[Ffmpeg-devel-irc] ffmpeg.log.20120225

burek burek021 at gmail.com
Sun Feb 26 02:05:02 CET 2012


[00:02] <vadim_> rainmaker1: 32 bits... this system was already installed on a test machine and i decided not to reinstall
[00:04] <rainmaker1> vadim_: still running, I will leave it like this in screen will see if it will crash until tommorow
[00:04] <vadim_> ok... let me know somehow ))
[00:05] <rainmaker1> vadim_: I'll be here tommorow
[00:05] <vadim_> rainmaker1: ok
[00:17] <Christophe_v_E> I've managed to activate ffmpeg - how do I run a command to it through php?
[00:21] <teratorn> Christophe_v_E: the same way you run any command in php, I guess
[00:24] <Christophe_v_E> so if I were to call an execute in php I could just for example put this in there and it would work? ffmpeg â¬i video.flv -acodec libfaac -ss 00:00:00 -t 00:03:00 -ar  -ab  -f flv -s videopart.flv
[00:30] <teratorn> Christophe_v_E: I don't use php, so I wouldn't know
[00:32] <Christophe_v_E> And if I wouldn't want to do it in php? Say I have a folder with flv's - with what could I split all the flv's in the folder into random pieces and put 20 random pieces together? Any help is appreciated :)
[00:32] <nextime> Christophe_v_E : probably you will have issue in managing the execution time
[00:32] <Christophe_v_E> In php I've set the execution time to unlimited .. I don't know if that'll account for everything, but I suppose so - when using xampp.
[00:33] <nextime> ouch
[00:33] <nextime> what a bad thing to do :)
[00:33] <Christophe_v_E> it's to run locally for a school program
[00:34] <nextime> Christophe_v_E : i have to admit that i'm a sort of anti-php taleban, so, do what you are doing and don't take care of my opinions on that
[00:34] <nextime> :)
[00:34] <Christophe_v_E> I'm willing to pay somebody to make something that will split the flv's up to random pieces. Then - with a seperate trigger put them together again as one video and save it, every new generation would have a different name of course.
[00:35] <Christophe_v_E> Depending on the price of course - if it's going to cost me too much for what you would make in half an hour I'd rather learn to do it myself for an entire day :)
[00:56] <rickk> hi
[00:56] <rickk> I have compiled avcodec with libx264 library. but the fuction avcodec_find_encoder cannot find CODEC_ID_H264.
[00:56] <rickk> ./ffmpeg -codecs|grep x264
[00:56] <rickk>   EV    libx264         libx264 H.264 / AVC / MPEG-4 AVC / MPEG-4 part 10
[00:56] <rickk>   EV    libx264rgb      libx264 H.264 / AVC / MPEG-4 AVC / MPEG-4 part 10 RGB
[01:07] <Christophe_v_E> Can somebody please explain me how I run a command. Thank you.
[01:12] <Christophe_v_E> Is there any guide at all on how to use this? It seems like the documentation dives right into commands while I don't even know what to run them in to get something done. I'd really like advise.
[01:15] <pasteeater> Christophe_v_E: your example will re-encode. you can also "copy and paste" with relative accuracy.
[01:15] <pasteeater> ffmpeg -i input.flv -ss 00:01:29.00 -t 15 -c copy output.flv
[01:15] <Christophe_v_E> I have no idea what I have to do with my example - where do I put it to get something done, pasteeater ?
[01:16] <Mavrik> why are you asking us PHP questions?
[01:16] <Christophe_v_E> doesn't have to be php at all -
[01:16] <Mavrik> for the sake of all your users and server security, go read the damn PHP documentation so you'll understand what you're doing
[01:16] <Christophe_v_E> .. that is .. rather friendly. I'm going to run it locally - for a small school thing.
[01:18] <pasteeater> are all flv in your directory going to have the same ss and t?
[01:18] <Christophe_v_E> they're all variable
[01:18] <Christophe_v_E> the flv's are different lengths, as should be the splitted parts, as should be the outputs
[01:19] <Christophe_v_E> But I'd really like to know where I would put a command .. literally. I haven't come accross anything telling me where I should go with my command .. - thanks for your interest :)
[01:30] <undercash> is it possible to stream a playlist with ffmpeg?
[01:50] <pasteeater> undercash: if it's a simple text file you can do it with a bash script.
[01:51] <pasteeater> maybe it would be a good SOC project.
[01:52] <undercash> using this cmd
[01:52] <undercash> ffmpeg -re -i "doc.avi" -vcodec libx264 -preset fast -crf 26 -acodec libfaac -ab 128 -ar 44100 -f flv rtmp://live.justin.tv/app/xxx
[01:52] <undercash> shouldn't this be a xml file like vlc
[01:53] <pasteeater> you probably want 128k instead of 128 (or omit -ab as 128k is default, IIRC).
[01:53] <undercash> oh yea
[01:53] <undercash> but how is supposed to be the text file? any exemple?
[01:56] <pasteeater> i meant use can have a text file of a list of your inputs and then "for file in ${inputfile} do", etc.
[01:56] <undercash> i d like just to use the terminal to stream..
[01:56] <pasteeater> maybe i'm not understanding what you're asking, but now i have to go back to work.
[01:57] <undercash> ok passteeater thx
[01:58] <qxt> From shell how can I get the length of a video.
[02:00] <qxt> Writing a script to do 2 pass encoding with a predetermined output size.
[02:03] <daimonion> Hi!
[02:03] <daimonion> Is anybody here?
[02:04] <qxt> Sort of . We are just acting like we are cool and ignoring ppl. SUP
[02:04] <sacarasc> qxt: Use the programme mediainfo. It's awesome at this kind of thing.
[02:05] <daimonion> Well, I have a problem compiling ffmpeg.
[02:05] <qxt> sacarasc I just remembered ffprobe
[02:05] <sacarasc> I prefer mediainfo for things like this.
[02:05] <daimonion> I think it is due to changes in vo_aacenc
[02:05] <qxt> daimonion what OS
[02:06] <daimonion> Arch Linux.
[02:06] <daimonion> qxt: Any other information needed?
[02:07] <sacarasc> daimonion: The output from make would be nice. :p
[02:08] <daimonion> http://pastie.org/3451503
[02:08] <qxt> daimonion I got a few scripts I wrote that will install upgrade remove ect for debian systems. Not 100% sure it will work with Arch (a great distro btw)
[02:08] <pasteeater> daimonion: what's wrong with ffmpeg-git from AUR?
[02:09] <daimonion> I don't know, I do not use it. :)
[02:09] <pasteeater> it works. adding libvo_aacenc support is easy.
[02:09] <daimonion> I use my own PKGBUILD, which compiles from here: http://www.ffmpeg.org/releases/ffmpeg-snapshot.tar.bz2
[02:09] <pasteeater> vo-aacenc PKGBUILD is also in AUR
[02:10] <daimonion> I know, I have the latest vo-aacenc and vo-amrwbenc installed.
[02:10] <pasteeater> all you have to do is add vo-aacenc to depends and '--enable-libvo-aacenc --enable-version3' to the ffmpeg configure if you use the AUR PKGBUILDs
[02:10] <daimonion> Which btw broke ffmpeg: ffmpeg: error while loading shared libraries: libvo-amrwbenc.so.0: cannot open shared object file: No such file or directory
[02:11] <daimonion> That's why I want to recompile, but I get error.
[02:11] <pasteeater> vo-* probably is built as shared by default
[02:11] <pasteeater> (the ones from AUR do not)
[02:12] <qxt> daimonion Ill put a automagic script up for ArchLinux tomorrow over at my pastebin http://pastebin.com/u/Onryo
[02:12] <qxt> Just because I like Arch =)
[02:12] <pasteeater> i don't understand what's wrong with using the simple AUR packages
[02:13] <daimonion> Here are my build flags: http://pastie.org/3451538
[02:13] <daimonion> OK, thanks qxt.
[02:14] <pasteeater> you don't need --enable-runtime-cpudetect unless you're a distro, cross-compiling, or going to be using the binary on other machines with different cpu
[02:14] <pasteeater> --enable-avfilter is no longer an option, IIRC
[02:14] <qxt> daimonion The Ubuntu guys baggered me a while back for this http://pastebin.com/uHZx6wJR but that is the last thing I am doing for Ubuntu. Use Debian myself
[02:15] <pasteeater> onryo: ffmpeg -i IronMan.mkv 2>&1 | awk '/Duration/{print $2}' | sed 's/,//g'
[02:15] <pasteeater> probably a million other ways to get the same. not sure if it will work for all inputs.
[02:15] <onryo> pasteeater yup you got a darn good memory. I asked you about speeds =)
[02:16] <onryo> pasteeater was the same clip. You had a 920 sandy bridge IIRC?
[02:16] <onryo> Well it was fast
[02:17] <daimonion> Sorry, my connection broke.
[02:18] <pasteeater> onryo: i7 860.
[02:18] <pasteeater> upgraded from an old P4 room heater.
[02:18] <daimonion> pasteeater: As you can see, ffmpeg-git from AUR uses --enable-nonfree build flag, which I don't want. Also, it doesn't use vo_aacenc.
[02:19] <pasteeater> pasteeater | all you have to do is add vo-aacenc to depends and '--enable-libvo-aacenc --enable-version3'
[02:19] <rickk> can anybody help me with av_log ?
[02:19] <pasteeater> nonfree is for faac. just remove faac from depends and nonfree libfaac from configure
[02:19] <daimonion> pasteeater: This is whyt I did in my own PKGBUILD script.
[02:20] <pasteeater> that's one of the reasons to have AUR. so you can modify the build scripts.
[02:20] <rickk> void avcodec_register(AVCodec *codec)
[02:20] <rickk> {
[02:20] <rickk>     AVCodec **p;
[02:20] <rickk>     avcodec_init();
[02:20] <rickk>     p = &first_avcodec;
[02:20] <rickk>     while (*p != NULL) {
[02:20] <rickk>     av_log(*p,1,"Name : %s CodecID 0x%x",(*p)->name,(*p)->id);
[02:20] <rickk>     p = &(*p)->next;
[02:20] <rickk> }
[02:20] <pasteeater> rickk: you should use pastebin.com for a multi line paste
[02:20] <rickk> sorry
[02:21] <pasteeater> daimonion: you take a working AUR PKGBUILD and modify it to suit your needs. it's usually easier and faster than doing it from scratch.
[02:21] <rickk> http://pastebin.com/EjTQqTNG
[02:22] <rickk> i cannot display av_log
[02:22] <rickk> i like to list all the codecs to be registered
[02:23] <daimonion> I don't understand what part you don't get: I did exactly that, the problem isn't in my PKGBUILD script, the problem is in the ffmpeg itself. ffmpeg wont compile since vo_aacenc and vo_amrwbenc upgrade.
[02:23] <pasteeater> i'll try
[02:26] <pasteeater> Unknown option "--enabke-libvo-amrwbenc"
[02:26] <pasteeater> durrrrr
[02:30] <pasteeater> daimonion: worksforme
[02:30] <pasteeater> don't build as shared
[02:41] <onryo> OK my script seems be be working. Calculates the b:v and b:a if need for predetermined "movie" sizes. My question is if there is anything wrong with the output format from this script or if it could be cleaner.
[02:41] <onryo> ffmpeg -i movieIn.avi -pass 1 -vcodec libx264 -preset veryslow -b:v 1226.72134943k -f matroska -sn -an /dev/null && ffmpeg -i movie.avi -pass 2 -vcodec libx264 -preset veryslow -b:v 1226.72134943k -acodec libfaac -ab 128k -ac 2 -sn -y movieOut.mkv
[02:41] <onryo> Yeah the floating points are there in the video etc
[02:42] <pasteeater> looks fine to me. check to see that the output is actually yuv420p which is probably what you want. if the input is dv it might pick yuv422 instead.
[02:43] <cali> is there any work on the Adobe F4F crap decoding?
[02:45] <rickk> can anyone help me...
[02:46] <rickk> i like to list all the codecs to be registered
[02:46] <rickk> http://pastebin.com/EjTQqTNG
[02:46] <onryo> pasteeater good idea. Going to put in a few checks. But seems to look good. I guess Ill drop this into qt to make it easy for others to use too. calculating 2 pass predetermined sizes is a drag.
[02:46] <rickk> I don't know y av_log doesn't show..
[02:46] <rickk> i set the log level as 0 though
[02:52] <pasteeater> onryo: you could just add '-pix_fmt yuv420p'
[02:53] <onryo> pasteeater Was thinking about that. Would have have to use that for both passes?
[02:54] <pasteeater> yes
[02:54] <onryo> thx
[02:55] <pasteeater> if you use -b:v, then you can also use -b:a for consistency instead of -ab </nit>.
[02:56] <onryo> -b:a 128k ... For some reason I think my computer can read my mind. Sloppy of me.
[02:59] <onryo> Does ffmpeg round off the floating points in the bit rate?
[02:59] <onryo> example my script calculated this for video -b:v 1226.72134943k
[03:10] <onryo> cout << "The size of a float is:\t\t" << sizeof(float) << " bytes.\n";
[03:10] <onryo> sry
[03:29] <onryo> pasteeater Is a planar YCbCr colorspace (pix fmt yuv420p) the best quality/compression for x264. Not going to say I am an expert in that. Wrote some C program a few years back for converting RGB 24bpp -> YUV420P 12bpp and got to admit I did not have a clue about what I was doing =)
[03:31] <onryo> Anybody who might have some advice is welcome with feedback.
[03:33] <pasteeater> i think a better determinant is what players you want to support. all can play yuv420.
[03:34] <pasteeater> techically, x264 does support RGB, but i don't think it's encouraged and most players only support YUV.
[03:34] <pasteeater> AFAIK
[03:40] <onryo> pasteeater Well its not a bad way to roll that is for sure. Not full of BS like wine tasters and say  "I get a hint of spring peaches" regarding colors I can't see anyway. The numbers add up. pix_fmt yuv420 planer it is.
[03:44] <onryo> Judging by my spelling its about time for me to go to bed. 4am again...I'll never learn. gn
[03:44] <pasteeater> full bouquet. earthy. diesel. leather. chalk.
[11:12] <guy> simple ffmpeg produces not what expected when simply trying to get file info
[11:12] <guy> https://gist.github.com/f65fb9b045c0096cb7a4
[11:12] <guy> am I doing something wrong?
[11:13] <guy> it is complaining that second file paramter is needed to start encoding. but I am simply trying to get audio info
[11:16] <mystica555_> -i is to designate input file, not give information. i personally recommend the 'mediainfo' package for what you are neeidng
[11:16] <mystica555_> *needing
[11:16] <JEEB> or ffprobe
[11:17] <JEEB> which comes with ffmpeg
[11:17] <JEEB> also you're getting audio information, no?
[11:17] <mystica555_> JEEB: good point
[11:20] <mystica555_> tbh i havent even messed with ffmpeg outside of the actual ffmpeg/ffplay commands
[12:12] <pierow> Hello, I was wondering if anyone could give me some help with AVCodecContext values for video and audio encoding? Is there a reference where I could find good default values for encoding different audio and video codecs?
[12:12] <pierow> Even just a point to somewhere in ffmpegs code would be very helpful :)
[12:13] <pierow> hmmm should I maybe be asking this in #ffmpeg-devel?
[12:30] <gpannwitz> http://pastebin.com/yYY9yVRw    I am trying to encode videos for burning to DVD, but the produced files are way too big.  I tried setting -b to something small, but it seems to be ignored.
[12:49] <sacarasc> gpannwitz: First, you're not using ffmpeg but libav... Anyway, remove -sameq.
[15:10] Action: insider is away: auto-away
[16:01] <superdump> ahoy
[16:09] <pwlnw> does ffmpeg or libav have option for calculate proportional output video size ? i need convert multiple videos and don't want calculate
[16:22] <Mavrik> pwlnw, yes
[16:22] <Mavrik> use the "scale" filter
[16:22] <Mavrik> -vf scale=-1:720
[17:11] <pwlnw> Mavrik so i need calculate scale ? i want just use like "not more than 320x200". and ffmpeg automatically calculate sizes
[17:37] <pwlnw> Mavrik seems not work. i have libav : avconv -i defy.flv  -y -v verbose -vf scale=-1:320  -codec:a libfaac -ac 2 -ar 48000 -vcodec libx264 -pre:v slow -profile:v high -level 30 out.x.flv
[17:38] <pwlnw> Mavrik it doesn't support scale calculation ?
[17:42] <pwlnw>  -vf scale=320:-1 - works ok. but -vf scale=-1:320 not
[17:58] <iive> don't use libav.
[18:00] <relaxed_> pwlnw: quote the filter
[18:02] <pwlnw> relaxed_ don't understand how
[18:02] <pwlnw> -vf 'scale=-1:320' ?
[18:03] <relaxed_> yes, does that work?
[18:04] <pwlnw> [libx264 @ 0x14a64e0] width not divisible by 2 (569x320)
[18:04] <pwlnw> better but not work
[18:06] <relaxed_> -vf 'scale=ceil(-1):320'
[18:07] <pwlnw> same
[18:09] <relaxed_> well, -vf 'scale=570:320'
[18:10] <pwlnw> but i can't calculate size any time
[18:11] <relaxed_> most of the time it works
[18:13] <cbreak> mplayer's -vf scale supports rounding to full 2 or 16 multiples
[18:14] <cbreak> maybe ffmpeg does too
[18:15] <relaxed_> ffmpeg only has -1, where mplayer has -2+
[18:21] <spalo> can we join two videos using ffmpeg?
[18:21] <relaxed> spalo: look at concat in the man page
[18:23] <spalo> thanks
[18:27] <pwlnw> i need send exactly coding x264 params which  i see on another video. what i need to do ? i can read it from programe MediaInfo. but can't send  for example 'qpmin'
[18:27] <pwlnw> or 'ref=4'
[18:27] <cbreak> pwlnw: strings -n10 moviefile
[18:28] <cbreak> that will give you the full details
[18:28] <cbreak> (maybe better than mediainfo)
[18:28] <pwlnw> i already get it. but can't find how to send from ffmpeg to x264 lib
[18:29] <pwlnw> not all params available from ffmpeg?
[18:29] <pwlnw> and mediainfo.exe can copy to text
[18:30] <pwlnw> many of params i find
[18:32] <pwlnw> left is qpmin=0 qpmax=69 .
[18:33] <pwlnw> --qpmin 10 --qpmax 50 : Unrecognized option '-qpmin'
[18:35] <gpannwitz> sacarasc: thanks so much :D
[18:35] <pwlnw> should be put options into preset file ?
[18:39] <pwlnw> x264 says  : Option qpmin not found.
[18:40] <pwlnw> but x264 --fullhelp | grep qpmin says       --qpmin <integer>       Set min QP [0]
[18:44] <pwlnw> http://mailman.videolan.org/pipermail/x264-devel/2010-November/007995.html. ok
[22:03] <NovaSt0rM> heyo
[22:06] <NovaSt0rM> i've googled the hell out of this but i still can't find an answer: when using a custom streaming IO buffer, and returning EOF when there's no more data, av_read_frame() returns a partial frame, but not an error.  I can clearly see that EOF is reached, but more data will be forthcoming.  Is there any way to force av_read_frame to re-try the entire read once more data is available?  Currently,
[22:06] <NovaSt0rM> it's returning another broken frame (the 2nd half of the first frame) on the next call.
[22:08] <NovaSt0rM> aaand i just noticed the topic.
[22:08] <Mavrik> ^^
[22:13] <NovaSt0rM> lol "Questions about using FFmpeg or developing with the libav* libraries should be asked in #ffmpeg"
[22:13] <NovaSt0rM> their topic sends me here :P
[22:14] <NovaSt0rM> wonder if using av_grow_packet to append the next packet to the previous broken frame would work cleanly
[22:15] <NovaSt0rM> seems kinda sketchy though
[22:16] <vadim_> NovaSt0rM: i think your questions should be addressed to developers in #ffmpeg-devel
[22:17] <Mavrik> NovaSt0rM, problem is we mostly don't know ins and outs of latest ffmpeg APIs :\
[22:20] <rainmaker1> vadim_: Hi, just to tell you that ffmpeg crashed after 30 min...
[22:22] <NovaSt0rM> the API base is quite a beast ..
[22:22] <vadim_> rainmaker1: bad news... thank you...
[22:23] <rainmaker1> vadim_: I have two ideas right now, first have you tried to get source other then network (ie. file)
[22:23] <rainmaker1> vadim_: second, haveyou tried vlan?
[22:23] <rainmaker1> vadim_: sorry VLC :D
[22:23] <vadim_> rainmaker1: yeap... it worked very well
[22:24] <vadim_> relaxed: i eman the file source
[22:24] <vadim_> i mean*
[22:25] <rainmaker1> vadim_: aaa the file source, that's interesting, I thought to make a setup (if feasible) to capture traffic using tcpdump then pipe it to ffmpeg
[22:29] <vadim_> rainmaker1: vlc - the anwser is yes and no )) yes - i used it a few monthes ago on windows as a all in one solution for transcoding  -> streaming, same udp source... it crashed irregularely, but i was unable to find out why...
[22:30] <vadim_> rainmaker1: and no - haven't tryed it under linux
[22:33] <vadim_> rainmaker1: on bug tracker there is a ticket regarding our truble, even i'm very unfamiliar with *nix programing, i think the idea described there makes sense... https://ffmpeg.org/trac/ffmpeg/ticket/609
[22:34] <rainmaker1> vadim_: I will try to specify pkt_size to see if it helps. I am running your example as it crashes more often then my :D
[22:35] <vadim_> rainmaker1: lol... ok... but i did such tests, this does not mean you should not do them ))) guess what was the result?..
[22:35] <rainmaker1> vadim_: yeah, everything worked just fine? :D
[22:36] <vadim_> rainmaker1: would i be here than? ))
[22:38] <rainmaker1> vadim_: Just kidding, but I think you just love to pull out your hair for the bug that just don't exist :)
[22:41] <vadim_> rainmaker1: really? ))
[22:45] <rainmaker1> vadim_: :))) Now to be serious a little :) Can you tell me if michaelni reporduced this bug?
[22:45] <vadim_> rainmaker1: yes
[22:45] <rainmaker1> vadim_: and?
[22:46] <vadim_> rainmaker1: and... made a patch, wich didn't help much....
[22:47] <rainmaker1> vadim_: and he knows about it, or that fixed the bug at his test server?
[22:48] <vadim_> rainmaker1: probably he did not wait for 30+ minutes to see the result ))
[22:51] <rainmaker1> vadim_: pitty...
[22:51] <vadim_> rainmaker1: in fact, he resolved the problem ))) it does not "hang" any more... it just exists...
[22:54] <rainmaker1> vadim_: you see, it's fixed :)
[22:56] <clockwize> hey, is there a way to make ffmpeg require no input and not output the progress of the conversion? if the file exists, i just want it to overwrite, etc (running ffmpeg from another piece of software)
[22:57] <rainmaker1> clockwize: -y ?
[22:58] <vadim_> clockwize: about output, -loglevel quiet
[22:58] <clockwize> ok, thanks guys
[23:02] <rainmaker1> vadim_: new record :) frame=45190 fps= 25 q=28.0 size=  219582kB time=00:30:05.64 bitrate= 996.2kbits/s dup=16 drop=0
[23:03] <vadim_> rainmaker1: what about this record?..
[23:03] <rainmaker1> vadim_: well it's beyound 30 min :)
[23:04] <vadim_> rainmaker1: aaa..... i see ))
[23:05] <rainmaker1> vadim_: about your example command, did you actually use sparse or dense mode for multicast?
[23:07] <vadim_> rainmaker1: if you'd tell me the difference, may be i will answer your question...
[23:08] <rainmaker1> vadim_: maybe I shoud first ask if your source is some webcam or mpeg2 ts ?
[23:08] <vadim_> rainmaker1: mpeg2 ts
[23:11] <rainmaker1> vadim_: ok, then when you first start ffmpeg (with your example) it must send igmp join message for a multicast group. The difference is, that in sparse mode you MUST send join bla blah where in dense mode you don't have
[23:11] <rainmaker1> vadim_: except in case your source is in the same vlan
[23:12] <rainmaker1> vadim_: then it's pretty much the same
[23:14] <rainmaker1> vadim_: right now I am using other program to send regular igmp join message as in my network I use sparse mode
[23:17] <vadim_> rainmaker1: does it mean that in sparse mode ffmpeg would not read from multicast if no other program sends igmp join message?
[23:21] <rainmaker1> vadim_: it means that ffmpeg must send igmp report message at regular intervals
[23:23] <rainmaker1> vadim_: it's late again, I'll put ffmpeg in screen and see if it will crash till tommorow
[23:24] <rainmaker1> vadim_: I will let you know here..
[23:24] <vadim_> rainmaker1: ok )) have to go to bed too )
[23:43] <lucas^> is there a list somewhere or description of which iPod Touch versions can actually play stuff encoded by -f ipod -vpre ipod320/640?
[23:45] <Mavrik> pretty much any that supports Level 3.0  and baseline it seems
[23:45] <Mavrik> should be all of the,
[23:48] <lucas^> what's the difference between 320/640? I'm unfamiliar with the settings listed in them
[23:51] <Mavrik> is there a reason you don't check yourself?
[23:53] <lucas^> the reason being, I have no idea what the significance of 'level 13' is over 'level 30'
[23:53] <lucas^> also, for whatever reason, ffmpeg shows the output video as high profile, which I'm assuming isn't baseline
[23:54] <Mavrik> nop
[23:54] <Mavrik> lucas^, "level" is a limit of macroblocks/s in a video
[23:54] <JEEB> levels are basically ways of telling the decoder how much RAM it will need to decode the stream
[23:54] <Mavrik> which limits your resolution + bitrate
[23:54] <Mavrik> lucas^, basically you take spec from the ipod you want to support and check which max level it supports
[23:55] <JEEB> they should all be level 3.0 (30) and baseline IIRC
[23:55] <JEEB> only the very first ones might've had problems with level 3
[23:55] <JEEB> resolution at 320xSOMETHING
[23:55] <Mavrik> yeah
[23:55] <lucas^> 240p I think
[23:55] <JEEB> 320x180/320x240 I guess?
[23:55] <Mavrik> JEEB, that's probably the reason the ipod320 profile limits level to 1.3, ipod640 to 3.0
[23:56] <lucas^> yeah I think I remember playing 640x480 h.264 encoded videos on this thing a year or two ago
[23:56] <JEEB> is the model newer or older than the iphone 3gs?
[23:56] <lucas^> a macroblock is basically an arbitrary chunk of video data that needs to be decoded within a certain timeframe for realtime playback?
[23:57] <Mavrik> lucas^, basically ANY iPod Touch will play ipod640 encoded videos
[23:57] <lucas^> I think it's newer, could be wrong though
[23:57] <Mavrik> (since even 1st gen Touches support Baseline 3.0)
[23:57] <lucas^> I think the big problem is that, for whatever reason, these videos are being encoded with -profile high
[23:58] <JEEB> if it's newer you could feed it pretty much anything that's within level 4/4.1 and high profile
[23:58] <lucas^> the iPad can (supposedly) support high-profile level 30 video as well?
[23:58] <Mavrik> For High profile you need the iPhone 4 gen Touches afaik
[23:58] <JEEB> high/level4|4.1 as well for ipad
[23:58] <JEEB> Mavrik, nope
[23:58] <JEEB> 3GS->
[23:59] <Mavrik> ah, awesome
[23:59] <Mavrik> thought 3GS only supported Main
[23:59] <JEEB> nah
[23:59] <lucas^> the recently encoded media isn't showing up, assumed that meant iTunes or whatever deemed it unplayable
[23:59] <JEEB> yes, iTunes doesn't like
[23:59] <JEEB> iTunes = will only accept what apple officially wants the device to "supporT"
[00:00] --- Sun Feb 26 2012


More information about the Ffmpeg-devel-irc mailing list