[Ffmpeg-devel-irc] ffmpeg.log.20120321

[burek]

[00:00] <crispycritter> (it was mostly to adapt to a different environment; I didn't intend to change semantics any.)
[00:02] <crispycritter> (I also get two perfectly good output frames; it fails while working on frame 3.)
[03:51] <Stevoman> Anybody out there?
[04:03] <Stevoman> Hey all, I am seeking some help with trying to encode to mp4/aac. My code works up to the point where I call avcodec_encode_audio2 (line 62), at which point my calling
[04:03] <Stevoman> code throws an AccessViolationException (Attempted to read or write protected memory). If someone could tell me what I am doing wrong, I would very much appreciate the help. Code can be found here: http://pastebin.com/YZsqumyf
[04:35] <jhericurl> is flv a good video/audio container?
[05:51] <jhericurl> what is the command to keep the same video but reencode PCM audio to  aac
[06:13] <emelo> Hi... anybody ???
[06:13] <emelo> I have a problem
[06:13] <emelo> if I do this:
[06:13] <emelo> ffmpeg -i rtsp://admin:admin@192.168.99.146/11 -vcodec copy -y -r 25 algo.mp4
[06:14] <emelo> it works..
[06:14] <emelo> if I do this:
[06:14] <emelo> ffmpeg -i rtsp://admin:admin@192.168.99.146/11 -vcodec copy -y -r 25 algo.ts
[06:14] <emelo> I get this error:
[06:14] <emelo> [mpegts @ 0x666660] H.264 bitstream malformed, no startcode found, use the h264_mp4toannexb bitstream filter
[06:14] <emelo> av_interleaved_write_frame(): Invalid data found when processing input
[06:15] <emelo> if.. I add the filter, I get this:
[06:15] <emelo> ffmpeg -i rtsp://admin:admin@192.168.99.146/11 -vcodec copy -y -r 25 -vbsf h264_mp4toannexb algo.ts
[06:15] <emelo> Failed to open bitstream filter h264_mp4toannexb for stream 0 with codec copy: Invalid argument
[06:15] <emelo> [mpegts @ 0x666660] H.264 bitstream malformed, no startcode found, use the h264_mp4toannexb bitstream filter
[06:15] <emelo> av_interleaved_write_frame(): Invalid data found when processing input
[06:24] <grepper> jhericurl: something like -c:v copy -c:a ac3 -b:a 96k -ar 22050
[06:24] <grepper> I think, not too used to the new syntax yet
[06:24] <jhericurl> ok
[06:25] <grepper> er
[06:26] <grepper> sorry, -c:a libfaac
[06:28] <jhericurl> does ffmpeg support apple-aac encoder?
[06:35] <grepper> no idea what that is precisely, is it a subset of aacplus ?
[06:35] <grepper> ffmpeg -codecs 2>&1 |grep aac
[06:38] <jhericurl> coreaac or something
[06:43] <grepper> core audio ?  That looks like an api rather than simply a codec
[06:48] <jhericurl>     AAC (CoreAudio?):
[06:48] <jhericurl>         Mac OS X users have the option to use Apple's Core Audio AAC encoder.
[06:48] <jhericurl>         It's known to be much better quality and is the default option for Mac users.
[06:58] <grepper> so the aac encoder that the core audio api uses
[06:58] <grepper> ffmpeg -codecs 2>&1 |grep aac
[06:59] <grepper> will show you the aac types your ffmpeg supports, choose one :)
[07:34] <cbsrobot> jhericurl: on mac use afconvert to use the apple aac encoder
[07:34] <jhericurl> i cannot use ffmpeg?
[07:34] <cbsrobot> use ffmpeg to add the audio to the video
[07:35] <jhericurl> what about on pc?
[07:35] <cbsrobot> not if you want to use the apple aac encoder
[07:35] <cbsrobot> pc - what is that :-)
[07:35] <jhericurl> pc versus mac
[07:35] <cbsrobot> no idea - but you could use the nero aac encoder
[07:36] <cbsrobot> it's also command line
[07:37] <jhericurl> so what do i type in ffmpeg first?
[07:38] <cbsrobot> what is in your source file ?
[07:38] <cbsrobot> in what you want to convert it ?
[07:39] <cbsrobot> else you could also try libva-aac from http://sourceforge.net/projects/opencore-amr/files/
[07:40] <cbsrobot> *libvo-aac
[07:40] <jhericurl> what is the command to keep the same video but reencode PCM audio to  aac
[07:41] <jhericurl> source.mov
[08:04] <cbsrobot> jhericurl: ffmpeg -i source.mov -vn -c:a pcm_s24le output.wav
[08:05] <cbsrobot> or pcm_s16le
[08:05] <cbsrobot> depends on the input - but as you do not want to pastebin it i don't know
[08:05] <cbsrobot> then run it through neroaac
[08:05] <cbsrobot> and put all together with
[08:07] <cbsrobot> ffmpeg -i sourve.mov -i neroaacoutput.aac -c copy -map 0:0 -amp 1:0 finaloutput.mov
[08:07] <cbsrobot> *source.mov
[09:57] <tangobravo> Hello all
[09:58] <tangobravo> I'm using ffmpeg to do some work with theora
[09:58] <tangobravo> and there is a bit of odd behaviour going on with duplicate frames
[09:59] <tangobravo> I'm still doing some testing, but it seems to me HEAD now generates correct theora output, with 0-byte packets for duplictes
[10:00] <tangobravo> Decoding in my application using libogg and libtheora now correctly sees them as duplicates
[10:01] <tangobravo> but then converting to something like png with ffmpeg doesn't get the right timestamps
[10:15] <tangobravo> in an older ffmpeg (before ffmpeg.c switched to encode2) the 0-byte packets from the encoder never made it into the stream
[10:15] <tangobravo> that messed up my decoding code, but somehow ffmpeg still knew there were duplicates
[11:30] <rly> Hi, why do you make basic mistakes that only a rookie would make in crystalhd?
[11:31] <rly> FYI, the latest git 'intended for users' doesn't actually build.
[11:31] <rly> And only because you redeclare a standard type...
[11:31] <rly> Why can't you just write in C instead of in pseudo-C?
[11:33] <juanmabc> i tend to reply "do it yourself, dude" :DDD
[11:33] <JEEB> Sounds like fun, and it seems through the git logs that a single guy was responsible for the whole crystalhd thing.
[11:33] <rly> Oh, great. Even your release doesn't work.
[11:33] <rly> Also because of that crystalhd thing that I don't even want to have.
[11:33] <JEEB> I'd recommend going over to the trac issue tracker
[11:33] <rly> Which compiler is supported for your releases?
[11:33] <JEEB> and herping a derp over there
[11:34] <Mavrik> rly: there's a reason "stable" builds exist
[11:34] <rly> Mavrik: this _is_ the stable build.
[11:34] <rly> Mavrik: and it doesn't work.
[11:34] <Mavrik> which one?
[11:34] <JEEB> I would think that gcc from version 4.2 upwards or so are recommended
[11:34] <JEEB> check what kind of things compile fine on fate
[11:34] <rly> Mavrik: "Freedom"
[11:34] <JEEB> http://fate.ffmpeg.org/
[11:35] <rly> I am using 4.4.5.
[11:35] <JEEB> gcc 4.4.5 (Debian 4.4.5-8) seems to have compiled just fine but it could be that it was configured without whatever you're trying to use
[11:35] <JEEB> (could be checked in fate)
[11:36] <JEEB> anyways, I recommend you calm down, take the issue onto the trac issue tracker and then herp a derp on #ffmpeg-devel
[11:36] <JEEB> (after you have reported it on trac)
[11:37] <JEEB> also, crystalhd was a hardware decoder thingy?
[11:37] <JEEB> --disable-hwaccels would disable all of them in configure
[11:37] <JEEB> (and there was IIRC ways to enable/disable specific hwaccels)
[11:38] <rly> /usr/include/libcrystalhd/bc_dts_types.h:99: error: conflicting types for uint64_t
[11:38] <rly> /usr/include/stdint.h:56: note: previous declaration of uint64_t was here
[11:39] <rly> It just happily redeclares that standard type judging from the error message.
[11:39] <JEEB> that sounds like the crystalhd library is actually in the wrong there o_O
[11:39] <JEEB> unless ffmpeg contains "libcrystalhd"
[11:40] <Mavrik> yeah, that sadly happens quite commonly - outside libs breaking builds
[11:40] <JEEB> as far as I can see, ffmpeg can use libcrystalhd
[11:40] <Mavrik> FATE doesn't test those
[11:40] <rly> Mavrik: you can get rid of that, you know?
[11:40] <Mavrik> ?
[11:40] <rly> Mavrik: just detect those situations and say "Looks like external library X has idiots on their team. Do you want to send them an e-mail saying that they should harden the fuck up? Yes/No?"
[11:41] <Mavrik> yeah, send that to the actual ffmpeg devs
[11:41] <rly> I think it is better if you have so many dependencies to import those source trees.
[11:41] <JEEB> not.. .really
[11:41] <rly> Or to only guarantee certain versions.
[11:42] <rly> So, you say 'our software works with exactly this combination of versions'.
[11:42] <JEEB> I mean, the libcystalhd project seems to be a semi-active project
[11:42] <JEEB> yeah
[11:42] <JEEB> version checks are very possible
[11:42] <JEEB> someone just has to implement them
[11:42] <JEEB> (given libcrystalhd has any way of showing its version)
[11:42] <rly> This is one of the very first things a project should do.
[11:42] <Mavrik> most of those libs don't have stable version numbering
[11:42] <rly> Mavrik: they have a version control system, right?
[11:42] <Mavrik> and alot of them aren't really active projects (see libaacplus for e.g.)
[11:43] <Mavrik> rly: not all of them
[11:43] <rly> Mavrik: all of those 'problems' have solutions.
[11:43] <JEEB> http://git.linuxtv.org/jarod/crystalhd.git
[11:43] <rly> The very notion of running all those tests on some test server is nice, but it's something users can report too.
[11:43] <JEEB> this seems to be thelibcrystalhd's git repo
[11:43] <Mavrik> rly: of course they do
[11:43] <Mavrik> they're just low priority
[11:43] <rly> You just need to integrate a messaging system in the build.
[11:44] <rly> Naturally, you first ask whether you can make a connection before you send something like that.
[11:44] <rly> But again, if you take external dependencies (usually already a bad idea) then you need to be prepared to setup infrastructure.
[11:45] <JEEB> anyways, I think there are currently two ways for you to fix your compilation: A) disable libcrystalhd B) try updating libcrystalhd
[11:45] <JEEB> unless you are wishing to add more checks into place
[11:45] <rly> JEEB: how can I disable libcrystalhd?
[11:45] <JEEB> it's a hwaccel, right?
[11:45] <JEEB> <JEEB> --disable-hwaccels would disable all of them in configure
[11:45] <JEEB> <JEEB> (and there was IIRC ways to enable/disable specific hwaccels)
[11:45] <rly> JEEB: I don't think it is counted as such.
[11:45] <JEEB> oh
[11:46] <Mavrik> hmm, it's not enabled by default though
[11:47] <rly> Mavrik: I am not enabling it.
[11:47] <rly> Mavrik: just ./configure and that's it.
[11:48] <Mavrik> ./configure --disable-crystalhd
[11:48] <Mavrik> there.
[11:48] <rly> The build system for libcrystalhd is also not really advanced.
[11:49] <Mavrik> there.MAINTAINERS:  crystalhd.c                           Philip Langdale
[11:49] <Mavrik> sending him a mail about your problem could do some good
[11:50] <rly> I think he fixed it in his git version.
[11:50] <rly> Perhaps we should get rid of this whole 'packaging' stuff, because its latency is just way too large.
[11:50] <JEEB> packages in distros are good for some things, less good for other
[11:51] <JEEB> multimedia tends to crawl behind usually
[11:51] <rly> Good package systems can be built, but there is just no money in it.
[11:51] <JEEB> also, crystalhd getting enabled by default is pretty funky
[11:51] <rly> I would pay 500 euros to get something decent built.
[11:52] <JEEB> esp. if there are distros that ship borked versions
[11:52] <rly> Except it requires another 1000 people to do the same.
[11:52] <rly> The rest of the build went fine (few warnings, though).
[11:53] <rly> I always compile with -Werror
[11:53] <JEEB> but all I can tell is that you can poke the guy who is maintaining the thing in ffmpeg and ask if certain versions can be identified in any way
[11:53] <rly> (and pendantic, etc.)
[11:53] <rly> pedantic*
[11:53] <rly> Especially for binary poking stuff you don't want to have off-by-one errors.
[11:54] <rly> Essentially ffmeg is a huge attack vector to put it in security terms.
[11:54] <rly> ffmpeg*
[11:54] <JEEB> well, unfortunately there are not only real warnings, but some misunderstandings in ffmpeg as well -- depending on the warning.
[11:54] <JEEB> (of course there's also a plenty of real warnings)
[11:55] <Mavrik> hmm
[11:55] <rly> The use of tarballs in the age of dvcs also escapes me.
[11:55] <Mavrik> I still can't reproduce the crystalhd automatic config
[11:57] <Mavrik> the problem with version checking is when you have projects like x264
[11:57] <Mavrik> which don't use version numbering
[11:57] <Mavrik> and you really don't want to keep people on lower versions
[12:01] <JEEB> x264 actually has API versions
[12:01] <JEEB> or ABI, or whatever
[12:01] <JEEB> so you can actually check for older versions if they are too old
[12:01] <JEEB> (or too new if you only support certain ranges)
[12:02] <JEEB> not all libraries do that
[12:03] <JEEB> http://git.videolan.org/gitweb.cgi?p=x264.git;a=blob;f=x264.h;h=efc61bc085a902ffff42ab3ffdbb1b6f1aec13b9;hb=HEAD#l44 <- this header line sets the API/ABI version
[12:03] <Mavrik> mhm, that's the problem - checking has to be one on library interface version
[12:03] <Mavrik> er, I mean on the library interface
[12:03] <Mavrik> and I'm not sure if the build system for ffmpeg supports that in an elegant way
[12:03] <JEEB> you're inluding x264.h
[12:04] <JEEB> in libx264
[12:04] <JEEB> (the libavcodec/libx264.c/h)
[12:04] <Mavrik> JEEB: yeah
[12:04] <JEEB> of course configure-time checks could be done as well
[12:04] <JEEB> I actually think configure might even have a check for libx264 alreadyt
[12:04] <Mavrik> JEEB: as I said, I don't know autotools enough to know if there's an easy way to produce a meaningful message for that
[12:04] <JEEB> hint: ffmpeg is not autotools
[12:04] <JEEB> it's just a shell script
[12:06] <Mavrik> *shrug* Not touching shell scripting with 10-foot pole
[12:06] <JEEB> :D
[12:06] <JEEB> I actually have become to like the x264 configure
[12:06] <JEEB> can't say I've read much of ffmpeg's
[12:06] <JEEB> both use shell scripts for configure
[12:07] <JEEB> and yeah, there's a check for libx264
[12:07] <JEEB> http://git.videolan.org/?p=ffmpeg.git;a=blob;f=configure;h=c6a4d17107943eb65d81535253b326e8c6ddf8a4;hb=HEAD#l3190
[12:07] <JEEB> "X264_BUILD >= 118"
[12:10] <JEEB> disabled crystalhd || check_lib libcrystalhd/libcrystalhd_if.h DtsCrystalHDVersion -lcrystalhd || disable crystalhd <- all that crystalhd has
[12:56] <tangobravo> Here's a test case for theora duplicate handling
[12:57] <tangobravo> Create 3 pngs for testing: convert -size 128x128 xc:red xc:green xc:blue orig/%d.png
[12:58] <tangobravo> ffmpeg to convert to a higher framerate, again using pngs so we can see the expected output
[12:58] <tangobravo> ffmpeg -r 1 -i orig/%d.png -r 25 increase-framerate/%d.png
[12:59] <tangobravo> Interestingly the first output frame is red, then 24 greens, then 25 blue, but let's just assume that's a small bug with the first frame when changing framerate
[13:00] <tangobravo> (adding more frames correctly puts 25 copies of each into the output)
[13:01] <tangobravo> now create a 25 fps theora file: ffmpeg -r 1 -i orig/%d.png -r 25 theora.ogv
[13:01] <tangobravo> and convert that to png with: ffmpeg -i theora.ogv ffmpeg-theora/%d.png
[13:02] <tangobravo> now the first frame is red, but then follow 13 green and the rest blue
[13:16] <tangobravo> My decoder code using libogg and libtheora decoees it correctly
[13:16] <tangobravo> I've dumped out the key information from the ogg pages and packets and it seems ffmpeg generates a valid stream
[13:17] <tangobravo> I think the issue with ffmpeg decoding it is due to the 0-byte packets, which represent duplicate frames in theora, not being correctly handled
[13:54] <Prirawien> Hi, I want to upload some audio files on Youtube, I try to encode them with "ffmpeg -i audiofile.mp3 -vn videofile.wmv", but YT say thar the video is incorrect
[13:54] <Prirawien> How can I do this ?
[13:55] <Tjoppen> I'm fairly sure youtube takes mp3
[13:55] <Tjoppen> ah, you might need some video to go with it
[13:56] <Tjoppen> just toss a background image in via some nle
[13:56] <Prirawien> Google say that we have ton convert audio file to video, but they give an advice for Windows and Mac, not unix
[13:56] <Prirawien> to*
[13:56] <Tjoppen> you could add -i image.jpeg and fiddle around a bit I think
[13:57] <Tjoppen> remember to add a large image you youtube encodes it with high audio bitrate
[13:57] <Tjoppen> if you have a 1920x1080 background you'll get better audio than if you have say 320x240
[13:58] <Prirawien> Ok, thanks you, I try
[14:04] <Prirawien> Tjoppen: I try "ffmpeg -i audiofile.mp3 -i myimage.jpg -vn videofile.wmv" but it does not work, YT always say that the video is invalid
[14:06] <ubitux> you're asking for a video without video
[14:06] <ubitux> -vn ’ video none
[14:06] <ubitux> if youtube takes mp3, you might just need to mux an image in it
[14:06] <ubitux> the feature was recently added
[14:09] <ubitux> check the section mp3 in the documentation on how to add an album cover
[14:13] <Prirawien> Thanks you, now I have a video whose length is 4s, but audio pist's length is 3min... I will found.
[14:22] <JEEB> there was a command line to match the length to the audio
[14:22] <JEEB> *command line option
[14:24] <ubitux> you may not need a video
[14:24] <ubitux> just an id3 cover in a mp3 might do the trick
[14:24] <ubitux> but i'm not a youtube user
[14:25] <Prirawien> JEEB: I search an option like this, I didn't found anything, so I parse the output of "ffmpeg -i myfile.mp3" to have duration, do you know what option is ?
[14:26] <JEEB> if I had known the option, I would've said
[14:26] <JEEB> happy hunting
[14:26] <Prirawien> Ahah, ok
[14:30] <cHiKNPad> hey whats a good ffmpeg gui for windows i have tried a few some are poorly written with options not in english and bad libraries ... :o|
[14:30] <grepper> Prirawien: you can use -shortest  -loop 1
[14:31] <Prirawien> grepper: If I do this, I have a little video without any sound
[14:32] <grepper> Prirawien: works here
[14:33] <Prirawien> Can you give me your entire command line, please ?
[14:33] <grepper> ffmpeg   -i bg.mp3 -shortest  -loop 1 -i bg.jpg  -y videofile.wmv
[14:33] <Prirawien> Thanks, I try
[14:37] <grepper> probably -shortest should go before the first -i FILE rather than the -i IMAGE but seems to work either way
[14:59] <Prirawien> Ok it works, thanks you
[14:59] <globus> I can a question for ffmpeg-api here ?
[15:01] <grepper> np
[15:03] <globus> Should I ask a question for the ffmpeg-api here ?
[15:05] <grepper> globus: you can ask
[15:10] <globus> I have used ffmpeg-api to encode yuv420p frames (using x264 lib) and stream them with rtp. I managed to export the appropriate SDP info using the avf_sdp_create(). However i decide (to get more control on the encoder settings) to encode the frames using the x264-api and then pass the encoded frames to ffmpeg-api and stream them with rtp. My question is how to generate/export the appropriate SDP info now.
[16:02] <globus> Can I use the ffmpeg-api only to stream x264 NALUs ?
[16:11] <globus> Is it possible to stream already encoded data from other lib like x264 using ffmpeg ? Any hint ?
[16:39] <globus> In other words how can i use the ffmpeg-api to stream over rtp encoded data that already are encoded by other lib ?
[16:40] <globus> I ahve seen this ff_rtp_send_h264(AVFormatContext *s1, const uint8_t *buf1, int size)  is it ok ?
[18:48] <sprzybilla> I am reverse engineering some encoding parameters I found, but cannot seem to find info on :"-cmp +chroma"  can someone show me where I can find docs on this?
[18:51] <jhericurl> what is best gui-frontend for ffmpeg?
[18:55] <bauerj> I'm just converting some AVI-files to MP4 with -vcodec copy and -acodec aac. Now I'm wondering, why the conversion is very slow, although it's only audio. I get 8 fps and a bitrate of 1045.5kbits/s if that's relevant.
[18:55] <bauerj> can I do anything to make this a little bit faster?
[20:18] <mad_> I am running: ffmpeg -y -i "Input.mkv" -ac 2 -acodec libmp3lame -ab 256000 -b 3500000 -s 1366x570 -vcodec mpeg4 delme.avi
[20:19] <mad_> I get a video with relatively low vidoe quality (does not look like 3,5mbit), sound is 64kbit and video sie 850x500
[20:19] <mad_> So ffmpeg is ignoring EVERYTHING I specify :P
[20:19] <mad_> Why?
[20:19] <mad_> (well, alomst everything)
[20:25] <mad_> Ahhhh
[20:25] <mad_> Fuck me!
[20:25] <mad_> I should play the actual output and not the old testfile!
[20:53] <sprzybilla> when you are using multiple passes do you need to specify the same outputfile?
[20:58] <relaxed> sprzybilla: no
[21:12] <sprzybilla> hrm, for some reason my audio gets chopped off on the second pass, even though I am using the exact same audio parameters that were in the first pass
[21:13] <sprzybilla> might be user error though SO MANY FLAGS!!
[21:13] <sprzybilla> :P
[21:40] <lrussell> hi
[21:40] <lrussell> can anyone help me?
[21:41] <pasteeater> lrussell: just ask your question. if someone knows they will answer
[21:41] <lrussell> I'm trying to make ffmpeg read a file from http and save it directly to an ftp server without saving it on the main server first. I tried "ftp://blahblahlah/MYVID.asf" but it didn't do anything. Reading the file is fine, but uploading it is a no-go.
[21:42] <lrussell> ex. ffmpeg -i "http://urlhere/Wildlife.wmv" -vcodec wmv2 -acodec wmav2 -f asf "ftp://ftphere/Wildlife.asf"
[21:43] <lrussell> does ffmpeg support this or is there any trick to do it?
[21:45] <pasteeater> lrussell: supported protocols are listed with: ffmpeg -protocols
[21:46] <lrussell> is that reading from the server? or writting to the server
[21:46] <pasteeater> the output is divided by input and output
[21:46] <lrussell> oh, kk
[21:47] <pasteeater> that was a shitty sentence...the output of -protocols is divided...
[21:47] <pasteeater> alternatively you could mount the target directory with sshfs and output to that.
[21:48] <pasteeater> i'm assuming you're using linux
[21:48] <lrussell> yeah, except I don't have direct shell access, just php exec()
[21:49] <pasteeater> i have no experience with that.
[21:49] <lrussell> I got the ffmpeg -protocols return
[21:49] <lrussell> no ftp... :(
[21:51] <lrussell> well, there goes my project. Is there any way you can use a linux ftp client that will directly pipe ffmpeg's video file to the server?
[21:54] <burek> lrussell, why do you need it like that?
[21:54] <lrussell> because I have a shared web host with permission from the server owner to convert videos, but not store them due to lack of space.
[21:55] <burek> and why do you need "live" ftp storing, while actually encoding
[21:56] <burek> can you encode the video and then pipe the file as an input to a ftp client
[21:58] <lrussell> I know, I could do that, but with limited space, it would be difficult.
[21:58] <burek> i see
[21:58] <burek> well
[21:58] <burek> you could use live streaming then
[21:58] <lrussell> ?
[21:58] <burek> send the output, using udp, to the "ftp" machine
[21:58] <burek> where you will start another "listeninng" ffmpeg process, which will "catch" udp stream and save it
[21:59] <lrussell> ffmpeg isn't installed on that server, and can't be.
[21:59] <burek> wget?
[21:59] <lrussell> wget? what?
[22:00] <burek> can you install wget there
[22:01] <lrussell> wget supports udp?
[22:01] <burek> no
[22:01] <burek> but you can use wget
[22:01] <burek> to retrieve http stream
[22:01] <burek> started by ffmpeg (or vlc)
[22:02] <burek> vlc can do it in 1 line
[22:02] <burek> so you might wanna try that
[22:02] <lrussell> the recieving server is linux, and I don't have root or anything on it
[22:03] <burek> receiving server needs ffmpeg or wget
[22:03] <lrussell> it has neither
[22:03] <burek> than pay the cheap hosting :)
[22:03] <lrussell> it actually isn't that cheap... -_-
[22:03] <burek> or ask them for a free account in exchange for some work, like writting faq section
[22:03] <burek> or something :)
[22:04] <lrussell> or ffmpeg needs ftp support. :P
[22:04] <burek> ffmpeg is not a allware
[22:04] <burek> it's a fastest encoder
[22:04] <burek> that's what it is
[22:04] <burek> and ftp does not have anything to do with multimedia in particular
[22:05] <PowerCC> relaxed: hi
[22:11] <sprzybilla> huh will marshall hangs out in here eh?
[22:15] <relaxed> PowerCC: hi
[22:15] <PowerCC> thanks for  your help the other day, simple -vf plugin call.
[22:16] <PowerCC> it's been working AWESOME!
[22:16] <PowerCC> I'm going to try and compile ffmpeg for my lenny nas.
[22:16] <relaxed> good to hear
[22:16] <PowerCC> just for push, not transcoding
[22:16] <PowerCC> believe it or not, I'm not a n00b ;)
[22:17] <PowerCC> been using mencoder for years, back in the days of -vo and framebuffer
[22:17] <PowerCC> just started to use ffmpeg which is basically big part of mencoder.
[22:17] <PowerCC> what do you use it mostly for?
[22:20] <relaxed> transcoding mostly
[22:20] <PowerCC> x.264 mostly?
[22:20] <relaxed> h.264, yes
[22:22] <PowerCC> very nice
[22:22] <PowerCC> too slow for my slow a*** iMAc ;)
[22:22] <PowerCC> 2.4GHz
[22:22] <PowerCC> could I set threads to 4 if I have a dual core multi-thread?
[22:22] <PowerCC> or is it still two?
[22:24] <relaxed> using "-threads 0" == 1.5 * the number of cores you have.
[22:25] <dj_who> hi, I'm using ffmpeg for cutting video, and my problem is that i cut (using -ss and -t ) a clip at e.g. 00:03:50.00 and video is cutted properly but it seems that in the output file at he beginning I have few sound samples (less than half second) that were in input file before 00:03:50.00
[22:26] <dj_who> i use -vcodec copy and -acodec copy
[22:26] <dj_who> can any one help?
[22:26] <PowerCC> dj_who: it sounds like a frame limit.
[22:26] <PowerCC> you should cut by frame, not timestamp.
[22:26] <dj_who> the video codec is h264
[22:26] <PowerCC> relaxed:  what do you think?
[22:27] <dj_who> how to cut by frame
[22:27] <dj_who> ?
[22:28] <dj_who> i have the cut time from mythtv and to match the correct time in ffmpeg i must make calculations:
[22:28] <PowerCC> 4kb time=0.28 nitrate = 118.1kbits
[22:29] <PowerCC> you will have to fetch the frames first.
[22:29] <dj_who> durrationdiff=ffmpegduration - mythdurationtime -1s
[22:29] <PowerCC> personally i use virtual dub but I know it's not ffmpeg.
[22:30] <PowerCC> relaxed:  so 0 = 1.5 * prcoessors?
[22:35] <dj_who> PowerCC i don' get it, how cut this? or maybe is beter software to cut mpegts(h264+ac3+mp2+DVB_subtitles)
[22:39] <judget> I am attempting to use ffmpeg to stream from an hdpvr to a crtmpserver so I can live stream to a JW player in a web page.
[22:39] <judget> here is the pastebin that shows what I am trying http://pastebin.com/A6xEdHds
[22:39] <judget> it seems to go through about 4 cycles of transcoding and stops
[22:39] <judget> I note one error about the audio
[22:39] <judget> any suggestions or helpm wpould be greatly appreciated
[23:54] <tekk> hey guys, is mkv container support built into ff now?
[23:55] <tekk> i basically want low level access to x264 inside mkv& without compiling in libx264/libmatroska
[00:00] --- Thu Mar 22 2012