

[00:00] <sesam> Mavrik, same here but in avid
[00:00] <Mavrik> t4nk637: audio? x264?
[00:00] <llogan> t4nk637: mp4 is a container format that can contain various video and audio formats
[00:00] <t4nk637> no audio
[00:00] <Mavrik> sesam: sadly, haven't found a decently cheap way of doing Premiere renders on a server :\
[00:01] <Mavrik> t4nk637: so, you'll need "image2, mp4" muxers, "libx264" encoder, "mjpeg" decoder
[00:01] <sesam> mavrik: isn't there a distrubuted render app that you can install on other computers? i know after effects does
[00:01] <llogan> Mavrik: i tend to use a frameserver from premiere to ffmpeg because i hate AME
[00:01] <Mavrik> t4nk637: disable ffserver, ffplay, ffprobe
[00:01] <Mavrik> t4nk637: and all devices and hwaccels
[00:02] <Mavrik> sesam: aftereffects yes, premiere no :\
[00:02] <llogan> --disable-everything --enable-foo for an extreme method
[00:02] <Mavrik> t4nk637: what llogan said, start with "--disable-everything" and enable just those muxers, encoder and decoder
[00:02] <t4nk637> ...yup, that sounds great...  I'd seen the --disable-everything --enable-goodstuffonly ...
[00:02] <llogan> there is --enable-small if size is a concern, but i've never really looked at that
[00:03] <t4nk637> awesome, the list you showed helps
[00:03] <Mavrik> --enable-small is ok if you need a small binary
[00:03] <Mavrik> but that will come at performance cost though
[00:03] <Mavrik> I suggest you leave that out if you're on a slow device and can afford a little flash.
[00:03] <t4nk637> processing will happen at night, so performance can be sacrificed.... cool
[00:03] <llogan> ...since you mentioned "embedded"
[00:04] <t4nk637> Kirkwood
[00:04] <t4nk637> = sortaslow
[00:04] <Mavrik> llogan: frameserver?
[00:04] <luc4> Mavrik: I see in the AVStream structure of my AVFormatContext there is a AVCodec, but that is not initialized. Do you know why I need to init that manually?
[00:04] <llogan> Mavrik: https://ffmpeg.org/trac/ffmpeg/wiki/FFmpegPremierePro
[00:04] <llogan> might work for avid too
[00:05] <sesam> llogan: that i will look up....
[00:05] <Mavrik> luc4: hmm, you should have AVStream's in the AVFormatContext and then AVCodecs attached to each stream
[00:05] <Mavrik> luc4: and you have to open the codec for each of those streams
[00:05] <Mavrik> llogan: oh, that looks interesting
[00:06] <Mavrik> thanks :)
[00:06] <luc4> Mavrik: oh, so maybe I just have to open? In the sample codes I see I also have to allocate so I was asking myself why.
[00:06] <Mavrik> hmm
[00:06] <t4nk637> thnx #@ llogan & Mavrik  ; byebye
[00:06] <llogan> Mavrik: feel free to edit/update the wiki page if something is not clear/correct
[00:06] <Mavrik> will do
[00:07] <llogan> i'm not usually the editor at my office, so i don't actually end up using premiere that much
[00:07] <llogan> thankfully
[00:07] <sesam> llogan: =)
[00:07] <klaxa> btw Mavrik, just checked with my logs, it was you i talked about the stuff with splitting h264 streams and all the shit :P
[00:07] <llogan> which reminds me i need to go get some more b-roll before it gets too dark
[00:08] <sesam> llogan: good luck
[00:08] <llogan> thanks. you too.
[00:08] <klaxa> and in the end i just took the I frames output by mkvtoolnix which seem to be IDR frames after all
[00:08] <klaxa> and i pass muxed files
[00:09] <llogan> Mavrik: oh, i noticed that one project file no longer shows the render queue...possibly related to debugmode frameserver...possibly not, so try with a backup first.
[00:09] <llogan> whatever the cause, it's terribly annoying.
[00:10] <llogan> i'll blame adobe
[00:10] <Mavrik> ^^
[00:10] <Mavrik> blaming Adobe is rarely wrong
[00:10] <llogan> "render queue" being the "export video" or whatever it's called.
[00:12] <llogan> i've also seen some audio/video sync issues, but i just ended up muxing them separately. i didn't get to investigate much.
[00:13] Action: llogan leaves for real
[01:20] <verdoc> say I've got "videofile.mp4" and it's encoded exactly how I want... what's the best way to create a "loop-version.mp4" that is the same video just looped X amount of times?
[01:22] <Mavrik> probably by a tool like mp4box
[01:22] <Mavrik> to concat it several times
[01:22] <verdoc> I tried that actually
[01:23] <verdoc> that output file didn't seem to have an updated length
[01:23] <seece> are -slice-max-size and -keyint options deprecated?
[01:24] <Mavrik> verdoc: hmm, possible
[01:24] <Mavrik> mp4 format isn't really built for easy concatenation :\
[01:25] <verdoc> I'm beginning to pick up on that
[01:25] <verdoc> ;)
[01:25] <verdoc> I tried this approach - http://revision3.com/forum/showthread.php?t=37586
[01:25] <Mavrik> the ugly and annoying hack is to remux to .ts, concat that and remux to mp4 :)
[01:25] <Mavrik> but mp4box should really be able to do that :\
[01:26] <klaxa> shouldn't ffmpeg itself have a loop functionality?
[01:26] <Mavrik> verdoc: changing to mpg will reencode the file
[01:26] <klaxa> it's referenced in the online manpage, but nowhere actually mentioned
[01:26] <Mavrik> it was deprecated or something
[01:26] <Mavrik> iirc
[01:27] <verdoc> do you have sample of the remux hack?
[01:27] <klaxa> -loop_input and -loop_output are marked as deprecated and one should use -loop instead, but -loop isn't explained anywhere
[01:27] <verdoc> yeah, I saw that in the man page too klaxa
[01:27] <verdoc> it looks like it's used to create the single image youtube music "video" songs
[01:27] <Mavrik> verdoc: ffmpeg -i <test.mp4> -codec copy -bsf:v h264_mp4toannexb output.ts
[01:27] <Mavrik> then concat files in any way
[01:28] <Mavrik> and do ffmpeg -i <test.ts> -codec copy output.mp4
[01:29] <verdoc> it didn't like that
[01:29] <Mavrik> no? :)
[01:29] <verdoc> I got dep warnings
[01:29] <verdoc> and it said -codec wasn't valid
[01:30] <Mavrik> just how old is your ffmpeg? :)
[01:31] <verdoc> ffmpeg 0.8.4-4:0.8.4-0ubuntu0.12.04.1
[01:31] <Mavrik> that would explain the errors
[01:31] <verdoc> *** THIS PROGRAM IS DEPRECATED ***
[01:31] <verdoc> This program is only provided for compatibility and will be removed in a future release. Please use avconv instead.
[01:31] <Mavrik> use "-vcodec copy -acodec copy" instead of "-codec copy"
[01:31] <JEEB> switch to the command 'avconv' in that case
[01:31] <JEEB> that's the updated binary with libav
[01:31] <JEEB> :)
[01:31] <verdoc> same syntax?
[01:31] <JEEB> yes
[01:32] <JEEB> (mostly)
[01:33] <JEEB> the biggest difference I know is how you set libx264 the more fine-grained specific settings (-x264opts vs specific -settings)
[01:33] <verdoc> `avconv -i <test.mp4> -codec copy -bsf:v h264_mp4toannexb output.ts` seems to be doing something
[01:38] <llogan> verdoc: -loop only works with image inputs and formats that support looping such as animated GIF
[01:46] <verdoc> that worked
[01:46] <verdoc> thanks guys
[02:26] <luc4> Hi! Can I somehow decode a 6 channels AAC stream to 2 channels?
[02:39] <pods_guest_974> hi what is the best way to detect a anime/porn movie ...by colors, histogram of colour?? ffmpeg can help to do this?
[02:46] <Tjoppen> those are two separate questions
[03:20] <limpc> probably means hentai, and theres no way to do that
[03:21] <limpc> if its not in the file info, you cant detect by a frame
[03:21] <limpc> cant tell the difference between Legend of the Overfiend, and Naruto :P
[03:40] <p4plus2> Is there any way to specify which jack ports for ffmpeg to record from?  Ideally selecting three channels (left, right, and a mono) and having them mix down to two channels(left+mono, right+mono)
[03:41] <p4plus2> (left and right generally being system or program sounds, mono being from a microphone)
[03:44] <p4plus2> actually I may have a guess as to how to make it work -- i'll try it out
[03:46] <p4plus2> hmm nothing
[03:47] <p4plus2> ok so it seems "ffmpeg -f jack -i name_here" creates a port for hooking ffmpeg up to
[03:48] <p4plus2> but is there a way to redirect sounds there through ffmpeg?
[03:52] <p4plus2> maybe I'll try jack capture and just merge the two sources later
[03:52] <p4plus2> last question then would be is there any way to pause a recording?
[03:54] <p4plus2> actually maybe STOP/CONT will work -- if that is even a safe action for ffmpeg...
[04:00] <raket> what about ffmpeg -f alsa -i plughw? or is your 1) soundcard uttercrap 2) you run a soundserver ? :-)
[04:03] <p4plus2> well I'd rather use jack so I can capture explicit programs and not some others
[04:05] <p4plus2> alsa, at least to my knowledge, doesn't have a simple way of allowing that  (maybe something more complicated with loopback devices could be done, but that requires all of my programs to allow me to easily use said device)
[04:05] <raket> it works, if you have a real soundcard with more then 1 channel
[04:06] <p4plus2> My sound card is (for now anyways) just a on board sound card with very limited options
[04:06] <raket> but it's very uncommon with such soundcards and no one seems to have documentet but it does in fact exist and does work. i've used this long long time ago
[04:07] <p4plus2> Ah, I'll try and dig something up in a bit -- gotta go eat dinner real quick though
[04:07] <p4plus2> Thanks for the info
[04:07] <raket> hope jack works! :-)
[04:07] <raket> (i have the same problem as you describe)
[04:09] <raket> anyone here? i need to know a video codec where i can force 8bit video
[04:29] <luc4> Hi! I'm trying to decode a AAC audio stream with S32 sample format, 6 channels 48000 samples per second, but this is what I get: ftp://luc4.homeip.net/output.pcm. I used this as a reference: http://ffmpeg.org/doxygen/trunk/decoding__encoding_8c_source.html#l00238. Any idea why that is the result?
[04:37] <DJJeff> can ffmpeg do mp4 to swf ?
[04:38] <DJJeff> I see mp4 to flv using winff
[04:39] <luc4> Seems possible.
[04:44] <DJJeff> well I got my video into flv just need it in swf heh
[04:44] <DJJeff> im impressed :-)
[04:46] <p4plus2> DJJeff: via google I found "ffmpeg -i input.flv -acodec copy -vcodec copy output.swf"
[04:47] <DJJeff> yes that worked.... lost quite a bit of quality
[04:48] <DJJeff> my mp4 file was 9MB the swf was only 1MB
[04:50] <DJJeff> just tried from mp4 to swf its complaining about sample rate or something
[04:55] <p4plus2> Hmm
[04:55] <p4plus2> maybe play with the settings where you converted to flv?
[04:59] <DJJeff> might work
[04:59] <DJJeff> going from mp4 to swf the filesize is actually getting large 30MB+
[05:00] <p4plus2> How was the quality?
[05:04] <DJJeff> dunno it ummm gave me warnings and I pressed ctrl c
[05:44] <DJJeff> seems like pushing the bitrate up to about 3000 in winff seems to maintain quality
[05:44] <DJJeff> from mp4 to flv
[05:44] <DJJeff> resulting in a 7MB flv from my 9MB mp4
[05:45] <DJJeff> and about a 3.5MB swf
[05:46] <DJJeff> alot of quality is lost from flv to swf :-(
[05:47] <DJJeff> oh I see the bitrate dropped to 1500
[06:02] <DJJeff> cant seem to get the bitrate from flv to swf to go higher then 1500
[06:02] <DJJeff> seems to ignore the -minrate all together
[11:38] <misterno> hey would i get better quality if i use source with certain resolution and then compress that or get source with higher resolution then lower it to same resolution as in previous example and then compress?
[11:39] <misterno> i'm talking about a game that can be resized
[11:39] <misterno> so in both example i would end up with equal resolution
[11:41] <Mavrik> misterno: resizing will be done before encoding in any case
[11:41] <Mavrik> so unless you use a noticably worse resizing algorithm there will be no difference
[11:41] <misterno> wouldn't game do better resizing itself?
[11:42] <misterno> likea  3d game
[11:42] <misterno> but your point is good so no detail will be taken from higher resolution source, it would be resized first
[11:43] <misterno> i'm just looking into how to record games in best possible quality using little amount of space
[11:45] <misterno> i see some streams on youtube at 320 or 480 that have quite good quality, then i see streams that are way too blurry at those resolutions
[11:46] <Mavrik> that's usually just differences in set bitrate of video
[11:46] <Mavrik> the resizing algorithm doesn't really make a noticable difference in content like games
[11:46] <misterno> maybe codec used too?
[11:49] <Mavrik> usually yes, but I doubt people use anything other than h.264 nowdays
[11:50] <misterno> i use afterburner to record video and it doesnt have h.264
[11:50] <misterno> it offers uncompressed, RTV1, MJPG and VFW compression
[11:52] <Fjorgynn> lol
[11:52] <misterno> do you know better app than afterburner?
[11:53] <BtbN> DXtory + vlc
[11:53] <BtbN> at least on windows
[11:53] <misterno> hmm never heard of dxtory i will check it out thanks
[11:54] <BtbN> it mapps screen capture to a virtual DirectShow device which VLC can use as source
[11:55] <Mavrik> misterno: hm, yea, uncompressed is usually the best for quality then
[11:55] <Mavrik> misterno: MJPG with really really relaxed settings is ok as well
[11:55] <misterno> oh wait i was wrong
[11:56] <misterno> because im new to this i didnt know i can setup VFW and then choose one of many codecs one which is x.264vfw
[11:56] <misterno> so it does support it in afterburner
[11:57] <misterno> uncompressed is unfortunatelly too taxing for my system which doesnt allow me to play game properly
[11:57] <misterno> so i have to use some kind of compression
[11:57] <misterno> also when i record 30 fps it just doesnt look nice
[11:57] <misterno> i prefer at 60 when it looks smooth
[11:59] <misterno> is there difference between H264 and h264?
[12:00] <sacarasc> The shift key.
[12:02] <misterno> why i have both offered in x264vfw ffmpeg configuration under "fourcc" ?
[12:02] <misterno> same goes for x264 and X264
[12:16] <Yulth> Hi everyone!
[12:17] <Yulth> Is there any way to store the mp3 id3 data into an mp4 container with HE-AAC audio and, furthermore, with accurate estimation of duration?
[12:18] <Yulth> I mean, the command on this pastebin, avoiding the lines reported also on the pastebin :) http://pastebin.com/8cGVqmWi
[12:19] <Yulth> any ideas?
[12:19] <durandal_1707> you tried and it did not work?
[12:21] <relaxed> Yulth: Isn't '-f adts' raw aac, and not mp4?
[12:22] <durandal_1707> and why mp3 is in output at all, please paste full and uncut output
[12:22] <Yulth> durandal_1707: it produces an mp4 (HE-AAC) audio file with no metadata from the mp3 file
[12:22] <Yulth> ok, let me 1 minute! :D
[12:22] <relaxed> remove -f adts
[12:23] <Yulth> relaxed: ffmpeg doesn't work if '-f adts' is removed, because doesn't recognize the format of the output audio...
[12:26] <Yulth> the pastebin with full output
[12:26] <Yulth> http://pastebin.com/ApWVFQAB
[12:28] <durandal_1707> remove -f adts
[12:30] <Yulth> so, here the error: http://pastebin.com/a3yUSUfM
[12:31] <durandal_1707> that is cover, if you do not care about cover, add -vn
[12:33] <Yulth> let me try it
[12:35] <Yulth> wooa!! it worked!!
[12:35] <Yulth> :)
[12:35] <Yulth> and any solution to solve the warning message: "Estimating duration from bitrate, this may be inaccurate" ? :D
[12:37] <durandal_1707> that is from demuxer level and does not affect muxing in any way
[12:38] <cytrinox> hi
[12:38] <durandal_1707> in other words mp4 duration should be exact to what is duration of decoded mp3 stream
[12:39] <Yulth> durandal_1707: but when it is converted and streamed in real time, Cannot produce errors when decoding by clients?
[12:40] <durandal_1707> no, decoding should be fine
[12:40] <cytrinox> simple question: is it possible to set the display width/height for matroska output files? something like -s hd1080 --display-size-foo=hd720?
[12:41] <Yulth> durandal_1707: ok, and it Cannot produces seek errors, isn't it?
[12:42] <durandal_1707> cytrinox: size of video or display size which define pixel real size?
[12:43] <durandal_1707> Yulth: what protocol for streaming, not all support seeking
[12:43] <Yulth> html
[12:43] <cytrinox> durandal_1707: I mean display size
[12:44] <Yulth> some audio players supports hatml header "Accept-Ranges: bytes"
[12:44] <Yulth> *html
[12:44] <cytrinox> |   + Display width: 1920
[12:44] <cytrinox> |   + Display height: 1040
[12:44] <cytrinox> these are properties of the output video track
[12:44] <durandal_1707> Yulth: html is not protocol for streaming
[12:45] <durandal_1707> cytrinox: and what is size of video?
[12:45] <cytrinox> the question is if it is possible to instruct ffmpeg to encode the video as something link hd1080, but set the display size to something different
[12:45] <cytrinox> |  + Video track
[12:45] <cytrinox> |   + Pixel width: 1920
[12:45] <cytrinox> |   + Pixel height: 1040
[12:46] <cytrinox> s/link/like
[12:46] <Yulth> durandal_1707: What alternative do you suggest for?
[12:46] <Yulth> rtsp? rtmp?
[12:46] <durandal_1707> if Display width/height is part of container than it should be supported
[12:46] <durandal_1707> Yulth: http, https, and so on, but not html
[12:47] <Yulth> durandal_1707: ok
[12:47] <luc4> Hi! I'm using the samples in the documentation to decode an aac stream, but I get something like this: ftp://luc4.homeip.net/output.pcm. I copy-pasted the code in the documentation and this is the result. Anyone who has any advice?
[12:47] <cytrinox> durandal_1707: yes, display width/height properties are part of matroska. but I have not found any switch how to set these information manually
[12:47] <Yulth> durandal_1707: really really grateful for your help. Thanks once again! See you! :D
[12:49] <durandal_1707> cytrinox: this is just metadata or there is special handling in mkv?
[12:51] <cytrinox> uh, I don't known how this is internally handled by matroska, but it can simply modified via mkvpropedit video.mkv --edit track:1 --set display-height=x
[12:51] <cytrinox> but my intention was to automatically produce the correct file with ffmpeg
[12:52] <durandal_1707> you mean the size was different from one in video stream?
[12:52] <cytrinox> eh?
[12:53] <durandal_1707> mkv have one video stream or several video streams, no?
[12:53] <cytrinox> I mean that ffmpeg always write the same display size into mkv as the real pixel size
[12:54] <cytrinox> and I want to have different (smaller) displays sizes
[12:54] <cytrinox> afaik other containers supports different display sizes, too
[12:55] <durandal_1707> paste ffmpeg output
[12:57] <luc4> Anyone who knows whether there is some known issue in AAC decoder maybe?
[12:58] <cytrinox> durandal_1707: what output? It's a theoretical question how to set different values for the display width/height properties directly with ffmpeg
[13:00] <Mavrik> luc4: shouldn't be, make sure you have correct samplerate and channel number (+ layout) set when playing (and creating) the pcm file
[13:02] <luc4> Mavrik: samplerate should be correct, 48000, there are 6 channels here, and this is correct, but for the moment I'm trying to write the PCM samples only from the first. I used the decoding_encoding.c sample. Maybe that is not ok for this kind of stream?
[13:03] <durandal_1707> cytrinox: it is hardcoded, you could open bug report to make it configurable
[13:04] <luc4> Mavrik: to be clear, I simply copy-pasted the decode sample method from that sample code and used on my aac stream.
[13:04] <Mavrik> that should work perfectly
[13:04] <Mavrik> I can't listen to your stream sadly (too slow download)
[13:05] <Mavrik> but usually the problem is with wrong samplerate and/or channel layout
[13:05] <luc4> Mavrik: I see this fwrite(decoded_frame->data[0], 1, data_size, outfile);
[13:06] <luc4> Mavrik: shouldn't it be fwrite(decoded_frame->data[0], 1, decoded_frame->linesize[0], outfile); ?
[13:08] <Mavrik> not in case of audio it seems
[13:08] <Mavrik> the call "av_samples_get_buffer_size" tells you how many bytes the passed number of samples at passed parrameters take
[13:09] <luc4> Mavrik: it is my understanding data_size is the size of the entire frame, but the frame should contain all the channels.
[13:09] <luc4> Mavrik: anyway not working either. I can send you the decoded sample somehow, it is less than 1MB.
[13:09] <Mavrik> data_size is the size of nb_samples samples when taking into account number of channels and sample format
[13:10] <luc4> Mavrik: oh ok, so that is not the issue...
[13:11] <luc4> Mavrik: that is the start of bug buck bunny from an aac stream. I can hear that audio more or less, but it is all messed up.
[13:12] <durandal_1707> luc4: what is not working, what you get as output?
[13:12] <luc4> durandal_1707: I placed a sample here ftp://luc4.homeip.net/output.pcm. I didn't know how to provide it.
[13:13] <durandal_1707> aac decoder is changed to use planar sample fmt output
[13:13] <durandal_1707> so if you use only data[0] you will get only 1st channel and it will be all messed up if you thing it is > 2 ch file
[13:13] <durandal_1707> *think
[13:13] <luc4> durandal_1707: oh...
[13:14] <durandal_1707> you need to manualy interleave
[13:14] <luc4> durandal_1707: oh... I'm sorry, it was my understand each channel was ok by itself...
[13:15] <luc4> durandal_1707: so I can't take a 6 channels aac and play only 1 channel?
[13:16] <durandal_1707> luc4: each channels is in separate plane for aac
[13:17] <durandal_1707> so using only data[0] you should have only one channel if you are using enough recent library
[13:18] <luc4> durandal_1707: yes, and forgive my ignorance :-) if I played only samples coming from data[0] as mono audio, should I expect a "correct" audio or not?
[13:19] <durandal_1707> you should expect correct audio if decoder is using planar sample format
[13:20] <luc4> durandal_1707: to be cleared, I used that sample code from the documentation to output to file only data[0]. Then I played that with aplay, specifying 48000Hz, 1 channel. Audio is like the one I provided.
[13:20] <durandal_1707> aplay?
[13:21] <luc4> durandal_1707: yes, I see that sample format for ffmpeg is float_planar: AV_SAMPLE_FMT_FLTP.
[13:21] <luc4> durandal_1707: yes, alsa player.
[13:22] <luc4> durandal_1707: the only thing that I'm not sure is that aplay is playing the samples as signed 32 bit little endian. ffmpeg instead outputs float.
[13:22] <durandal_1707> that is the problem
[13:23] <luc4> durandal_1707: ah ok! But do you have any advice on how to convert float to signed 32?
[13:24] <luc4> durandal_1707: my complete code should output the audio using OpenMAX on embedded devices. But I don't seem to be able to play float with that, so I suppose I have to convert or ask ffmpeg to output something different. Is it possible?
[13:25] <durandal_1707> yes, ffmpeg use aresample filter which use libswresample
[13:26] <luc4> durandal_1707: I just tried to play the file using aplay output_ok.pcm -f FLOAT_LE -r 48000 -c 1. More or less same messed up audio.
[13:28] <luc4> durandal_1707: it seems that with float only the volume is different.
[13:29] <durandal_1707> how many channels original aac have?
[13:29] <luc4> durandal_1707: 6
[13:31] <luc4> durandal_1707: it is simply a stream extracted from the big buck bunny video sample using ffmpeg.
[13:32] <durandal_1707> how? / paste log output
[13:32] <luc4> durandal_1707: of the extraction?
[13:33] <luc4> durandal_1707: the aac stream seems correct. I can play it with other players and it is perfect.
[13:35] <luc4> durandal_1707: this is the extraction output http://pastebin.com/BCZQdsZ7
[13:37] <durandal_1707> than you are either removing some samples or writing uninitialized data
[13:38] <durandal_1707> perhaps you need to paste code you use
[13:39] <luc4> durandal_1707: I'm surely doing something wrong, this is the second day I use ffmpeg :-) but I just copy-pasted the decoding_encoding.c function...
[13:40] <luc4> durandal_1707: that sample was the output of the code from decoding_encoding.c.
[13:42] <durandal_1707> luc4: make sure that you write exactly what you get
[13:43] <durandal_1707> if you write only one channel samples than data_size should be divided by 6
[13:44] <luc4> durandal_1707: that is what I asked before to Mavrik...
[13:45] <p131> Hi everyone !
[13:45] <p131> I'm playing with concat filter with no luck :(
[13:46] <p131> ffmpeg -y  -i 1264  -i "Sans nom-1.avi" -s 320x240 -filter_complex  "[0:0],scale=320:240[first];[1:0],scale=320:240[second];[first][second]concat=n=2:v=1 [v]" -s 320x240  -map "[v]" out.mpg
[13:46] <p131> => no such filter ''
[13:46] <p131> I know it's a simple syntax issue, but i cannot get into it
[13:46] <durandal_1707> ffmpeg -filters list concat?
[13:47] <p131> yepo
[13:47] <p131> The issue came from the moment i tried to chain the scale filter
[13:47] <p131> (cause concat need files to be at the same res)
[13:49] <p131> I must have made a dummy syntax error, but i dont understand where
[13:49] <smooki> hi
[13:50] <smooki> can ffmpeg do dts connect ?
[13:50] <smooki> or dolby digital live ?
[13:54] <luc4> durandal_1707: seems to work! But shouldn't decoded_frame->linesize[0] be equal to data_size/6.0?
[13:55] <durandal_1707> it is not?
[13:56] <luc4> durandal_1707: let me check the values precisely.
[13:57] <smooki> anyone know maybe if any software can do DTS Connect please ?
[13:59] <luc4> durandal_1707: linesize[0] is data_size/3.0.
[14:00] <luc4> durandal_1707: also, since that function is writing down only the first channel, shouldn't it be changed to data_size/6.0?
[14:01] <durandal_1707> luc4: example is broken/ needs some love to work with recent changes....
[14:03] <luc4> durandal_1707: ok, but what about linesize[0]? Shouldn't it be equal to data_size/6?
[14:04] <durandal_1707> luc4: what if you use len/6 instead?
[14:04] <luc4> durandal_1707: do you mean data_size/6?
[14:04] <durandal_1707> no
[14:05] <luc4> durandal_1707: ah the value returned by the decode function.
[14:06] <luc4> durandal_1707: len == 1272.
[14:08] <luc4> durandal_1707: I will use data_size/channels, but from the documentation it is my understanding linesize[0] should be equal to data_size/6.
[14:18] <Fjorgynn> can the mp4 container contain subtitles?
[14:18] <JEEB> 3gpp timed text, yes
[14:19] <JEEB> or well, mpeg-4 timed text which is IIRC based on 3gpp timed text
[14:36] <hub2> Hi,I made a program in C++ which encodes files to AAC but I got error "The encoder 'aac' is experimental but experimental codecs are not enabled, add '-strict -2' if you want to use it." and I don't know what to do. Have you any ideas how to solve it?
[14:37] <Fjorgynn> libvo_aacenc
[14:37] <Fjorgynn> hub2: that's the encoder
[14:38] <Fjorgynn> codec whatever
[14:41] <hub2> Fjorgynn:Yes I know that's about encoder but i don't know how to "add '-strict -2' " or solve it any other way?
[14:42] <Fjorgynn> if you use -c:a libvo_aacenc you don't have to add some strict whatever
[14:42] <Fjorgynn> why not add -strict -2 in the line and try otherwise?
[14:43] <durandal_1707> it is C++ program
[14:43] <Fjorgynn> so it's not using ffmpeg?
[14:45] <hub2> in what line i should add it,it's a C++ program?
[14:50] <Mavrik> hub2: if I understand correctly, you have a C++ encoder which calls ffmpeg libraries directly?
[14:52] <hub2> Yes
[14:53] <Mavrik> hub2: basically, the catch is AAC encoder in ffmpeg is still pretty experimental, that's why it's usually better to use fdk_aac, libvo_aacenc or libfaac
[14:53] <Mavrik> hub2: the "proper" way to fix your issue is to recompile ffmpeg libraries with libfdk_aac enabled and then recompile the C++ program
[14:54] <Mavrik> this will make it choose the external encoder
[14:54] <hub2> Thanks i'll try that
[14:55] <Fjorgynn> is libfaac better than libvo_aacenc, and how can I get it in Windows? :(
[14:56] <Mavrik> hub2: the other option is to set your AVCodecContext->strict_std_compliance to FF_COMPLIANCE_EXPERIMENTAL
[14:56] <Mavrik> Fjorgynn: no, it's not really
[14:56] <Mavrik> Fjorgynn: use libfdk_aac if you can
[14:56] <Mavrik> it's by far the best one
[14:57] <Fjorgynn> aha
[14:58] <Fjorgynn> yay, managed to get the size down to 5-6 MB /minute.
[14:59] <Fjorgynn> Mavrik: okay, I have no idea what the difference is really
[14:59] <Fjorgynn> between libfdk_aac and libvo_aacenc
[15:00] <Mavrik> Fjorgynn: HE-AACv2 support, multi-channel support and the encoder itself it noticably better (better audio quality)
[15:01] <Fjorgynn> okay
[15:21] <Fjorgynn> seems like changing the sample rate from 48 to 44.1 doesn't make a lot of difference
[15:27] <luc4> Hi! I'm currently playing a stream by reading the file (single stream) and passing data to the decoder. Can I simply extract encoded data so that I can pass that to a hardware decoder?
[15:55] <JEEB> <Fjorgynn> if you use -c:a libvo_aacenc you don't have to add some strict whatever <- switch to fdk-aac if you only ca
[15:55] <JEEB> *can
[15:55] <JEEB> vo_aacenc isn't really better than the internal aac encoder
[15:56] <JEEB> and yeah, I see Mavrik already noted that fdk-aac is preferred if you're going to use some library
[16:01] <Fjorgynn> JEEB: so how can I do it on Windows?
[16:01] <Fjorgynn> if I don't want to compile anything
[16:02] <Fjorgynn> got the latest ffmpeg
[16:03] <espr3ss0> hi again
[16:03] <Fjorgynn> Mavrik:
[16:14] <fatpony> i have a general question about two pass x264 encoding
[16:15] <fatpony> should i include all options in the first pass?
[16:15] <fatpony> like ref and subme?
[16:16] <JEEB> Fjorgynn, fdk-aac seems to be not LGPL or GPL compliant license-wise so you'd indeed have to compile
[16:16] <JEEB> (patent clause in the license)
[16:18] <Fjorgynn> I also has besweet
[16:18] <espr3ss0> Q: does anyone know if this works ? ... "http://letsneverdie.net/blog/?p=75"
[16:18] <Fjorgynn> fatpony: what about 2 pass?
[16:18] <JEEB> fatpony, uhh why are you setting specific x264 settings?
[16:19] <JEEB> presets should be fine in like 90%+ of all cases
[16:19] <fatpony> JEEB: i want a specific filesize for the output
[16:19] <Fjorgynn> JEEB: and how do I compile it on Windows?
[16:19] <JEEB> ...that is setting an average bit rate
[16:19] <JEEB> those settings have NOTHING TO DO WITH FILE SIZE
[16:19] <JEEB> (which you mentioned)
[16:20] <fatpony> let's say i'm using two passes anyway
[16:20] <fatpony> should i set the subme and so on parameters on the first pass?
[16:20] <JEEB> well yes, that is just -preset yourpreset -pass 1 -b:v yourbitrate and then -preset yourpreset -pass 2 -b:v yourbitrate
[16:21] <fatpony> okay
[16:21] <JEEB> use preset and -pass 1 and -pass 2
[16:21] <JEEB> that way you get automated "fast first pass settings" with the first pass
[16:21] <JEEB> while keeping everything else the same
[16:21] <JEEB> :V
[16:21] <fatpony> i've seen examples where both presets and ref/subme are specified
[16:21] <JEEB> http://mewiki.project357.com/wiki/X264_Settings#preset
[16:21] <JEEB> list of presets for libx264
[16:22] <Fjorgynn> seems like I found something
[16:22] <JEEB> fatpony, some people want to use specific values with specific presets, that's their choice -- I'm pretty sure you have no reason to blindly follow
[16:22] <JEEB> setting ref can help with ffmpeg because libx264 doesn't limit reference frames to a certain level if you are setting it
[16:23] <JEEB> so if you do -level 40 it will set level 4.0 in the header, but not actually limit the reference frame count
[16:23] <JEEB> that's the only case I really think where setting some setting manually really matters
[16:23] <JEEB> but if you are not setting a level
[16:23] <JEEB> x264 will auto-decide based on your general settings
[16:24] <fatpony> yeah i changed ref to respect level 4
[16:24] <fatpony> .1
[16:24] <JEEB> Fjorgynn, grab the latest git, autoreconf -fiv, see the configure script generated and then ./configure --enable-static --disable-shared --prefix=/your/prefix
[16:24] <JEEB> yes, that makes sense
[16:24] <JEEB> too bad libx264 doesn't do that automatically
[16:24] <JEEB> only x264cli does
[16:24] <fatpony> where are these presets stored?
[16:24] <JEEB> inside libx264
[16:24] <JEEB> you access them with -preset
[16:24] <fatpony> ah ok
[16:24] <fatpony> so it's ok to use -preset slow -x264opts ref=4?
[16:25] <JEEB> yes
[16:25] <JEEB> preset sets the defaults
[16:25] <JEEB> then you limit ref
[16:25] <JEEB> just make sure you set the level then too I guess
[16:25] <fatpony> oh yeah it's in my commandline
[16:26] <JEEB> so yeah, it'd be something like -c:v libx264 -pass 1 -preset slow -level 41 -x264opts ref=4 and -c:v libx264 -pass 2 -preset slow -level 41 -x264opts ref=4
[16:27] <JEEB> the -pass 1 sets "fast first pass settings"
[16:27] <fatpony> oh you set level through ffmpeg?
[16:27] <fatpony> i added it to x264opts
[16:28] <JEEB> yes, there's a level setting in ffmpeg as far as I know
[16:28] <fatpony> i'll use that then
[16:28] <JEEB> if I'm incorrect do note
[16:29] <Fjorgynn> JEEB: easy for you to say
[16:29] <espr3ss0> Q: anyone know what FAM is ?
[16:29] <JEEB> Fjorgynn, possibly
[16:30] <JEEB> you could use qtaac or neroaacenc via command line if you want
[16:30] <JEEB> *qaac was it?
[16:30] <JEEB> first needs QuickTime's certain components installed
[16:30] <JEEB> and then mux the AAC stream into whatever you want
[16:31] <fatpony> well thanks for your help JEEB, i'll let it encode now :)
[16:56] <Fjorgynn> JEEB: can't do that
[17:03] <JEEB> well then you don't encode to AAC :D
[17:03] <JEEB> because all of your alternatives atm are crap
[17:04] <Fjorgynn> JEEB: then why are there no aac alternatives already compiled?
[17:04] <Fjorgynn> "home made"
[17:04] <JEEB> because most of the good alternatives available are either not GPL or LGPL compatible, so you can't distribute builds with them
[17:05] <JEEB> and the internal one lacks an active developer
[17:05] <JEEB> if you want to sponsor the internal one's development, feel free. You'll just have to find an active maintainer for it
[17:07] <JEEB> right now if you can't compile fdk-aac the easiest way is either to use QT's encoder via qaac or use nero's encoder via neroaacenc
[17:07] <JEEB> you can pipe raw audio into both as far as I know
[17:09] <Scorcerer> hello
[17:14] <Scorcerer> is there a linux tool to softcode *.srt subtitles into an .avi container ?
[17:15] <Scorcerer> i can use aviaddxsubs on windows, but it takes time, GUI, and can't be automated easily
[17:16] <Fjorgynn> installed msys but yeah
[17:17] <JEEB> that isn't really supported widely so I recommend you'd use a better container if you need to have subtitles as data in there Scorcerer
[17:17] <Fjorgynn> read something about the ass filter
[17:17] <JEEB> that's hardsubbing
[17:17] <Fjorgynn> aha
[17:17] <JEEB> he wanted to put the srt into the container
[17:17] <JEEB> separate track
[17:17] <Scorcerer> i only need it for PS3
[17:18] <JEEB> yeah, the PS3 seems to support it in some weird way
[17:18] <Scorcerer> roght now in a directory i have an .avi movie, .txt subs and .rst subs
[17:18] <Scorcerer> but thse separate files aren't streamed over DLNA
[17:19] <Scorcerer> and it doesn't support .mkv
[17:19] <Fjorgynn> mp4?
[17:19] <JEEB> mp4 with timed text would work, but I have no idea if his solution works with that
[17:19] <Scorcerer> it doesn't matter what codec i use, if its .mkv container it won't play
[17:20] <Scorcerer> but i'm not 100% sure now
[17:20] <JEEB> or if PS3 won't support timed text
[17:20] <JEEB> also ffmpeg cannot convert other formats into timed text so you'd have to use GPAC (mp4box) for it
[17:20] <JEEB> (other subtitle formats)
[17:20] <Scorcerer> it's ok for now, as my ps3 is in repair, but i have another question
[17:21] <Scorcerer> i have bunch of movies encoded in divx, and to play them on RaspberryPI i need to transode them into x264
[17:22] <Scorcerer> one episode weights ~350MB
[17:22] <Scorcerer> is there easy way to transcode that ?
[17:22] <Fjorgynn> git clone git://source.ffmpeg.org/ffmpeg.git ffmpeg
[17:22] <Fjorgynn> can't get this to work
[17:22] <JEEB> you need git basically
[17:22] <JEEB> you should be able to grab msysgit from its project
[17:22] <JEEB> I usually grab the 7z
[17:22] <JEEB> and extract it, and then add the /bin of it to the end of the PATH
[17:22] <JEEB> (of msys shell)
[17:23] <JEEB> (or you might just grab the installer and it'll handle global PATH addition for it)
[17:23] <JEEB> http://code.google.com/p/msysgit/downloads/list
[17:23] <Scorcerer> i have ffmpeg 0.10.6 installed on my media server (which runs on linux)
[17:23] <JEEB> the ones that don't say msysgit and netinstall are what you want
[17:23] <Scorcerer> isn't it enough ?
[17:24] <JEEB> Scorcerer, if it has libx264 linked in, yes
[17:24] <JEEB> also the format is H.264
[17:24] <Scorcerer> yes, it does
[17:24] <JEEB> the encoder is called x264
[17:24] <Scorcerer> i also read i have to use 2-pass encoding if i want to getacceptable quality
[17:25] <Scorcerer> but do i need to specify all options like HV size and others ?
[17:25] <JEEB> eh
[17:25] <JEEB> if you want good quality you EITHER use crf (quality-based) OR 2pass average bit rate
[17:25] <JEEB> crf gives you somewhat stable set quality
[17:26] <JEEB> average bit rate gives you an average bit rate over the whole video
[17:26] <JEEB> both work from the same algorithm
[17:26] <JEEB> so if you don't need specific size
[17:26] <JEEB> you set crf
[17:26] <JEEB> and one pass needed
[17:26] <JEEB> also what do you mean with "HV size and others?"
[17:27] <Scorcerer> like, horizontal/vertical size of input/output video stream
[17:27] <JEEB> output is the same as input by default
[17:28] <JEEB> unless you need to change that, there is no need to scale anything
[17:29] <Scorcerer> ah, so it's ok to jus use x264 and set crf to medium, for, like, one episode, and if i'll find quality acceptable, run it on all others ?
[17:29] <Scorcerer> wondering how long will it take to transcode one episode :)
[17:29] <JEEB> just use -ss and -t
[17:29] <JEEB> to seek to a certain part and then encode X seconds
[17:30] <JEEB> 25-30 seconds of material should be fine for a test
[17:30] <JEEB> do first like this: ffmpeg -i input -ss <Amount to seek> -t <Time to encode> -c:v libx264 -crf 23 -an out.mp4
[17:31] <JEEB> then if that looks good, push the crf value higher
[17:31] <JEEB> if it looks bad, push it lower
[17:31] <JEEB> by one or so
[17:31] <JEEB> then just keep doing that until you have the highest crf value that still looks good to you
[17:31] <klaxa> -ss is pretty slow though :S
[17:31] <JEEB> oh right
[17:31] <JEEB> you could set that before -i
[17:31] <JEEB> it's faster that way
[17:31] <JEEB> not exact, but faster
[17:32] <klaxa> oh... if only i new that earlier :P
[17:32] <JEEB> after that you can start testing -preset values
[17:32] <JEEB> http://mewiki.project357.com/wiki/X264_Settings#preset
[17:32] <JEEB> medium is the default
[17:32] <JEEB> find the slowest that is still fast enough for you
[17:33] <JEEB> (everything but ultrafast and placebo should be "OK" to use)
[17:33] <JEEB> (placebo and ultrafast are usually for benchmarks)
[17:33] <Scorcerer> the truth is, i would want to keep quality as-is, but get similar or smaller filesize - for now, each episode has 350MB
[17:33] <JEEB> then crf mode is the best for that
[17:34] <JEEB> because it lets you select the highest crf value that still looks good
[17:34] <JEEB> that way you can compress exactly as much as you can while keeping the quality you want
[17:34] <Scorcerer> would like to try crf, as it's easier to batch-transcode many files, and i thuink it would be faster than 2-pass
[17:34] <Scorcerer> ah, ok
[17:34] <JEEB> average bit rate mode is made for exactly that -- when file size is more important than the quality
[17:34] <Scorcerer> on to the tests then, thanks for help :)
[17:35] <JEEB> crf is when you want to set the "quality"
[17:35] <JEEB> and then the presets will help you tweak it
[17:35] <JEEB> (compression level)
[17:35] <JEEB> for the PS3 you also would end up using -level 41 and limit reference frames most probably
[17:35] <JEEB> unfortunately libx264 doesn't do that automatically
[17:36] <JEEB> the x264 command line encoder sets reference frame counts according to the resolution :<
[17:36] <JEEB> (and level you set)
[17:37] <Scorcerer> i think that's too much for me now, gotta do some trial-and error tests and then i'll be ready to learn some more :)
[17:37] <JEEB> well, for now
[17:37] <JEEB> just do what I noted up in the log
[17:38] <JEEB> encode 25-30 seconds of content
[17:38] <JEEB> with various crf values set
[17:38] <JEEB> look at the results, see if it looks good enough
[17:38] <JEEB> adjust crf value until you've found the highest that looks good enough for you
[17:41] <Fjorgynn> If I download this source http://ffmpeg.zeranoe.com/builds/
[17:42] <JEEB> IIRC that hasn't been updated for quite a while
[17:42] <JEEB> oh, it has
[17:42] <JEEB> someone poked zeranoe
[17:42] <Fjorgynn> 2013-01-04?
[17:42] <JEEB> although I would still get the ffmpeg sources from git
[17:42] <Fjorgynn> ;)
[17:43] <Fjorgynn> JEEB: but then I must download all stuff also, not only x264 and aac, right?
[17:43] <JEEB> just install/setup msysgit one way or another and be done with it :P
[17:43] <JEEB> Fjorgynn, you'd have to in any other case too afaik
[17:43] <JEEB> uhh, what
[17:43] <JEEB> what exactly do you want to do with ffmpeg
[17:44] <JEEB> if you just want to encode H.264 and AAC (and quite a few other formats), you only need yasm -> x264 -> fdk-aac -> ffmpeg
[17:44] <Fjorgynn> convert stuff to H.264 with aac.
[17:44] <JEEB> that's all you need then
[17:44] <JEEB> yasm for assembler stuff
[17:44] <JEEB> x264 for H.264 encoding
[17:44] <JEEB> fdk-aac for AAC audio encoding
[17:44] <JEEB> ffmpeg for... you know, ffmpeg :P
[17:44] <Fjorgynn> yasm... hmm
[17:44] <JEEB> yasm is simple to grab from the main site of theirs
[17:45] <JEEB> you just need the yasm.exe to be in PATH
[17:45] <JEEB> so you can call it from msys with 'yasm'
[17:45] <Fjorgynn> never heard of yasm before
[17:45] <JEEB> it's a rewrite of nasm
[17:45] <Fjorgynn> never heard of that either
[17:45] <JEEB> basically, assembly code to binaries
[17:45] <JEEB> both x264 and ffmpeg EXTENSIVELY use assembly code
[17:45] <JEEB> so you will want it
[17:45] <JEEB> unless you wan to be slow
[17:45] <Fjorgynn> and how do I include it in the ffmpeg?
[17:46] <JEEB> just have yasm available from the msys prompt
[17:46] <JEEB> put yasm.exe into the msys PATH
[17:46] <JEEB> that's it
[17:47] <JEEB> it's available from yasm's site
[17:47] <JEEB> http://yasm.tortall.net/Download.html
[17:47] <JEEB> <JEEB> you just need the yasm.exe to be in PATH
[17:47] <JEEB> <JEEB> so you can call it from msys with 'yasm'
[17:47] <Fjorgynn> and how do I put it in the PATH? I don't know something like that. Should be more documentation about it
[17:48] <JEEB> ...
[17:48] <JEEB> PATH is not something msys-specific
[17:48] <JEEB> neither *nix specific
[17:49] <JEEB> try 'echo $PATH'
[17:49] <Fjorgynn> http://ffmpeg.org/trac/ffmpeg/wiki/MingwCompilationGuide
[17:49] <JEEB> in msys's shell
[17:49] <JEEB> it will show the PATH of your current msys shell
[17:49] <JEEB> PATH is where the system (in this case msys shell) looks for binaries to run
[17:50] <Fjorgynn> there are already yasm in the mingw msys stuff I am using :o
[17:50] <JEEB> well that's great then
[17:50] <JEEB> check yasm --version
[17:50] <JEEB> to check the version
[17:51] <Fjorgynn> 1.2.0
[17:51] <JEEB> great
[17:51] <JEEB> up-to-date too
[17:51] <Fjorgynn> http://pastebin.com/EdEsMUuS
[17:51] <Fjorgynn> using this script thingy
[17:51] <Fjorgynn> http://ffmpeg.org/trac/ffmpeg/wiki/MingwCompilationGuide
[17:51] <Fjorgynn> and this
[17:52] <Fjorgynn> or yeah not much that last but
[17:52] <JEEB> ugh
[17:52] <JEEB> forgets about --prefix
[17:52] <Fjorgynn> googled it :/
[17:52] <JEEB> also memalign hack is no longer neede
[17:52] <JEEB> *needed
[17:52] <JEEB> w32threads are default now
[17:52] <JEEB> can you pastebin what you get when you run 'ls -alh /mingw' ?
[17:53] <JEEB> not paste on the channel, but pastebin and link here
[17:53] <Fjorgynn> http://pastebin.ca/2299627
[17:54] <JEEB> I will guess that if you do gcc -v the default gcc will be i686-w64-mingw32, right?
[17:55] <Scorcerer> what can i do when audio is going out-of sync with video (i'm using c:a copy right now)?
[17:55] <Fjorgynn> something like that JEEB
[17:55] <JEEB> ok
[17:56] <Fjorgynn> Target: i686-w64-mingw32
[17:56] <JEEB> Fjorgynn, thus your default --prefix should be --prefix=/mingw/i686-w64ming32
[17:56] <JEEB> argh
[17:56] <JEEB> forgot the second dash
[17:56] <JEEB> --prefix=/mingw/i686-w64-mingw32
[17:57] <JEEB> now first of all grab x264 from git with git clone git://git.videolan.org/x264.git and it should create a folder called 'x264' into the current folder and start putting stuff into it
[17:57] <JEEB> that is, if you now finally have git around :P
[17:58] <Fjorgynn> libiconv-2.dll is missing
[17:58] <JEEB> you failed to set up git properly
[17:58] <Fjorgynn> will look after it
[17:59] <JEEB> basically when you get x264's sources around, you do ./configure --enable-static --prefix=/mingw/i686-w64-mingw32 (see the i686-w64-mingw32 subfolder under /mingw , that is the folder for mingw stuff for that architecture, we will follow that)
[18:00] <JEEB> oh right, you should also disable cli
[18:00] <JEEB> ./configure --disable-cli --enable-static --prefix=/mingw/i686-w64-mingw32
[18:00] <JEEB> should configure x264 without the command line encoder, with a static library and install it into i686-w64-mingw32
[18:00] <JEEB> then you do 'make'
[18:01] <JEEB> and then if that finishes correctly, 'make install' to install it into the prefix
[18:02] <Fjorgynn> because now I have several folders
[18:02] <Fjorgynn> directories
[18:13] <Fjorgynn> so
[18:22] <Fjorgynn>  JEEB ERROR: libx264 not found
[18:24] <Fjorgynn> will try again
[18:25] <Scorcerer> JEEB: transcoding didn't stop after full movie length, it it problem with file or i messed something up?
[18:25] <Scorcerer> i'll paste cmd, sec
[18:26] <Fjorgynn> ah, seems to be git that was the problem there
[18:26] <Scorcerer> ffmpeg -i movie.avi -c:v libx264 -preset slow -crf 22 -c:a copy test_22.avi
[18:27] <Fjorgynn> Scorcerer: what's the problem with that?
[18:28] <Scorcerer> quality was nice, filesize was really small (need to try with crf 18) audio was totally out-of-sync and it didn't stop transcoding after 1:36:xx which was end of movie...
[18:29] <Fjorgynn> :/
[18:29] <Scorcerer> i used mplayer to check how long it is, and before i pushed q it was 1:52:21 :/
[18:34] <Fjorgynn> should work
[18:34] <Fjorgynn> I've got a problem compiling ffmpeg myself
[18:36] <Scorcerer> i don't think it's a problem witch ffmpeg itself, rather badly encoded file :/
[18:36] <JEEB> Fjorgynn, if you used the correct prefix with everything, then --extra-cflags="-I/mingw/i686-w64-mingw32" and --extra-ldflags="-L/mingw-/i686-w64-mingw32" should work for x264
[18:37] <JEEB> although if you have pkg-config around it'd be simpler to just set PKG_CONFIG_PATH=/mingw/i686-w64-mingw32/lib/pkgconfig before ./configure on the line (PKG_CONFIG_PATH=... ./configure --options)
[18:39] <Fjorgynn> now I can configure ffmpeg atleast
[18:39] <DJJeff> anyone know how to ffmpeg from .flv to .swf and keep same quality ?
[18:40] <Fjorgynn> lol flash
[18:40] <DJJeff> ?
[18:40] <Fjorgynn> ffmpeg -i input.flv output.swf
[18:40] <Fjorgynn> doesn't it work?
[18:41] <DJJeff> it does but I loose alot of quality
[18:41] <DJJeff> my 7MB flv becomes 1MB swf
[18:41] <DJJeff> the flv bitrate was 5000 the swf cant go higher then 1500
[18:42] <DJJeff> not sure why
[18:42] <Fjorgynn> $ make
[18:42] <Fjorgynn> common.mak:138: *** missing separator.  Stop.
[18:42] <Fjorgynn> JEEB: :(
[18:42] <JEEB> hmm
[18:42] <JEEB> inb4 CRLF makefiles
[18:42] <JEEB> turn off autocrlf in git and re-clone
[18:42] <Fjorgynn> what?
[18:43] <Fjorgynn> it is ffmpeg
[18:43] <JEEB> yes
[18:43] <JEEB> but that sounds like a usual problem conceived by having the sources cloned with Windows CRLF end lines
[18:43] <JEEB> and not just LF *nix end lines
[18:43] <JEEB> I'm currently trying to grab the github help page regarding it from my browser's history
[18:43] <JEEB> https://help.github.com/articles/dealing-with-line-endings#platform-windows
[18:44] <JEEB> there
[18:44] <JEEB> you want to set autocrlf to 'false'
[18:44] <JEEB> not true
[18:44] <Fjorgynn> how?
[18:44] <JEEB> open the link I provided?
[18:44] <JEEB> it's the first command line noted in there
[18:44] <JEEB> as far as I can see
[18:45] <Fjorgynn> git config --global core.autocrlf false
[18:45] <JEEB> yes
[18:45] <JEEB> Then see the "Re-normalizing a repo"
[18:45] <JEEB> part
[18:45] <JEEB> so you don't have to re-clone
[18:46] <Fjorgynn> okay
[18:47] <JEEB> except for the commit part of course
[18:47] <JEEB> you don't need to do that
[18:47] <JEEB> you just need to do the first two commands there
[18:47] <Fjorgynn> okay
[18:48] <Fjorgynn> now I'll configure and make again
[18:48] <Fjorgynn> DJJeff: can't you set the -b:v 5M ?
[18:50] <Fjorgynn> And we have a winner!
[18:53] <Scorcerer> what does that mean: Switching to NI mode, due to poor interleaving ?
[18:53] <JEEB> uh-oh
[18:53] <JEEB> can you pastebin the whole log of that file onto a pastebin provider, and then link it here?
[18:54] <Scorcerer> sure, one sec
[18:55] <Scorcerer> http://pastebin.ca/2299655
[18:56] <JEEB> welp
[18:56] <Scorcerer> that is for other file, audio desync persists here, dunno about going beyond EOF
[18:57] <JEEB> it isn't going beyond EOF
[18:57] <JEEB> that file is just...
[18:57] <JEEB> you might want to pass it through virtualdub or so in the stream copy mode for both video and audio
[18:57] <Scorcerer> so it's broken ok :)
[18:58] <JEEB> the first hint is
[18:58] <JEEB> [avi @ 0x634fb0] scale/rate is 0/0 which is invalid. (This file has been generated by broken software.)
[18:58] <JEEB> then you've set the audio bit rate way too low
[18:58] <JEEB> remember that bit rate is set in bits in ffmpeg
[18:58] <JEEB> not in kilobits
[18:59] <JEEB> you were setting it to 192 while I think you meant 192k
[18:59] <Scorcerer> didn't know that, thanks
[18:59] <Scorcerer> yup
[18:59] <Scorcerer> anyway
[18:59] <Scorcerer> do i have to use virtualdub, or is there linux-cmd tool to do this ?
[18:59] <JEEB> anyways, you might want to pass such files through something like vdub or possibly nandub depending on how broken they are
[19:00] <JEEB> well, 1) vdub has command line interface and 2) it works under wine
[19:00] <JEEB> so you can use it under linux
[19:00] <JEEB> if you want
[19:00] <hotwings> is ffmpeg git down at the moment?
[19:00] <Scorcerer> don't have windows here, except for VM
[19:00] <Scorcerer> ah, ok
[19:00] <Scorcerer> i'll try, thanks
[19:00] <JEEB> see the virtualdub documentation basically I guess
[19:00] <JEEB> hotwings, seems up for me http://git.videolan.org/?p=ffmpeg.git
[19:00] <JEEB> at least the http part of it
[19:01] <Fjorgynn> ...
[19:01] <hotwings> getting fatal: unable to connect to source.ffmpeg.org:
[19:01] <hotwings> with git clone
[19:01] <JEEB> hotwings, that domain leads to the videolan repo anyways
[19:01] <JEEB> so use the videolan url
[19:02] <JEEB> probably borked DNS entry or something
[19:02] <Fjorgynn> now I must have libx264-129.dll in the same directory
[19:02] <Fjorgynn> but whatever
[19:02] <hotwings> ok thanks JEEB
[19:02] Action: Fjorgynn hugs JEEB *thanks*
[19:02] <JEEB> Fjorgynn, you did --enable-shared and not static did you not?
[19:02] <JEEB> with x264's configure
[19:03] <JEEB> or you already had a libx264.dll.a in the library lookup path
[19:03] <JEEB> which is the shared library
[19:03] <JEEB> if you have libx264.a and libx264.dll.a in the same folder, remove the .dll.a one
[19:03] <JEEB> to make sure it's not used
[19:05] <Fjorgynn> aha
[19:09] <Fjorgynn> I see
[19:10] <Fjorgynn> JEEB: is it presets and the ffmpeg.exe I need?
[19:10] <Fjorgynn> or just the ffmpeg.exe?
[19:10] <JEEB> usually you don't even need the ffpreset files
[19:10] <Fjorgynn> :)
[19:10] <JEEB> because libx264's presets are inside it
[19:10] <JEEB> so they're within the binary
[19:11] <Fjorgynn> and with this build I can convert avi mpeg4 to x264 (h.264) ?
[19:11] <Fjorgynn> now I must learn how to use the fdk-aac :o
[19:12] <Fjorgynn> now it compiles my new binaries :D
[19:14] <JEEB> yes, most if not all decoders are in ffmpeg itself
[19:14] <JEEB> and you just need to encode H.264
[19:14] <JEEB> and you have libx264 for that
[19:14] <JEEB> and fdk-aac for aac
[19:17] <Fjorgynn> is it -q:a 1-5 instead of -b:a 128k I should use now or what?
[19:17] <JEEB> whichever you want
[19:17] <Fjorgynn> I see
[19:17] <Fjorgynn> I've noticed that youtube and vbr isn't friends :(
[19:20] <JEEB> shouldn't have anything to do with that
[19:21] <Fjorgynn> hmm it still complains about that file...
[19:22] <Fjorgynn> but I recompiled x264 and ffmpeg
[19:22] <Fjorgynn> ./configure --disable-cli --enable-static --prefix=/mingw/i686-w64-mingw32
[19:22] <JEEB> that's fine, now you should have libx264.a in /mingw/i686-w64-mingw32/lib
[19:23] <JEEB> and x264.h in /mingw/i686-w64-mingw32/include
[19:23] <DJJeff> im really having a hard time with flv to swf quality here is my output
[19:23] <DJJeff> http://pastie.org/pastes/5627484/text
[19:23] <Fjorgynn> JEEB: and?
[19:24] <JEEB> do you have any other libx264.something files in either /mingw/lib or /mingw/i686-w64-mingw32/lib?
[19:24] <JEEB> ending with .a
[19:24] <JEEB> esp. libx264.dll.a
[19:25] <Fjorgynn> yes in the i686-w6...
[19:26] <Fjorgynn> not any more
[19:27] <JEEB> you should only have your own libx264.a which is the static library around, basically
[19:31] <Scorcerer> i'll forgo virtyaldub and try use avidemux :)
[19:31] <JEEB> avidemux... I've seen nothing good from it
[19:31] <JEEB> if you just need to remux
[19:31] <JEEB> use vdub or possibly nandub
[19:32] <JEEB> it'll take care of most of those hacky files if it is possible
[19:34] <Scorcerer> do i need decompressor for usign direct stream copy mode then ?
[19:34] <Scorcerer> or just load file, set mode to direct stream copy andt save ?
[19:38] <Fjorgynn> JEEB: one problem
[19:38] <Fjorgynn> ./configure --disable-cli --enable-static --prefix=/mingw/i686-w64-mingw32
[19:38] <Fjorgynn> if I do that (redownloaded x264) I get an error saying that libx264 isn't foud
[19:38] <Fjorgynn> when I try to configure ffmpeg
[19:39] <Fjorgynn> so yeah
[19:40] <JEEB> Fjorgynn, how do you configure ffmpeg :P
[19:41] <JEEB> your x264 configuration seems just fine
[19:41] <DJJeff> ffmpeg has hard time going from flv to swf is there other programs that work better in ubuntu/debian ?
[19:41] <Fjorgynn> ./configure --enable-libfdk-aac --enable-libx264 --enable-gpl --enable-version3 --enable-nonfree
[19:42] <Fjorgynn> http://pastebin.ca/2299663
[19:42] <Fjorgynn> x264
[19:42] <JEEB> PKG_CONFIG_PATH=/mingw/i686-w64-mingw32/lib/pkgconfig ./configure --enable-libfdk-aac --enable-libx264 --enable-gpl --enable-version3 --enable-nonfree --extra-cflags="-I/mingw/i686-w64-mingw32/include" --extra-ldflags="-L/mingw/i686-w64-mingw32/lib"
[19:42] <JEEB> try this
[19:43] <sacarasc> --disable-cli? :o
[19:43] <JEEB> he doesn't need the x264 cli
[19:43] <JEEB> only the library
[19:44] <sacarasc> Oh, that's x264's configure.
[19:44] <sacarasc> I got confused, sorry.
[19:45] <Scorcerer> JEEB: looks like VirtualDub worked (at least on one file - testing in ffmpeg now) - no warnings in ffmpeg so far.
[19:47] <Scorcerer> but audio is still out of sync :/
[19:48] <Fjorgynn> same problem JEEB
[19:48] <JEEB> pastebin config.log
[19:51] <Scorcerer> JEEB: Streams in files are numbered from 0 or from 1 ?
[19:51] <Fjorgynn> JEEB: http://pastebin.ca/2299666
[19:51] <JEEB> Scorcerer, yes
[19:51] <JEEB> from zero
[19:51] <JEEB> both open inputs and streams in inputs
[19:52] <Scorcerer> ok, so then the error i got from VirD is from subtitle stream, nothing to worry about then :)
[19:52] <Scorcerer> "AVI: Stream 2 has an invalid sample rate. Substituting 15 samples/sec as a placeholder
[19:53] <Scorcerer> "
[19:53] <JEEB> Fjorgynn, ugh seems like x264 chose pthreads and not win32threads... since it was available.
[19:53] <JEEB> --enable-win32thread
[19:53] <JEEB> add this to x264's configure
[19:54] <JEEB> to make it use standard win32 threads
[19:54] <JEEB> for some reason if x264 can find pthreads it will go there instead of just keeping to win32threads by default
[20:01] Action: Fjorgynn gives JEEB a big kiss
[20:02] <Fjorgynn> ffmpeg version N-48482-g6717d1a Copyright (c) 2000-2013 the FFmpeg developers built on Jan  5 2013 19:19:10 with gcc 4.7.2 (GCC) configuration: --enable-libfdk-aac --enable-libx264 --enable-gpl --enable-vers
[20:02] <Fjorgynn> ion3 --enable-nonfree
[20:04] <DJJeff> avi to mp4 is pretty impressive in terms of quality and filesize
[20:07] <Fjorgynn> yeah
[20:07] <Scorcerer> am i wrong, or avi is just a container in which you can store anything you want (like x264)?
[20:08] <sacarasc> Yes, that is correct.
[20:08] <sacarasc> Well, not anything, but some things.
[20:09] <Fjorgynn> why are people using libx264 with a i container?
[20:09] <Fjorgynn> AVI
[20:09] <Fjorgynn> that is
[20:10] <Scorcerer> for me, it's just that i have every movie in an avi
[20:10] <Scorcerer> what container it should be stored in, then? mp4 ?
[20:11] <sacarasc> Whatever you want, but don't reencode for reencoding's sake.
[20:12] <Fjorgynn> all new stuff I make are mp4
[20:17] <JEEB> AVI + B-frames is a very bad combo (be it MPEG-4 Part 2 or H.264)
[20:17] <JEEB> for MPEG-4 Part 2 (divx/xvid etc.) AVI ended up being the "de facto" container
[20:18] <JEEB> but at least with H.264 people switched to other containers
[20:18] <Scorcerer> so the test encoding ended, it looked like this : http://pastebin.ca/2299669
[20:19] <Scorcerer> i had to manually stop it after i checked it's legth which was 2:10:xx, instead of stopping after, like, 1:36:xx
[20:19] <Scorcerer> :(
[20:20] <Scorcerer> audio was not only not synced, but also fast played from beginning after every skip (forward/backward)
[20:20] <Scorcerer> but video was great :)
[20:20] <Scorcerer> not sure what to do now :/
[20:27] <Fjorgynn> Scorcerer: copy the audio stream?
[20:28] <Fjorgynn> I dunno
[20:29] <Scorcerer> tried that, didn't work :(
[20:30] <Scorcerer> for now, i'll use virtualDub on all files
[20:30] <Scorcerer> and then i'll think what to do :)
[20:34] <Fjorgynn> btw flac seems to cut wav files in hal
[20:34] <Fjorgynn> half
[20:40] <Scorcerer> still won't stop after normal end of video :/
[20:41] <JEEB> it is probably getting the packets and how long the whole thing is supposed to be wrong
[20:42] <JEEB> nandub /might/ be needed for VBR audio in avi
[20:42] <JEEB> (welcome to the garden of hacks that is "normal stuff in AVI"
[20:42] <Scorcerer> that's why i'm migrating wrom this divx %#*%^
[20:46] <Fjorgynn> is there a way to se what -q:a I should have?
[20:46] <Fjorgynn> for aac?
[20:49] <JEEB> "I should have"?
[20:51] <espr3ss0> Q: i saw when looking through some web-pages but have forgotten what ... do i need to with 'make' "install-lib / install-man" for any of the following ? ... "gettext / glib / gpac / libiconv / pkg-config / x264"
[20:52] <Fjorgynn> JEEB: what I understand there is 1-5, what does that mean and how do I find out about that?
[20:53] <JEEB> no idea (´
[20:53] <Fjorgynn> -q:a seems not to work at all
[20:55] <Scorcerer> nandub didn't help :(
[20:55] <Scorcerer> damn
[20:55] <JEEB> can you extract the audio track from the avi?
[20:56] <JEEB> avconv -i derp.avi -vn -c:a copy out.ac3 if it's ac3
[20:57] <Scorcerer> i don't have avconv yet, sec
[20:57] <JEEB> argh
[20:57] <JEEB> ffmpeg I mean
[20:57] <JEEB> sorry, getting tired :P
[20:57] <JEEB> (as well as multitasking)
[20:57] <Scorcerer> sure, no problem :)
[20:58] <Scorcerer> that's strange
[20:58] <Fjorgynn> -vbr seems to be good
[20:58] <Scorcerer> now time= was equal to total length of movie
[20:59] <Scorcerer> and earlier (when transcoding all at once) it wasn't
[20:59] <Scorcerer> it extracted ok
[21:00] <Scorcerer> maybe i should extract video stream too ?
[21:04] <JEEB> nah
[21:04] <JEEB> just encode the video (-an is "audio none")
[21:05] <JEEB> or wait
[21:05] <Scorcerer> i'm doing it now, into out.avi
[21:05] <Scorcerer> done
[21:05] <Scorcerer> and time wal also good (time=01:36:37)
[21:05] <Scorcerer> was*
[21:05] <JEEB> ffmpeg -i input.avi -i out.ac3 -map 0:1 -map 1:1 -c copy out.mp4
[21:06] <JEEB> ugh
[21:06] <JEEB> -map 0:0 and -map 1:0
[21:06] <JEEB> I meant
[21:06] <Scorcerer> one sec
[21:06] <JEEB> that should pick the video from the avi, and audio from the ac3 file
[21:08] <Scorcerer> [mp4 @ 0x645630] track 0: could not find tag, codec not currently supported in container
[21:08] <Scorcerer> shouldn't it be 0:0 and 0:1 ?
[21:08] <JEEB> pastebin full log
[21:08] <JEEB> and no
[21:08] <JEEB> because you have two inputs
[21:09] <Scorcerer> ah, right
[21:09] <JEEB> so basically the first track from inputs 0 (probably the video) and 1 (the ac3 file)
[21:09] <Scorcerer> http://pastebin.ca/2299681
[21:10] <JEEB> ok
[21:10] <JEEB> msmpeg4 is not ok into mp4
[21:10] <JEEB> try dot-mkv
[21:10] <Scorcerer> ok, now it works
[21:11] <Scorcerer> ok, done with right time again
[21:11] <JEEB> yeah, taht thing just meant that the format X can't be put into mp4
[21:11] <JEEB> *that
[21:11] <Scorcerer> now transcode in x264 ?
[21:11] <Scorcerer> to*
[21:11] <JEEB> if that mkv is fine, you can transcode with those two inputs :)
[21:13] <Scorcerer> no video in mplayer
[21:13] <Scorcerer> one sec
[21:13] <developer> can ffmpeg encode video with srt or ass  subs from mkv ?
[21:16] <sacarasc> developer: Yes.
[21:16] <sacarasc> -vf ass, IIRC.
[21:16] <JEEB> not sure if that could load the subtitles straight from the file
[21:17] <JEEB> (as in, muxed stream from input)
[21:17] <JEEB> if yes, great
[21:18] <Scorcerer> JEEB: no video output from .mkv, but i'm transcoding anyway, mplayer spits out smth like this: http://pastebin.ca/2299684
[21:19] <developer> would be really good to just encode with the subs already from mkv instead of muxin out the tracks and remuxin :)
[21:19] <JEEB> Scorcerer, looks like mplayer just fails at playing it :P
[21:19] <Scorcerer> yeah, i think so too
[21:20] <Scorcerer> one more question, what does "time=xx:xx:xx.xx" mean? place in file where ffmpeg is now ?
[21:26] <espr3ss0> Q: can mp4v2 be used instead of gpac ?
[21:27] <sacarasc> I don't think you need either for ffmpeg.
[21:29] <espr3ss0> ok
[21:29] <espr3ss0> thanks
[21:37] <spaam> ubitux: does your HDS demuxer handle encrypted streams?
[21:42] <durandal_1707> HDS?
[21:42] <spaam> durandal_1707: HTTP Dynamic Streaming
[21:42] <spaam> durandal_1707: Adobe thing :)
[21:47] <durandal_1707> first it is not demuxer and second it is not ubitux demuxer
[21:52] <durandal_1707> and ubitux did not do much with it because company for which he worked lost interest for hds
[21:53] <spaam> ok :)
[22:19] <doublehp> i can fully read a file with mplayer and vlc, including ST, but ffmpeg complains: Decoder (codec text) not found for input stream #0:3
[22:22] <doublehp> http://pastebin.ca/2299687
[22:24] <doublehp> how is it possible that both players can read it fine, but ffmpeg won't find the codec ???
[22:25] <doublehp> it may be a broken dep in Gentoo, but, i don't know which one
[22:25] <JEEB> I'm /pretty/ sure you didn't mean to re-encode there
[22:26] <JEEB> and that just means that it can't convert from srt to ass subtitles
[22:27] <doublehp> JEEB i did not start yet to work on the recoding aspect; i do want to recode, to make file smaller; any help welcome :) I wanna keep the 2 audios and the subs
[22:28] <JEEB> -map 0 maps all tracks from input 0 (the first -i you have), otherwise only the "best" track will be grabbed of both video/audio/subtitles into output. If you want to copy audio and subtitles, set -c:a copy -c:s copy
[22:28] <JEEB> and I don't really know what you mean with "make the file smaller". The input already seems to be an x264-encoded H.264 stream
[22:29] <JEEB> I mean, you won't get it much smaller without losing quality unless you resize it down
[22:30] <doublehp> input is 16GB; i want output to be around 1GB
[22:33] <JEEB> How that will look will depend on what kind of content there is in that movie then (its compression complexity), as well as how well it was originally already compressed. The x264 wording in there means that someone most probably already once has encoded that from a source with x264
[22:33] <JEEB> I recommend you use -ss and -t to seek to a certain part that actually has content, and then encode 25-30 seconds worth of content with -c:v libx264 -crf 23
[22:33] <JEEB> crf 23 is also the default for libx264
[22:34] <JEEB> (-ss is seek, and -t is "how much to encode time-wise, basically")
[22:34] <JEEB> if the output looks good, you make the crf value higher
[22:34] <JEEB> if it looks bad, make it somewhat lower
[22:35] <JEEB> do that for a while and you will find the highest crf value that still gives you quality that you like
[22:35] <JEEB> after that you can start setting up the preset, all of them are listed http://mewiki.project357.com/wiki/X264_Settings#preset <- here, and they are set with the -preset setting
[22:35] <JEEB> medium is the default
[22:36] <JEEB> slower presets = better compression
[22:36] <JEEB> you can then try a few and see what is the slowest that is still fast enough for you
[22:36] <JEEB> after that you can try encoding with those settings
[22:36] <JEEB> if it will still end up too big, then you have to resize
[22:37] <JEEB> but that's it so far
[22:38] <JEEB> (you could just set the bit rate to a certain size, but that would prioritize "I want this size!" instead of quality
[22:38] <JEEB> (and for bit rate based encoding you would need two passes)
[22:38] <doublehp> JEEB it added a second new sub ....
[22:38] <JEEB> haha, do show the log
[22:39] <JEEB> (with the command as well as output)
[22:40] <JEEB> also btw, I do recommend you build a newer ffmpeg :) Esp. if you're in an environment like gentoo where you have to build your stuff yourself anyways
[22:41] <doublehp> http://pastebin.ca/2299689
[22:41] <Scorcerer> JEEB: it works now, splitting audio and video seems to do the trick
[22:41] <JEEB> Scorcerer, seems like the audio track demuxing barfed and trying to extract it separately first was a good idea :)
[22:42] <JEEB> doublehp, I see only one subtitle track?
[22:42] <JEEB> or wait
[22:42] <doublehp> one in input is fine; but mplayer founds two in output
[22:42] <JEEB> uhh
[22:42] <JEEB> Stream #0:3(fre): Subtitle: text (default) Stream #0:4(fre): Subtitle: text
[22:43] <JEEB> you have two subtitle tracks in the damn input!
[22:44] <doublehp> won't be the worst issue even in a 1GB file; forget it.
[22:44] <JEEB> in other words, it didn't /add/ any subtitle tracks
[22:44] <JEEB> there were two to begin with
[22:44] <JEEB> and yes
[22:44] <JEEB> subtitle tracks don't really matter
[22:44] <JEEB> they are really small in comparison to pretty much anything else
[22:45] <JEEB> unless you are putting base64-encoded data there
[22:45] <doublehp> i need to crop top and bottom, and then, resize to, let say, twice smaller
[22:45] <Scorcerer> JEEB: yup, that's the way to do it, but i still have questions, fo egzample: do i have to provide .ac3 for ac3 sound and mp3 if it will be encoded im mpeg2layer3, or there is some container i can use to store it in?
[22:45] <Scorcerer> universal container*
[22:45] <JEEB> mkv I guess?
[22:46] <Scorcerer> ah, ok
[22:46] <Scorcerer> i would like to create one for loop to transcode everyfile in directory over, say a week :)
[22:46] <JEEB> doublehp, http://ffmpeg.org/ffmpeg-filters.html#crop and http://ffmpeg.org/ffmpeg-filters.html#scale
[22:46] <JEEB> both have examples, too :)
[22:47] <doublehp> found -vf crop=
[22:47] <JEEB> yes, that's the documentation for it that I just linked
[22:47] <JEEB> the other filter is the scale filter
[22:47] <JEEB> you first crop, then scale
[23:03] <DJJeff> im really having a hard time with flv to swf quality here is my output
[23:03] <DJJeff> http://pastie.org/pastes/5627484/text
[23:06] <doublehp> my movie is 1920x800, that's 2.4; which smaller sizes respect this ratio ?
[23:07] <doublehp> 1152 480
[23:19] <klaxa> 940x400?
[23:19] <klaxa> uh...
[23:19] <klaxa> 960x400
[23:19] <doublehp> yup, could do
[23:20] <doublehp> but i think that 1152 is a standard for laptops; never saw 960x400 as native size for any screen
[23:22] <klaxa> yeah generally using 480 for horizontal pixels is a good value (i guess?)
[23:23] <doublehp> hmmm
[23:23] <doublehp> you are right , with 960x400, interpolation will be nicer
[23:24] <klaxa> well it's just half then horizontal size
[23:25] <doublehp> now; how do i copy audio 1 ; but recode audio 2 ?
[23:25] <klaxa> look at -map
[23:26] <hub2> Hi,Is there a C function that checks how many frames left in queue?
[23:29] <doublehp> this soft is wonderfull :) <3
[23:48] <doublehp> JEEB sub 0 is a fake empty one; i am removing it. thanks for help. I have written every thing in a text file for future use. I have spent long time on it, not just for this file, but wrote my tuto for next time :)
[23:49] <doublehp> i think that, to see the difference between a 16GB and 1GB file, i would need a 42" screen; i only got 19" :P
[23:50] <sacarasc> I think you need an eye test. :p
[00:00] --- Sun Jan  6 2013