[Ffmpeg-devel-irc] ffmpeg.log.20111213

[burek]

[00:01] <luke-jr> beandog: I'll be updating that bug with results when I find out if I can enable NEON
[00:02] <beandog> Hmm, okay, I'm looking for why it was turned off
[00:05] <beandog> https://bugs.gentoo.org/show_bug.cgi?id=349814
[00:06] <beandog> looks like it got switched in the 0.8 branch
[00:06] <beandog> in 0.7 it only disables them if you pass custom-cflags
[00:07] <luke-jr> beandog: I don't see hwo that bug is relevant
[00:07] Action: luke-jr notes 0.8 doesn't use custom CFLAGS no matter what you do :/
[00:07] <beandog> alright, ignore it then :)
[00:07] <beandog> hmm ... that got dropped too
[00:07] <beandog> very odd.
[00:11] <beandog> Okay I see what happened.
[00:14] <luke-jr> beandog: --extra-cflags='-mfpu=neon' without --disable-optimizations builds a NEON-enabled ffmpeg for me
[00:14] <luke-jr> beandog: ?
[00:14] <beandog> luke-jr: https://bugs.gentoo.org/show_bug.cgi?id=394493
[00:15] <beandog> I think it was just an oversight
[00:16] <luke-jr> beandog: your bug should block 392451 ;)
[00:16] <beandog> yahp
[00:17] <beandog> thx
[00:21] <beandog> thanks for the  bug report
[00:21] <luke-jr> np
[00:22] <luke-jr> (yay, a friendly Gentoo developer! :P)
[00:23] Action: mystica555_ wonders when darkice will get bumped..
[00:23] <beandog> luke-jr: :D
[00:26] <luke-jr> beandog: I plan to add test results to my bug for 0.8 soonish
[00:35] <walisser> is there a reason why I get little to no MP speedup with bframes enabled (MPEG4)
[00:42] <_hc> I'm trying to figure out how to use the new audio filters along with a video
[00:43] <_hc> I have the filters working but it only then outputs the video, I want to output both video and audio
[00:43] <_hc> and apply filters to both
[00:43] <_hc> I'm using master from git
[00:43] <_hc> and doing
[00:43] <_hc> ./ffmpeg -f lavfi -i "amovie=midr-mjpeg.mov,volume=5" -acodec copy -y /tmp/output.mov
[00:43] <_hc> that works fine
[00:46] <_hc> but the result is only audio
[00:46] <_hc> but I haven't found a way to make it keep the video
[00:46] <pasteeater> walisser: what's "MP"?
[00:57] <_hc> ah, I found it:
[00:57] <_hc> ./ffmpeg -f lavfi -i "movie=midr-mjpeg.mov" -f lavfi -i "amovie=midr-mjpeg.mov,volume=5" -acodec copy -y /tmp/output.mov
[02:05] <Shimmy> Is there an FFMPEG option to print help for specific format/codec?
[02:08] <Shimmy> I want to get some info on how to use the AMV format (which is now supported in version 0.9)
[02:38] <keeperofdakeys> Shimmy: everything uses that same options, unless it is an external library like lame or x264
[03:00] <Shimmy> Anyone knows if there is a parameter/argument to get help about specific format/codec?
[03:02] <Seph> there isn't. It's all in -h
[03:15] <Shimmy> Anyone has any idea how to use the AMV encoder included in FFmpeg version 0.9 (Harmony)?
[03:18] <Shimmy> @Seph, I output the whole help to a file, searched for my specifc format, and it doesn't have any specs, when I'm trying to use it here is what happens: http://pastebin.com/RtC3bMzC
[03:20] <Shimmy> The AMV codec is now supported in version .9 as stated here http://ffmpeg.org/#pr9
[03:20] <Shimmy> But I don't know how on earth to use it
[03:22] <Shimmy> I also tried simple ffmpeg -i sample.wmv -f amv output.amv
[03:22] <Shimmy> and it doesn't work: http://pastebin.com/46syEpUQ
[03:25] <dashcloud> just a general comment here: is there any particular reason you build with --enable-decoder ?
[03:28] <Shimmy> @dashcloud are you talking to me?
[03:28] <dashcloud> yes
[03:31] <dashcloud> I'll try to see if I can get amv to encode- but I believe you don't use -f
[03:38] <Shimmy> As per why I configured it --enable-decoder, it's because I made a batch that reads all available encoders/muxers/filter etc. and sets them all --enable-decoder=ENCODER at configure
[03:38] <dashcloud> you don't need to enable any decoders that don't have lib in front
[03:38] <Shimmy> I just wanted to make sure I have a version with all goodies possible
[03:39] <dashcloud> I'm also pretty sure that unless you happen to use --disable-* first, you'll get them all anyway
[03:39] <Shimmy> what ever
[03:39] <Shimmy> as per the AMV issue it happens anywhere
[03:40] <Shimmy> Take a look: http://pastebin.com/iVCDzGhm
[03:40] <Shimmy> This is a simple ffmpeg build with just -enable-gpl
[03:42] <Shimmy> And I'm encountering the same errors when trying to convert to AMV
[03:48] <relaxed> Shimmy: did you try a different input?
[03:49] <pasteeater> Shimmy: you could try 'ffmpeg -i input -c:v amv output.mkv' or output.avi
[03:51] <dashcloud> do you have an existing amv on hand? I found some in the samples archive, and this one has amv video (yuvj420p) and adpcm_ima_amv
[03:52] <Shimmy> @dashcloud - yes I do - check it out: http://pastebin.com/ASJH45mB
[03:55] <relaxed> hmm, just built most recent git and I'm getting " D A D  adpcm_ima_amv   ADPCM IMA AMV"
[03:56] <dashcloud> it appears there's multiple generations of amv- do you need the oldest, or can your player handle some of the newer ones?
[03:56] <relaxed> maybe it meant the amv video codec is now supported.
[03:56] <relaxed> and not the format as a whole "-f amv"
[03:57] <dashcloud> because it can generate the video (though apparently not in a straight .amv file- you need to use .avi and then re-name it)
[03:58] <relaxed> Shimmy: what does `ffmpeg -codecs 2>&1| grep ima_amv` return?
[03:59] <Shimmy> @dashcloud I don't have the player handy. I just know my output file has to stick to the working example http://pastebin.com/ASJH45mB
[04:01] <dashcloud> if you need something now, you may have to use this instead: http://code.google.com/p/amv-codec-tools/
[04:01] <Shimmy> @relaxed 1) I tried diff input - same error 2) -f amv is indeed not supported just plain
[04:02] <Shimmy> @pasteeater I tried your line and it works, but I'm not sure my hardware will play it (as mentioned, I don't have it handy)
[04:03] <relaxed> Shimmy: well, `ffmpeg -codecs` should list "D A D  adpcm_ima_amv   ADPCM IMA AMV" which means you can't encode amv audio. So I'm guess it's not a true amv format?
[04:04] <Shimmy> I believe you're right :(
[04:04] <Shimmy> I'm affraid so :'(
[04:04] <dashcloud> if you look here: http://samples.mplayerhq.hu/amv/ there's amvs with a number of other video/audio codecs in them
[04:04] <beastwick> hello dash!
[04:05] <beastwick> I now have a real problem you can possibly help me fix
[04:05] <beastwick> if you have time :D
[04:05] <dashcloud> and if you really do need that format, you'll probably have to use http://code.google.com/p/amv-codec-tools/
[04:05] <beastwick> oh sry you are helping someone already
[04:05] <dashcloud> which is old, but it does offer amv support builtin
[04:05] <relaxed> Shimmy: Don't worry, I'm sure they'll add support.
[04:05] <dashcloud> just ask your question and if someone can answer it, they will
[04:05] <Shimmy> @dashcloud thanks for the link, I'll ask my customer to check which AMVs work on his device
[04:05] <Shimmy> The question is how long it's gonna take :)
[04:06] <Shimmy> @beastwick how ya doin
[04:06] <beastwick> I am well
[04:06] <beastwick> let me see if I can sort it out solo
[04:07] <beastwick> I will put on pastebin
[04:07] <beastwick> http://pastebin.com/7MXtjFH0
[04:07] <Shimmy> My main problem is, I'm a .NET dev i built a converter using FFmpeg and then the customer asked for additional support for AMV now my whole thing is already based on ffmpeg
[04:10] <Shimmy> I found this library which managed to convert to amv with this: http://code.google.com/p/amv-codec-tools/ but since I donno C I don't know how to merge it with FFmpeg cuz it's supported input format is limited according to its supplied FFmpeg version;
[04:11] <Shimmy> Maybe I should combine it with the new FFmpeg, I have no clue how tho, I don't C at all, just the basics newbies
[04:13] <relaxed> you decode and pipe yuv4mpeg/wav from a recent ffmpeg to the old amv-codec-tools ffmpeg and encode that way. (yeah, it's horrible but would work)
[04:14] <relaxed> a temp solution until ffmpeg has total support.
[04:18] <beastwick> Would anyone be able to show me what to remove in the following script so I only grab VIDEO and not AUDIO?
[04:18] <beastwick> I am doing this off of an example
[04:19] <beastwick> http://pastebin.com/0f7GqCC3
[04:20] <beastwick> oh and would i just do -b:v 500K for 500 KB/S bit rate?
[04:25] <dashcloud> you could try this: http://pastebin.com/3AnEdKJY
[04:30] <beastwick> WOOT!
[04:30] <beastwick> it works
[04:30] <beastwick> check out twitch.tv/gamebeasttv  password is pass to see how awesome u r!
[04:30] <beastwick> now I gotta figure out audio - I don't want pulse as everythign will come in through line-in via a hardware mixer
[04:32] <beastwick> oops
[04:32] <beastwick> pls dont steal my stream lol -.-
[04:41] <beastwick> dash how can I make sure it doesn't go above 600 KB/SEC upload?
[04:42] <beastwick> just add -b:v 500K  ^ same question as above
[04:42] <beastwick> I have poor internet, and I need to make sure.
[04:59] <buhman> http://absdev.org/ffmpeg.png 
[04:59] <buhman> erm, I pushed some key-combination, and ffmpeg started giving me all sorts of bizzare hex
[04:59] <buhman> what key was that and how do I tell it to stop?
[05:00] <relaxed> buhman: hit "?" and look around
[05:21] <Shimmy> Found how to use amv codec
[05:21] <Shimmy> I'm using amv-codec-tools http://code.google.com/p/amv-codec-tools with the new ffmpeg
[05:55] <GordonFreeman> http://ivebeentrippineversince.tumblr.com/
[05:55] <GordonFreeman> some pics are mega c00l here <;
[06:50] <plushka> Hi all! Who is alive?
[06:50] <plushka> Pls need help with building
[06:56] <loon> Hello all.
[06:57] <loon> i am having an issue installing libx264 with ffmpeg 
[07:10] <burek> what kind of issue
[07:11] <burek> plushka, what's the problem
[07:12] <plushka> burek, With this bugs, at compile "ERROR: libx264 version must be >= 0.118." - This
[07:12] <burek> its not a bug
[07:12] <burek> your libx264 is old
[07:12] <burek> probably you used your package manager to install it
[07:12] <burek> ages ago
[07:12] <burek> you need to update it
[07:16] <plushka> burek, But how, if (>>> No outdated packages were found on your system.) mm?
[07:18] <burek> apt-get update
[07:18] <burek> apt-get dist-upgrade
[07:18] <plushka> gentoo 
[07:18] <burek> then make uninstall
[07:19] <burek> and git clone..
[07:19] <plushka> pls write links
[07:19] <plushka> on git 
[07:19] <burek> http://pastebin.com/LhRkXhzZ
[07:19] <burek> dont forget ldconfig
[07:20] <plushka> ld.so.conf edit? Write path to .so?
[07:21] <drv> the version in gentoo ~x86 is new enough i think, no need to build it by hand
[07:22] <plushka> error after ./configure -> make http://pastebin.com/GRngussq
[07:23] <plushka> drv, and ~64?
[07:23] <drv> yeah, same version
[07:25] <burek> only git version is new enough :)
[07:25] <plushka> That is no difference, still with his hands?
[07:26] <burek> ?
[07:26] <drv> latest ~amd64 and ~x86 version is 0.118 api
[07:26] <drv> http://packages.gentoo.org/package/media-libs/x264
[07:26] <burek> if you compile latest ffmpeg with libx264, obviously you need latest libx264
[07:26] <burek> because new features are being added all the time, and old one deprecated, etc
[07:27] <plushka> burek, git repo with x264 not compile
[07:29] <burek> you can't compile latest libx264 from git?
[07:29] <plushka> yepp
[07:29] <burek> can you please use pastebin.com, to show your command line and its output?
[07:29] <plushka> one minute
[07:31] <plushka> Is it end make.compile http://pastebin.com/7dnSsdPF
[07:33] <burek> i dont see your cmd line there
[07:34] <drv> if this is amd64, i'm not sure where the -march=i686 is coming from
[07:35] <plushka> burek, cmd line - ammm?
[07:35] <plushka> What need?
[07:37] <plushka> You need a cmd that I run?
[07:39] <burek> yes
[07:39] <burek> cmd + complete output
[07:41] <plushka> Look at http://pastebin.com/JAaYGL46 please
[07:41] <jasta> i'm trying to confirm that ffmpeg does not handle RTCP packets from a peer when ffmpeg is operating as the streaming server, is that correct?
[07:44] <burek> plushka, I can't help you without cmd+output
[07:44] <burek> i need to see what version of gcc you have
[07:44] <burek> and what kind of config gets created for your machine before compile
[07:45] <burek> so, please provide cmd+full output (not just the error)
[07:45] <burek> if you are concerned about privacy, make your paste private and expire in 10 minutes
[07:45] <plushka> burek, listen i exec  ./configure --enable-static --enable-shared --system-libx264
[07:46] <plushka> And make - done
[07:46] <plushka> what is mean?
[07:46] <burek> i don't understand you
[07:46] <plushka> Ohhh my bad english
[07:46] <plushka> Minute
[07:47] <burek> anyway, please join #videolan
[07:47] <burek> and ask your question there
[07:47] <burek> because they are developing libx264
[07:47] <burek> and they could provide more help then we could
[07:47] <jasta> hmm.  how on earth am i supposed to detect a dead peer when operating as a server of RTP/RTSP
[07:48] <jasta> im using ffmpeg for the protocol support, but i realized there's this terrible issue in my server where i will encode and send packets forever to dead peers.
[07:48] <plushka> I did ./configure --enable-static --enable-shared --system-libx264. And after that libx264 assembled correctly. Faithful if this config?
[07:48] <burek> read more on those protocols and understand how they work?
[07:48] <jasta> i would have thought RTCP could be used for this, but it _seems_ that ffmpeg discards these packets when operating as the server
[07:49] <burek> plushka, what did you configure? ffmpeg or x264?
[07:49] <jasta> burek: it's not actually the protocol that is the concern.  i'm fairly certain RTCP is used for this.  the issue is what ffmpeg does with these packets to expose this.
[07:49] <plushka> x264
[07:49] <burek> plushka, http://pastebin.com/LhRkXhzZ
[07:50] <burek> jasta, did you read the source code for rtcp support in ffmpeg?
[07:51] <plushka> burek, i know. But i add --system-libx264 in line ./configure. After that make done without error
[07:51] <burek> why would you add that if nobody told you to do so plushka ?
[07:52] <plushka> Is it correct?
[07:52] <burek> no
[07:52] <plushka> ok. 
[07:52] <jasta> burek: i did.  it very much suggests to me that ffmpeg discards RTCP packets when operating as the server.
[07:53] <jasta> i was hoping to have a developer confirm this.  there is much code dealing with RTCP packets in ffmpeg, but it all _appears_ to me to be relevant only to ffmpeg operating as an RTP/RTSP client.
[07:53] <burek> http://git.videolan.org/?p=ffmpeg.git&a=search&h=HEAD&st=commit&s=rtcp
[07:53] <burek> see who is developing that part of the code
[07:53] <burek> and try asking them
[07:53] <jasta> ok thank you
[07:54] <plushka> However, with this flag to build the library fine, then I normally set ffmpeg
[07:54] <plushka> burek, thx
[07:54] <burek> :beer:
[09:46] <burek> did any of you guys use EasyCAP? is it worth something?
[11:48] <Shimmy> Anyone with experience on the AMV format? The new 0.9 version states that it supports it but it doesn't
[11:48] <Shimmy> I tried to amv-codec-tools from here, but then it doesn't work with the new version either
[11:49] <JEEB> well, the "AMV video encoder" is there, not "full AMV bleh bleh support"
[11:49] <JEEB> and I think you got the -vcodec set
[11:49] <JEEB> anyways, dot-amv is just an avi with certain fields' values changed
[11:49] <JEEB> in the header
[11:49] <JEEB> I think I told you to try it damn it
[11:50] <JEEB> to output to avi
[11:50] <JEEB> anyways, you are getting money out of this and the format sucks, I have no obligation to help any more :P
[11:50] <JEEB> I used enough time trying to help you yesterday
[11:51] <JEEB> (note: the audio looked like normal PCM/WAV/whatever, shouldn't be hard to recreate)
[11:52] <Shimmy> @JEEB wanna get paid?
[11:52] <Shimmy> And do this job for me?
[11:55] <mystica555_> audio is ima-adpcm with some slight modifications
[11:55] <mystica555_> according to wiki
[11:55] <JEEB> yeah
[11:57] <Shimmy> how can i contact u
[11:57] <Shimmy> u forgot that the adpcm_ima_amv is not supported for encoding
[11:58] <Shimmy> got priv
[11:58] <JEEB> the audio can be handled other ways if explicitly needed
[11:59] <Shimmy> even it's for hardware player?
[12:00] <speedo666> hi guys
[12:00] <speedo666> little question,  i want to extract frames from file
[12:00] <speedo666> but the sequence of png filename is always starting with 0
[12:00] <JEEB> more like with using different things to do the audio and the video, but I'm looking at how it was implemented in the "dead for long" google code project
[12:01] <speedo666> can i change that to othert number? starting seq diff than 0
[12:01] <speedo666> ?
[12:01] <MT`> hey, I'm trying to do a x264 2 pass encode, but it seems that whatever profile/preset/tune I choose has crf enabled on the first pass. I can't find what has happened to presets like 'fast_firstpass' - any ideas?
[12:01] <JEEB> MT`, presets are now used from libx264
[12:02] <JEEB> http://mewiki.project357.com/wiki/X264_Settings#preset
[12:02] <JEEB> and the default rate control mode in libx264 has now been crf for a while
[12:02] <MT`> thanks, I can use google too :(
[12:02] <JEEB> so you need to set a bitrate if you explicitly need it
[12:03] <JEEB> the presets shouldn't affect your rate control mode
[12:03] <MT`> I dont want to set a bitrate, I just want to do a first pass to achieve a better quality encode
[12:03] <JEEB> you are making no sense
[12:03] <MT`> ok, 6 months ago, with x264/ffmpeg, I could do a 2 pass x264 encode
[12:03] <JEEB> two pass encoding gives pretty much the same quality as one pass crf, the difference being in what you exactly want to achieve
[12:04] <JEEB> two-pass is for achieving a certain file size (bitrate)
[12:04] <MT`> now I cannot, the ffmpeg profile that did it 6 months ago  is gone
[12:04] <JEEB> you can still do two-pass encoding
[12:04] <MT`> and there is no equivalent firstpass profile in x264 that does the same
[12:04] <JEEB> you have to set that you are doing a first pass?
[12:05] <MT`> yes, I pass -pass 1 to ffmpeg
[12:05] <JEEB> can you pastebin your exact ffmpeg command lines and their output?
[12:05] <JEEB> also, if you just want quality
[12:05] <JEEB> you don't need two-pass
[12:05] <JEEB> just use the highest crf value that looks good >_>
[12:06] <MT`> hmm
[12:06] <JEEB> also, I am commenting like this because you clearly are either confused or not sure what you are doing as I read from "<MT`> I dont want to set a bitrate, I just want to do a first pass to achieve a better quality encode"
[12:07] <MT`> that is so condescending
[12:07] <MT`> ...
[12:07] <JEEB> because two passes are only needed in the case of "I want the best possible quality while easily setting a file size (bitrate) to which the encode will be done."
[12:08] <JEEB> (you could achieve the same result by finding the crf value that gives you that size, but let's just say that you usually don't want to do X test encodes to find it :) )
[12:08] <JEEB> "I just want quality" => find the highest crf value that still looks good for you, use it. no need to set passes or anything, just set crf and preset and you're off.
[12:09] <MT`> 2pass allows the encoder to base early frames on the knowledge garnered in the earlier run
[12:09] <MT`> crf does the same thing by averaging over a much smaller timespan
[12:09] <MT`> I want the highest quality with the lowest possible bitrate, which 2 pass should give me
[12:10] <JEEB> the thing is, when you are not limiting your encode bitrate-wise, you don't really need the data you get from a first pas
[12:10] <JEEB> *pass
[12:10] <JEEB> I wish there was D_S or someone here who could argue you down, but unfortunately they are not here :)
[12:10] <JEEB> "I want my encode to end up exactly of size X/bitrate X and look good" => two-pass bitrate encoding
[12:11] <MT`> OK, let me ask a different question: does ffmpeg only allow me to configure x264 using profile/preset/tune, or can I explicitly provide encoding options
[12:11] <JEEB> with bitrate encoding we have to know which parts of the clip need more bitrate, and which less to make an optimized encode when hitting a certain size
[12:12] <JEEB> you can provide libx264 settings, yes
[12:12] <JEEB> While with crf you are globally just setting a certain rate factor
[12:12] <JEEB> thus there is no need to optimize anything
[12:12] <JEEB> you are optimizing for a different result :3
[12:12] <MT`> JEEB: look, you're clearly sold on this, you're not convincing me
[12:12] <MT`> i dont want to waste your time
[12:12] <JEEB> ok, you can ask Dark_Shikari
[12:12] <JEEB> or any other dev
[12:13] <JEEB> and they will say the same
[12:13] <MT`> fine
[12:13] <JEEB> that if you just care about quality, use crf
[12:13] <JEEB> and if you want to hit (in a relatively short period of time) a certain bitrate with good quality
[12:13] <JEEB> then 2pass
[12:13] <MT`> I dont just care about quality
[12:13] <MT`> thats the entire point :(
[12:14] <JEEB> well, you are caring about compression, and I don't see crf being any worse off than bitrate-based
[12:14] <JEEB> the same algorithm is just being optimized for a different thing
[12:18] <JEEB> because you can hit the same file size with crf and get same/better by 0.x%/worse by 0.x% results, but that just doesn't make sense if you just want to hit a certain file size, now does it (in my testing with SSIM and actually munching through crf values crf came up better, but as the troll there said it _can_ be 0.blehbleh% better or worse, not a notable difference when you have an actual end result you want to achieve)?
[12:21] <JEEB> and you are free to check whatever I am saying on any of the developers of x264
[12:21] <JEEB> :3
[14:08] <Shimmy> Any news?
[14:09] <GordonFreeman> do you like fractals?
[14:10] <Shimmy> u askin me?
[14:45] <Ceida> hi , i m having problem on webm codec , i want to reset image decode to "all image" to clean all. vpx_codec_decode(&decoder, compressed_video_buffer, sizeof(compressed_video_buffer), 0, 0) <= when this function fails , i want to reset codec , how can i reset ? someone know?
[18:54] <TimN> drawtext plus timecode whoop whoop, thanks ubitux:
[19:04] <ubitux> TimN: woot at least one user ;)
[19:08] <TimN> ubitux: one *big time* user, and given the possible confusion over the syntax with the use of : in the timecode I have added an example to https://ffmpeg.org/trac/ffmpeg/wiki/FilteringGuide
[19:09] <ubitux> ah nice one
[19:09] <ubitux> maybe you could even send a patch to ffmpeg-devel to add an example in the documentation?
[19:10] <ubitux> but well not that important
[19:10] <ubitux> otoh, you may change your example to deal with ntsc material and drop frame
[19:10] <ubitux> ’ ';'
[19:11] <ubitux> it will also show you can set the rate in rationnal form
[19:11] <ubitux> 30000/1001
[19:11] <TimN> I am working on patches to the filter doc as discussed earlier, its just that other work has been rather pressing recently, and include NTSC
[19:11] <ubitux> anyway, thanks :)
[19:11] <ubitux> what were you using so far for the timecode btw?
[19:12] <TimN> Well for odds and ends Telestream..... :( but I have just been asked about doing several thousand hours worth so your patch is timely....
[19:13] <ubitux> i meant how did you burn it before :)
[19:14] <TimN> You can burn in using Telestream, or using a hardware box on the incoming stream....
[19:14] <ubitux> oh ok
[19:52] <symtab> hi
[19:53] <symtab> i'm using this: http://pastebin.com/LG4jxY3e
[19:53] <symtab> to convert my video to a format supportted on most mobile phones
[19:53] <symtab> however recent ffmpeg versions dont have the baseline preset
[19:53] <symtab> any ideas how i can achieve exactly the same with current ffmpeg version 0.9
[19:53] <symtab> ?
[19:55] <JEEB> use libx264 features?
[19:55] <JEEB> -profile baseline -level 3 -preset <your libx264 speed/compression preset>
[19:55] <JEEB> http://mewiki.project357.com/wiki/X264_Settings#profile
[19:56] <JEEB> ^ the documentation'ish thing for the x264 settings
[19:56] <JEEB> in ffmpeg you can at least control the main ones
[19:56] <JEEB> profile, preset, level
[19:59] <symtab> let me try that
[19:59] <symtab> thank you
[20:03] <symtab> i get some errors: http://pastebin.com/wVLPFP6S
[20:03] <symtab> any ideas?
[20:03] <symtab> [NULL @ 0x98a4bc0] [Eval @ 0xbfbc4908] Undefined constant or missing '(' in 'baseline'
[20:04] <JEEB> you make me sad with that command line :)
[20:05] <symtab> i'm sorry
[20:05] <symtab> :)))
[20:05] <JEEB> you can drop most of the x264-related things
[20:05] <symtab> its the only thing that works on all mobile phones i've tested
[20:05] <MT`> symtab: the problem is that it is attempting to apply the profile to libfaac as well
[20:05] <MT`> symtab: you need to prefix it so that it is just for video codec
[20:05] <JEEB> symtab, you don't need to set them manually any more
[20:05] <symtab> i found with google to use -vprofile
[20:05] <symtab> using that now
[20:06] <JEEB> profile and level set those things now
[20:06] <JEEB> lemme see
[20:06] <symtab> i will find time to test without all the options (but its not easy, because i only have android 2.3)
[20:06] <JEEB> I assure you
[20:06] <MT`> the exact same thing hit me yesterday
[20:06] <symtab> -vprofile works
[20:06] <JEEB> -vprofile baseline -level 3 works
[20:06] <MT`> -profile:v would also work, and is the new cool afaict
[20:06] <symtab>     Stream #0:0(eng): Video: h264 (Constrained Baseline) (avc1 / 0x31637661), yuv420p, 320x176 [SAR 44:45 DAR 16:9], 372 kb/s, 25 fps, 25 tbr, 25 tbn, 50 tbc
[20:06] <symtab> PERFECT
[20:07] <JEEB> sigh...
[20:07] <symtab> -level 1.3
[20:07] <symtab> -level 3 wont work on all phones
[20:07] <symtab> only new ones
[20:07] <JEEB> uhh
[20:07] <JEEB> level 3 has worked on any phone that can handle baseline during the last two-three years
[20:07] <JEEB> but sure
[20:07] <JEEB> you can drop the level lower if you want
[20:07] <JEEB> the one thing I'm sure about is that you don't need to play with
[20:07] <JEEB> -flags +loop -cmp +chroma -partitions 0 -subq 1 -trellis 0 -refs 1 -coder 0 -me_range 16 -g 300 -keyint_min 25 -sc_threshold 40 -i_qfactor 0.71
[20:08] <JEEB> -rc_eq 'blurCplx^(1-qComp)' -qcomp 0.6 -qmin 10 -qmax 51 -qdiff 4
[20:08] <JEEB> these
[20:08] <JEEB> you can keep your maxrate and bufsize
[20:08] <symtab> so drop everything except maxrate and bufsize?
[20:09] <symtab> everything related to x264 
[20:09] <JEEB> yes, because the libx264 wrapper in ffmpeg is no longer stupid :3
[20:09] <JEEB> you have more chances of ruining stuff than making stuff work :D
[20:09] <JEEB> (by setting that stuff manually)
[20:10] <symtab> ok
[20:10] <symtab> thank you
[20:10] <JEEB> note: the braindeadness of the ffmpeg's libx264 wrapper was what kept me from using libx264 with it for a long time
[20:10] <JEEB> although... I still don't use it if I can use x264 itself
[20:13] <JEEB> and I think it was trying to apply the -profile to the audio encoder because it was set after it :)
[20:14] <symtab> yes
[20:14] <symtab> using -vprofile works correctly
[20:15] <JEEB> ffmpeg -i 15.mp4 -vcodec libx264 -preset slow -profile baseline -level 1.3 -b 375k -maxrate 10M -bufsize 10M -threads 0 -aspect 16:9 -s 320x176 -acodec libfaac -ac 2 -ab 128k -ar 44100 15.mobile.mp4
[20:15] <JEEB> this should work as well
[20:16] <JEEB> (also, libx264 used to be less nice to be used as well, f.ex. the level flag and maybe even the profile flags were just setting flags in the stream, not actually keeping settings inside those limitations)
[20:16] <JEEB> been a year or more now that it started controlling those things :3
[20:20] <MT`> JEEB: position is irrelevant, if you dont specify a stream specifier for a profile, it applies it to all streams (well, certainly a+v)
[20:20] <JEEB> ah
[20:20] <JEEB> so they changed that
[20:21] <MT`> i tried all the variations working that one out ;)
[20:21] <JEEB> I still remember when profile and preset very pretty much libx264-specific things
[20:21] <JEEB> IIRC
[20:21] <JEEB> and vprofile or whatever that loaded those files was for that
[20:21] <MT`> yep
[20:22] <JEEB> shows how much I use avconv/ffmpeg for encoding
[20:22] <MT`> I dont know much about the history of x264, but I think the profiles were ffmpeg, or at least the profiles I was using
[20:22] <JEEB> the first profiles were
[20:22] <MT`> and they simply specified different x264 flags
[20:22] <JEEB> yes
[20:22] <JEEB> then libx264 got its own presets
[20:22] <JEEB> (note: in H.264 profiles are pretty much feature sets)
[20:23] <MT`> yep, certainly makes things easier to grok
[20:23] <JEEB> so ffmpeg then got a setting (-preset) to load the libx264 presets IIRC
[20:23] <JEEB> and then a few other settings IIRC
[20:23] <JEEB> and thus the old preset files were pretty much made unneeded
[20:24] <MT`> still need them for libvpx, but yeah
[20:24] <MT`> and of course, the ffmpeg settings sometimes had different names to the ones in x264
[20:25] <JEEB> yeah
[20:25] <JEEB> +1 reason why I would've not used ffmpeg for libx264 encoding back then
[20:47] <Seph> Anyway I can get a list of all the libx264 presets?
[20:49] <JEEB> http://mewiki.project357.com/wiki/X264_Settings#profile
[20:49] <JEEB> presets are under the profiles
[20:49] <JEEB> also you can IIRC get them with x264 --fullhelp | grep "placebo"
[20:50] <JEEB> because it should grab the line with placebo on it, which is the most "you don't really want to use this for anything else but benchmarks" preset
[20:51] <Seph> yeah just found x264 --fullhelp listed under ffmpeg -h
[20:54] <Seph> Any reason you suggested using -threads 0?
[21:03] <JEEB> Seph, unless that is default in the ffmpeg wrapper now as well
[21:03] <JEEB> I guess it is since it's the libx264 default
[21:03] <JEEB> d'oh
[21:07] <Seph> does it disable threading or allow libx264 to choose? I've always had threads set to 8 as that's how many (virtual) cores my cpu has
[21:08] <JEEB> --threads 0 in libx264 context means "auto"
[21:09] <Seph> excellent then
[21:10] <Seph> I just updated ffmpeg and libx264 so I figured I should probably get rid of my old manual settings and just use a preset
[21:10] <Seph> next up is getting ffmpeg to support fonts and subtitles :/
[22:05] <rocketmagnet> hi all, os there the oposite switch to -y ????
[22:05] <Ave> dont use -y and it wont overwrite without asking
[22:05] <rocketmagnet> is there
[22:07] <pasteeater> TimN: what does "(change the : to a ; before the frame count)_________________________________________________________" mean?
[22:20] <cbsrobot> pasteeater: I think TimN tries to draw a line in asciiart to the ; 
[22:22] <pasteeater> ah, i see. duh.
[22:38] <dowdle> Ok, so -ss will let me jump to a particular place to start encoding.  Is there (what is the) flag that will let me start at the beginning of a video but stop encoding before the end of the video?
[22:39] <dowdle> Basically I have a video that is two videos in one, and I want to chop it into two.  I can encode the second half easily with -ss but I can't seem how to encode the first half and stop where the second one begins.
[22:50] <pasteeater> dowdle: -t option
[22:51] <pasteeater> ffmpeg -i input -ss 10 -t 20... will skip first 20 seconds and make output 10 seconds long.
[22:52] <dowdle> pasteeater: Thanks.  I'll give that a try.
[23:00] <cbsrobot> pasteeater: it's other way round -no ?
[23:01] <cbsrobot> skip first 10 seconds and make output 20 seconds long
[23:13] <Seph> Have I misunderstood something, shouldn't ffmpeg support ssa/ass subtitles?
[23:13] <Seph> And mp4 as well.
[23:14] <Seph> http://pastie.org/private/0larhdufpbenhei3acftg <- see line 55 and 78
[23:16] <pasteeater> cbsrobot: yes, i transposed my numberos
[23:21] <alpha--> hello
[23:21] <alpha--> I'm converting some .flac files to .mp3 files using this command: ffmpeg -i in.flac -vn -ab 192k -acodec libmp3lame out.mp3
[23:21] <alpha--> so the output bitrate is set to 192 kbit/s
[23:22] <alpha--> but when I play the resulting file with mplayer, it detects it as AUDIO: 44100 Hz, 2 ch, s16le, 48.0 kbit/3.40% (ratio: 6000->176400)
[23:22] <alpha--> 48 kbit/s
[23:22] <alpha--> is my command wrong, or is mplayer guessing wrong or what?
[23:25] <pasteeater> alpha--: you can find out the average bitrate of the file manually
[23:26] <pasteeater> (3.25 MB * 8192 [converts MB to kilobits]) / 240 seconds = ~110 kilobits/s
[23:26] <pasteeater> for example
[23:26] <ubitux> Seph: please also copy the command line you used&
[23:26] <alpha--> hmm 1 sec
[23:26] <ubitux> Seph: it looks like you're remuxing, ass filter is usable this way: -vf ass=/path/to/file.ass
[23:27] <ubitux> but maybe you actually want a remuxing&
[23:27] <pasteeater> does mp4 support ass?
[23:27] <jbwiv> I've tried a few times to reduce a video size on ubuntu with ffmpeg by reducing the bitrate, but no matter what i use the file size ends up growing. based on past irc discussions, I believe this many have something to do with ubuntu's build or use of libav. are there any distros out there which package a full-blown tried and true non-libav version of ffmpeg?
[23:27] <pasteeater> jbwiv: ubuntu if you compile ffmpeg.
[23:27] <pasteeater> http://ubuntuforums.org/showthread.php?t=786095
[23:28] <pasteeater> also, use a pastebin site to show your ffmpeg command and the complete console output
[23:28] <alpha--> yes pasteeater, it adds up correctly that way
[23:29] <alpha--> I also discovered another problem, mplayer can not detect file duration correctly
[23:29] <alpha--> but vlc ... something is wrong with my mp3
[23:29] <alpha--> vlc can*
[23:29] <jbwiv> pasteeater, ok, yeah, did that before but all the files are at home. I'll do that when I get there tonight. thanks
[23:30] <pasteeater> alpha--: are you using an old mplayer (or ffmpeg too)?
[23:30] <alpha--> ffmpeg 0.7.5
[23:30] <alpha--> MPlayer SVN-r32624-4.4.4
[23:39] <Seph> ubitux: sorry, not trying to mux it in
[23:39] <Seph> I want the subtitle embedded into the mp4 container so that it's a soft subtitle
[23:39] <ubitux> yes that's what i meant
[23:39] <ubitux> no one can guess your command line
[23:39] <ubitux> so no one can help you fixing it
[23:40] <Seph> http://pastie.org/3012862 <- this is the command line
[23:40] <Seph> yeah I thought it was included in the ffmpeg output since the configure argument is.
[23:40] <Seph> the subtitle is in the $base (mkv) file being read.
[23:46] <Seph> hmm no, it actually seems mp4 does not support ass subtitles. Have to be converted to timed text format
[23:48] <walisser> I am trying to make MPEG 4 ASP Level 5, compliant file, what rc-buffer-size should be used? Wikipedia says 1792 kb but I can't find a combination of bitrate flags that will work with that
[00:00] --- Wed Dec 14 2011