[Ffmpeg-devel-irc] ffmpeg.log.20130514

burek burek021 at gmail.com
Wed May 15 02:05:01 CEST 2013


[01:44] <reindeer> Hmm..how can i pass http stream url to lavfi filter 'amovie=http://mystream'
[01:48] <llogan> reindeer: i don't know but depending on what you're trying to do you may not need amovie.
[01:49] <reindeer> well, i was testing around the fancy filters and though of trying it for live input..
[01:49] <reindeer> http://ffmpeg.org/trac/ffmpeg/wiki/FancyFilteringExamples
[01:50] <llogan> which example?
[01:50] <reindeer> at showwaves and showspectrum
[01:50] <llogan> with ffplay or ffmpeg?
[01:50] <pFitz> hello, got a problem installing ffmpeg. after building/installing x264, fdk-aac ffmpeg gives me an error of not able to find libx264
[01:51] <reindeer> llogan ffplay
[01:51] <llogan> pFitz: use a pastebin service to show your ffmpeg configure options and the tail end of config.log
[01:51] <sacarasc_> pFitz: Do you have 2 copies of libx264 installed?
[01:52] <pFitz> I am not sure. how would I find out?
[01:53] <reindeer> llogan: i tried regular mp4 file and it played ok, but http url gives me: "Missing key or no key/value separator found after key 'http://mystream'
[01:53] <reindeer> and then i tried single/double etc quotes
[01:53] <pFitz> sorry, not to great with linux commands. how would I show my ffmpeg config options
[01:54] <llogan> history | grep configure
[01:56] <llogan> reindeer: im not sure if (a)movie can accept url, but that would be a good feature request if it does not
[01:57] <reindeer> llogan: hmm..oh ok
[01:57] <pFitz> do you want me to paste the history right in the chat?
[01:57] <llogan> no. use pastebin.com or similar
[01:58] <pFitz> ah, thats what you meant about pastebin....
[01:58] <llogan> also include the output of: tail ~/ffmpeg/config.log
[01:58] <pFitz> give me a min.
[02:00] <reindeer> llogan: filter manual states about 'movie' filter that it does not have to be a file, also device or a stream (accessed through some protocol) is said to be working
[02:01] <reindeer> and 'amovie' should be the same but "selects and audio stream by default"
[02:02] <pFitz> ok loagn, sorry. with pastebin, do i give you the raw, embed or?
[02:02] <llogan> raw is fine
[02:02] <reindeer> llogan: so i guess it should work OOB by not with my cmdline skills..
[02:03] <reindeer> *but
[02:03] <pFitz> http://pastebin.com/raw.php?i=P32VG3eB
[02:03] <pFitz> you want the url I assume?
[02:04] <llogan> yes.
[02:06] <durandal11707> reindeer: it does not work for me and i cant get it to work either
[02:06] <pFitz> wasnt sure how much of the end you needed on the log file
[02:06] <reindeer> durandal11707: copy that
[02:10] <llogan> pFitz: did you actually install libx264?
[02:10] <llogan> see https://ffmpeg.org/trac/ffmpeg/wiki/UbuntuCompilationGuide
[02:10] <llogan> (i'll be updating that page to a non-system install very soon)
[02:10] <pFitz>  I went threw the steps. so I thought  I did lol
[02:11] <pFitz> that might be the guide I was useing
[02:12] <llogan> but your history does not look like that is the case
[02:12] <pFitz> my history doesnt show I installed x264?
[02:12] <llogan> unnecessary sudo, and the guide does not "make install"
[02:13] <reindeer> durandal11707: I tried to "implement" this logic to getting it work, but ran out of skill..   http://ffmpeg.org/trac/ffmpeg/wiki/How%20to%20change%20audio%20volume%20up-down%20with%20FFmpeg
[02:13] <pFitz> running it as root was the only way I could get it to work. I would get permmision denied errors
[02:14] <llogan> you shouldn't need sudo for ./configure in most usage
[02:14] <durandal11707> reindeer: what you want to do?
[02:15] <pFitz> I did have one problem trying to install x264. was giveing me an error that yasm was not up to date so  I built the newest version from an other guide and then x264 "seemed" to install ok
[02:15] <llogan> anyway, you should have x264.h in /usr/include or /use/local/include
[02:15] <reindeer> durandal11707: show the "showspectrum" filter on to top of an input http stream
[02:16] <pFitz> what would be causeing me to have permission denied errors trying to ./configure?
[02:16] <durandal11707> reindeer: i told you that using streams with (a)movie is not possible, its bug
[02:17] <pFitz> Would I run into more problems if i go back to installing x264 again?
[02:17] <reindeer> durandal11707: well you did not say it was a bug, i though you were just as surprised as i was for it not to work..
[02:17] <llogan> pFitz: the permissions of the directory that contains configure or the user who owns it
[02:18] <reindeer> durandal11707: ..but it should be possible to use filters on any input right? so this could be done some other way perhaps
[02:18] <llogan> pFitz: it should be easy enough to copy and paste the commands in the guide i linked to
[02:19] <pFitz> ok, let me change the ownership of the x264 directory and try installing it again
[02:19] <durandal11707> reindeer: yes, download it
[02:21] <llogan> pFitz: probably easier to start over and follow the guide word-for-word
[02:22] <reindeer> durandal11707: yep.. :)
[02:22] <llogan> reindeer: or make a "wish" in the bug tracker
[02:22] <pFitz> http://pastebin.com/pQEX69W5 <- is what I got when I tried running ./configure again for x264
[02:23] <pFitz> I pretty much did, thats whats frustrating
[02:24] <llogan> you added sudo and used "make install" instead of checkinstall...unless that history was for something else
[02:25] <pFitz> I think that was for yasm
[02:25] <pFitz> not part of this
[02:25] <llogan> the guide has yasm compile instructions
[02:25] <pFitz> but your right, I did run sudo
[02:25] <llogan> what is your ubuntu version?
[02:26] <pFitz> 12.04
[02:26] <reindeer> llogan: well im sure there would be a way to do this using current builds, like fe. using an intermidiary file/buffer to disk and read that, but those methods are out of my skills atm : -/
[02:27] <pFitz> it gave me an error saying that I had ver 1.1 and min was 1.2
[02:28] <pFitz> i found an other guide to build and install 1.2 after that I didnt see any more errors
[02:28] <llogan> you're following the guide for 13.04. see the first paragraph for a link to the 12.04 version
[02:28] <pFitz> lol
[02:28] <pFitz> geez
[02:29] <pFitz> well ok. sorry to waste your time. I'll try again for 12.04 :)
[02:29] <llogan> i may have to address that issue. other users may not read that paragraph
[02:32] <reindeer> llogan: and basically because this is a dev channel, there is not a much better source for information at grasp on how to achieve things with ffmpeg, which im really happy about!
[02:34] <pFitz> right, I didnt even pay attention. I just went right into the steps
[02:34] <llogan> i think a lot of people do
[02:35] <pFitz> I have to agree with reindeer, its nice to beable to chat to someone that actually wrote the page
[02:35] <pFitz> or maybe not wrote. not sure. but you know what I mean
[02:35] <llogan> reindeer: this isn't a dev channel (that's #ffmpeg-devel). this is the channel for users.
[02:38] <reindeer> llogan: oh well, but u see i get a crazy idea, do some google, cant get it to work and end up here hopefully for the working solution, but hey its me and my crazy ideas! I know maybe 5% of the stuff that ffmpeg is capable of, so got a lot to learn and by coming here to talk, i can quite fast find out if this or that will even be possible! :)
[02:38] <durandal11707> if i spend more time answering questions here, i have less time to work on ffmpeg
[02:41] <pFitz> logan, I got to installing x264 again. coping the whole box and I am getting permmsion denied errors again. what should I do?
[02:41] <reindeer> durandal11707: but you decide how you spend your time, or what questions you want to answer. If they are dev related, then it would be good to talk about them, but if im asking "what is good pc to run four ffmpeg streams to downstears" dont waste your time on me ;)
[02:42] <pFitz> specificly right here patrick at ubuntuserver:~/fdk-aac$ ./configure --disable-shared
[02:42] <pFitz> ./configure: line 1881: config.log: Permission denied
[02:42] <pFitz> ./configure: line 1891: config.log: Permission denied
[02:43] <llogan> pFitz: i dont know.
[02:44] <llogan> what's with your permissions?
[02:46] <pFitz> i have no clue...  could it be something with ownership of build-essential, automake and the rest?
[02:46] <durandal11707> you created/modified files with root access, and now are trying same with lower access
[02:48] <pFitz> more then likely. not an easy fix
[02:49] <pFitz> maybe reinstall the OS? lol
[02:53] <llogan> pFitz: sudo chown -R pFitz:pFitz x264
[02:53] <llogan> or simply delete the x264 directory and start over
[02:54] <llogan> although you may want to: cd ~/x264 && sudo make uninstall
[02:54] <llogan> then delete or whatever
[02:55] <pFitz> ill try that thanks
[03:02] <pFitz> removed everything starting guide over
[03:08] <pFitz> what the heck... I start with yasm and it gives me "no rule to make target 'install' "
[03:08] <pFitz> and aborts
[03:12] <llogan> show your commands and the console output on pastebin.com
[03:15] <pFitz> one sec
[03:17] <pFitz> I can't go far enough back. I'm using putty to ssh to the server. it seems to be a permission problem still
[03:17] <pFitz> http://pastebin.com/HnZZYi7C
[03:17] <pFitz> is what i got
[03:19] <llogan> check to see if you have proper permissions for your current directory
[03:25] <pFitz> dr-xr-xr-x 7 patrick        patrick        4096 May 13 16:06 patrick
[03:25] <pFitz> means i dont have write access correct?
[03:25] <pFitz> wounder how i did that lol
[03:28] <llogan> chmod 755 /home/patrick
[03:28] <pFitz> thanks, was trying to do that but was one directory behind
[03:29] <pFitz> looks like that took care of that problem
[03:30] <llogan> probably the result of blindly using sudo
[03:31] <pFitz> yeah, i didnt realize somethings you dont need root to do
[03:31] <pFitz> im sure that could fix alot of my problems not having write access
[03:31] <pFitz> thats how I learn :)
[03:45] <llogan> pFitz: now is it working as expected?
[03:46] <llogan> by the way i changed the 13.04 guide so it performs a local install so no more system install and no more interfering with repo junk
[03:48] <pFitz> sorry, went to finish painting my living room. I think so far it is
[03:48] <pFitz> havent seen any errors. im to aac codec
[03:49] <pFitz> is that how 12.04+ setup with local install?
[03:51] <llogan> the 12.04 guide installs to "system" and can interfere with the shitty fake "ffmpeg" from the repository (and therefore anything that depends on said shitty, fake ffmpeg).
[03:52] <pFitz> gotcha
[03:53] <pFitz> thats what started me redoing my ffmpeg, i kept getting missing header errors stuff, trying to figure out why forked-daapd keeps stoping playback in itunes
[03:56] <relaxed> llogan: pong
[03:57] <llogan> relaxed: i updated the guide to make a "local" install. let me know if you see anything dumb.
[03:57] <relaxed> url?
[03:58] <llogan> https://ffmpeg.org/trac/ffmpeg/wiki/UbuntuCompilationGuide
[04:00] <relaxed> did you test this method?
[04:00] <llogan> yes
[04:00] <llogan> man page placement can be better though
[04:03] <relaxed> I'll look at it more once I'm at work.
[04:03] <llogan> ok. thanks.
[04:14] <Userdsfdsf> hi
[04:36] <reindeer> hi again, i got a small prob again lets see
[04:37] <reindeer> i got two inputs, that both need to be written onto disc and streamed out
[04:39] <reindeer> something like this:  -i input1 -c copy ./out1 -i input2 -c copy ./out2  -map 0:0 -map 1:0 -f matroska pipe:| vlc -
[04:40] <reindeer> the -map are all crazy cant figure it out and the output stream needs to have both tracks combined
[04:42] <reindeer> bascially i just cant figure out how to do the mapping, so that -map 0:0 goes to the file AND the output stream. Any idea how i can get this to work
[04:44] <reindeer> this is what i was doing before i tried to play with the FancyFilter testing..
[04:44] <reindeer> llogan still awake? :)
[04:58] <reindeer> hmm..i changed matroska to 'adts' and it seems to only support single audio track, but with matroska, i almost get it to work and then the playback is chopping up every second..it..cuts..a..little.piece..of..the..music! :) And im only streaming low quality 128kbit mp3 ..
[05:50] <pFitz> logan still around?
[08:18] <aruns> hi, i have installed ffmpeg from source on a couple of computers on my LAN. when running ffmpeg ... -f rtp rtp://<my-ip>:<port>, i'm able to do ffplay rtp://localhost:<port> and view the stream. but when i try and do ffplay -i out.sdp i'm getting errors about could not find codec parameters for stream 0
[08:19] <aruns> i'm creating out.sdp by dumping the output of the ffmpeg command
[08:21] <aruns> the command i'm using to send the stream - ffmpeg -re -f v4l2 -i /dev/video0 -s 320x240 -r 29.97 -vcodec libx264 -pix_fmt yuv420p -s 320x240  -tune zerolatency -profile baseline -level 30 -flags +global_header  -f rtp rtp://0.0.0.0:4008 > out.sdp
[08:22] <aruns> sorry to clarify i'm able to do ffplay rtp://localhost:<port> from my local machine, but i'm unable to get it to work with ffplay -i out.sdp from any other machine. anyone know what might be going on or have any pointers ? tia
[09:07] <aruns> anyone have any success doing a point to point stream using rtp and the latest development version of ffmpeg ?
[10:57] <aruns> i'm trying to stream using ffmpeg - ffplay over rtp. this is the latest developmental version of ffmpeg. it works when doing both on the same machine, but gives an error about unable to find codec parameter: unspecifed size when running ffplay from another machine. here is the pastie - http://pastebin.com/pw8NDLRj
[11:02] <luc4> Paranoialmaniac: sorry, can I disturb you?
[11:11] <luc4> Hi! Is there any guide on how to mux a h264 stream?
[11:43] <Paranoialmaniac> luc4: jük
[11:53] <luc4> Paranoialmaniac: hi! Do you remember the issue I was having yesterday muxing an h264 bytestream?
[11:53] <Paranoialmaniac> yes
[11:54] <burek> luc4 what's the big deal about muxing h264 stream?
[11:54] <luc4> Paranoialmaniac: you told me the stream was avcc, but I see in the stream that the start prefix for nals is present. Doesn't this mean that it is annex b?
[11:55] <JEEB> luc4, he meant that the avcc is the way it goes into "mp4" type containers
[11:55] <JEEB> and that you were doing it wrong then
[11:55] <luc4> burek: I'm trying to mux h264, but after I get my mp4, I get a file that is somehow broken at some point but I can't understand exactly what is wrong. Clearly I made something wrong using ffmpeg.
[11:55] <JEEB> burek, he's using the API
[11:56] <JEEB> luc4, anyways yes -- it's pretty sure you were getting Annex B from the thing as that's what raw H.264 streams generally are passed as
[11:56] <burek> oh i see
[11:57] <JEEB> there's a bit stream filter available that can convert Annex B to AVCc
[11:57] <JEEB> if I recall correctly
[11:57] <luc4> JEEB: so I have to get all the pps and sps, remove from the stream manually and put it into the mp4?
[11:57] <aruns> fflogger, i posted the problem again and linked this pastie - http://pastebin.com/pw8NDLRj is that enough information or do you need more ?
[11:57] <burek> aruns, complete output log
[11:58] <Paranoialmaniac> JEEB: no, you said such bsf is not present yesterday
[11:58] <JEEB> no, I just said what you said was the opposite
[11:58] <JEEB> although
[11:58] <JEEB> now that I see...
[11:58] <JEEB> -bsfs has no h264_annexbtomp4 :<
[11:58] <JEEB> only h264_mp4toannexb
[11:59] <JEEB> luc4, basically you will have to follow MPEG-4 (14496) Part 15 and create the avcc structure and feed it to the libavformat "mp4" muxer :s
[11:59] <Paranoialmaniac> i think libavformat is not suitable for annexb to avcff format
[12:01] <luc4> JEEB: ok, thanks. But do I also need to set something specific to libavformat or is it getting the information from the stream I pass?
[12:01] <JEEB> not fully sure how libavformat wants the data
[12:02] <JEEB> I guess having a read at libavformat/movenc.c and friends could be helpful
[12:03] <aruns> burek, fflogger - here is the pastie with the command and the complete output of ffplay with -debug 5 - http://pastebin.com/h4ByLkZw
[12:03] <Paranoialmaniac> it is pity that movenc.c doesn't help since it writes avcc as just binary string
[12:04] <JEEB> lol
[12:05] <Paranoialmaniac> the h264 parser can generate extradata for h264, but i don't know how to convert it to avcc format
[12:07] <Paranoialmaniac> oh ff_isom_write_avcc()
[12:08] <luc4> JEEB: but is this what muxing.c sample code does? It seems pretty simple from there...
[12:08] <Paranoialmaniac> but this is not a public function :(
[12:09] <Paranoialmaniac> hm
[12:13] <Paranoialmaniac> luc4: apparently, you should set extradata generated from the parser after all. extradata is annexb format but it will be written into avcC automatically
[12:14] <luc4> Paranoialmaniac: now I tried to convert my mp4 to avi, and seems to be pretty good...
[12:38] <xlinkz0> i get a lot of [mp4 @ 0x1cf1da0] st:0 PTS: 1314816 DTS: 1313792 < 1324244 invalid, clipping
[12:39] <xlinkz0> when trying to concat two videos
[12:39] <xlinkz0> the only difference in the codec that i can see is their
[12:39] <xlinkz0> 'tbn' is different
[12:41] <JEEB> it's not that tbn, at that point the timestamps are normalized against each other so any of the timebases shouldn't matter
[12:41] <JEEB> it just means that "the concat thingy fails at life", and it might be trying to hide that :P
[12:42] <xlinkz0> well how can i set it just to test if it's working?
[12:42] <JEEB> set what?
[12:43] <xlinkz0> tbn
[12:43] <xlinkz0> this is ffprobe output from both videos : http://codepad.org/OCqpRZBZ
[12:44] <JEEB> it's one of the timebases found from various metadata, and it's not even sure if those get used. And the real problem is not the damn timebase, but the fact that suddenly the concat thingy is getting a decoding timestamp that is smaller than the previous one, which shouldn't happen.
[12:44] <JEEB> in other words, it's failing to compensate for the fact that there are multiple sources, possibly
[12:44] <JEEB> can't be sure without looking at the code and I won't be doing that
[12:48] <xlinkz0> hmm simply doing ffmpeg -i video.mp4 video.mp4
[12:48] <xlinkz0> seems to fix it :\
[12:49] <JEEB> because you are re-encoding, yes
[12:49] <JEEB> quite naturally
[12:49] <JEEB> the fancy pantsy timestamps get not needed on the output side :P
[12:50] <JEEB> (the delay etc.)
[12:50] <JEEB> same thing as if you had originally just cut with re-encoding methinks? :)
[12:52] <xlinkz0> this is not related to yesterdays discussion
[12:52] <xlinkz0> i had to add some intro and outro movies and they weren't 'compatible' with the mp4's produced by ffmpeg
[12:52] <xlinkz0> i solved yesterdays cut problems by reencoding with -g 25 -keyint 25
[12:53] <xlinkz0> the files had constant RAPs
[12:53] <xlinkz0> have*
[12:53] <JEEB> keyint-min you mean, I guess?
[12:53] <xlinkz0> yes
[12:53] <xlinkz0> even though x264 reported it as 13 instead of 25
[12:53] <xlinkz0> it did produce constant keyframe interval
[12:54] <JEEB> then it did it by chance, and you really set keyint and not keyint-min, I guess?
[12:54] <xlinkz0> sec
[12:54] <xlinkz0> -r 25 -keyint_min 25 -g 25
[12:55] <JEEB> the number it reported should be the one it used internally
[12:55] <xlinkz0> it did it by chance.. to about 200 GB of video :)
[12:55] <JEEB> also there was another option via which you could just disable the "intelligent" keyframe decision algo
[12:55] <JEEB> don't remember which but the x264 settings wiki page should have it there, too
[12:56] <JEEB> in that case RAPs would come up automagically only at every keyint
[12:57] <xlinkz0> i'll look it up next time, thanks
[13:00] <xlinkz0> is there any way to 'cloud reencode' ?
[13:00] <xlinkz0> would've helped me a lot the other day to be able to buy something like this
[13:03] <JEEB> well, there are some services that have created their own re-encoding services on their server racks, but as I don't really know any that do exactly what you need in the way you want etc. I am not going to name any  (lol "cloud", most things you can get "on the cloud" as VMs are rather expensive if you plan on encoding with them)
[13:04] <JEEB> I know at least one video company that does indeed use "cloud" VMs for encoding, but only if their capacity suddenly goes over the expectations. Because they generally are either underpowered, overexpensive, or both to use them all the time :P
[13:05] <xlinkz0> well i'd only need them at certain events
[13:05] <xlinkz0> when i have like i did this weekend
[13:05] <xlinkz0> like 200GB of video that needs to be reencoded
[13:05] <xlinkz0> and preferably within a couple hours
[13:06] <JEEB> anyways, in most cases it's way more cost-effective to just get a "proper" dedicated box for it, or boxes, depending on the amount etc.
[13:06] <zap0> transfering 200GB woud take a few hours
[13:07] <zap0> as in.. uploading it and re-downloading it from "the cloud"
[13:07] <xlinkz0> internet landlines are cheap in my country
[13:08] <exutux> hi guys
[13:08] <StarGolder> i have a problem sound with ffmpeg. convert vob to mp4... after 10-20 seconds sounds mutes
[13:08] <exutux> there is a way to remove DRM from mp3 files?
[13:08] <xlinkz0> i could get 200 MB/s for something like 200 $
[13:08] <exutux> with ffmpeg obvs
[13:09] <JEEB> exutux, most probably no :P ffmpeg doesn't generally handle any kind of DRM
[13:09] <JEEB> xlinkz0, ~2 gigabits? Are you sure? :P
[13:09] <JEEB> 200mbit/s is more realistic
[13:09] <xlinkz0> a home subscription of 10$ gives you 10 MB/s
[13:10] <xlinkz0> yeas, mega BYTES
[13:10] <xlinkz0> i realistically download with that speed from local datacenters
[13:10] <exutux> uhm ok JEEB thanks... btw do you know some ( linux ) mode for to do it? can you write me in pvt, maybe is better :D
[13:10] <JEEB> exutux, no, I do not. Usually specific to the service you got them off of.
[13:10] <JEEB> Go Google and have fun
[13:11] <exutux> JEEB: I done it...but there are sw for windows :(
[13:11] <StarGolder> ffmpeg -i source -ac 2 out.mp4 ... try that but sound was gone after 10-20 seconds...
[13:14] <StarGolder> pls help me...
[13:24] <luc4> JEEB: I read this: http://stackoverflow.com/questions/15897849/c-ffmpeg-not-writing-avcc-box-information. Do you think this is correct? So I can place my annex b stream into my mp4 and I also read the header of my I-frames and put into extradata.
[13:28] <JEEB> the SPS and PPS (among other things) going into the extradata sounds like it'd supposed to be done, yes. But I'd prefer you rather consult part 15 on this stuff
[13:29] <JEEB> and see what ffmpeg spews out
[13:29] <JEEB> Annex B in "mp4" is something I've seen in the wild, but it makes me want to spank someone hard
[13:40] <luc4> JEEB: I see. I'll try :-) Do you know if I am supposed to simply copy the SPS and PPS like with memcpy?
[13:41] <JEEB> you could see how the internal function ff_isom_write_avcc() does related things and follow suite?
[13:41] <luc4> JEEB: yes, I'm reading that portion now, thanks.
[13:46] <aruns> i have compiled ffmpeg 1.2.1 on two machines and i'm attempting to stream via rtp. i'm able to play the video on the local machine using ffplay, but i'm getting an error about unspecified size when trying ffplay from the remote machine. here is the pastie - http://pastebin.com/h4ByLkZw anyone know what might be going on or have suggestions to try? tia
[13:49] <burek> aruns, analyzeduration is specified in microseconds
[13:50] <aruns> burek, thanks. any idea how high i should reasonably try ?
[13:51] <burek> you shouldn't in general
[13:51] <burek> what is the version of ffplay on your localhost
[13:51] <aruns> the only reason i'm doing it because the error message says to try it
[13:52] <aruns> i've now compiled 1.2.1 on both
[13:52] <aruns> i was using git on both earlier
[13:53] <burek> can you provide the log from your ffplay on localhost that works, too?
[13:53] <aruns> sure
[13:56] <aruns> burek, this the ffplay from localhost - http://pastebin.com/hqtRXmc4 it throws a few warnings but after that last warning the video starts to play
[14:01] <burek> hmh
[14:01] <burek> in your sdp file
[14:01] <burek> you have a private ip address
[14:01] <burek> which will work from your localhost
[14:01] <burek> but might not be visible outside your home lan
[14:01] <aruns> i tried changing both of them to the LAN ip
[14:01] <aruns> i'm doing this only on the LAN
[14:01] <burek> o=- 0 0 IN IP4 127.0.0.1
[14:01] <burek> 127.0.0.1 also won't work over the network
[14:02] <burek> you might see on wikipedia about sdp file structure
[14:02] <burek> to get the idea what ip means what
[14:02] <burek> and change it accordingly
[14:02] <aruns> ok thanks, taking a look
[14:11] <aruns> burek, apparently o= refers to the session originator, so the IP address of 127.0.0.1 ought to work since it would denote the local machine. anyway i tried to give it the public IP address of the remote machine and still got the same error
[14:12] <burek> there is more than one wrong ip inside sdp
[14:13] <Devedse> Does anyone know if the website x264.nl has other or the same builds as "JEEB's builds"?
[14:15] <aruns> the line o= ... now has 192.168.1.12 - the machine that is trying to run ffplay and c= ... has 192.168.1.6 - the IP of the machine running ffmpeg , could you point out the problem ?
[14:19] <Jay7> hello! I'm trying to compile freej on my ArchLinux, but it fails on AVFormatParameters, which was deprecated as I understand
[14:19] <Jay7> is there any hint how to rewrite this?
[14:20] <burek> aruns http://www.hjp.at/doc/rfc/rfc4317.txt
[14:20] <burek> Jay7 did you ask freej developers?
[14:21] <Jay7> I'm not sure they are 'exists' :)
[14:21] <xlinkz0> why was the yasm step removed from the compilation guide?
[14:24] <xlinkz0> seriously wtf the old compilation guide was much better
[14:25] <burek> which one is an "old" one xlinkz0?
[14:26] <xlinkz0> the one with checkinstall
[14:26] <burek> wiki? ubuntucompilationguide?
[14:28] <aruns> the current install guide has checkinstall http://ffmpeg.org/trac/ffmpeg/wiki/UbuntuCompilationGuide it also links different guides for different versions of ubuntu
[14:28] <xlinkz0> aruns: it doesn't even have the word checkinstall in it
[14:29] <xlinkz0> i'm trying to compile un debian squeeze
[14:29] <burek> xlinkz0, check the history
[14:29] <burek> on the wiki, i mean
[14:29] <aruns> well someone has just changed it in that case. i used that guide an hour ago
[14:29] <xlinkz0> quantal seems to have the old one
[14:29] <burek> llogan removed checkinstall for a reason
[14:29] <burek> i think it was last night
[14:30] <xlinkz0> dunno what reason but it worked flawlessly on debian
[14:30] <burek> scroll up this chat, or type !logs
[14:30] <xlinkz0> i'll just use quantal thanks
[14:30] <burek> ok
[14:31] <aruns> burek, i tried using the IP of the machine running ffmpeg in both places, so similar to host.atlanta.example.com in the sample config shown in that RFC
[14:31] <aruns> both for o=... and c=...
[14:31] <burek> btw, if you only need x264 in ffmpeg, regardless of the linux flavor, you can see this too: https://ffmpeg.org/trac/ffmpeg/wiki/How%20to%20quickly%20compile%20FFmpeg%20with%20libx264%20(x264%2C%20H.264)
[14:34] <xlinkz0> thanks, good to know
[16:00] <Devedse> Does anyone know if the website x264.nl has other or the same builds as "JEEB's builds"?
[16:12] <lmf40> Devedse: what are you looking for?
[16:12] <lmf40> Devedse: better build it yourself
[16:18] <whatadewitt> howdy! im trying to convert a video to animated gif and i cant get any sort of decent quality. everything i find on google is a similar problem with no real solution... anyone in here able to help me?
[16:20] <Devedse> lmf40, I see no advantage in building it myself, that's why I was looking for a build from the internet and was wondering wether to use JEEB's build or the build from x264.nl
[16:22] <lmf40> Devedse: when you build it yourself you will at least have the latest revision
[16:25] <Devedse> lmf40, well, I could build it myself but according to JEEB's website he added a lot of "patches" to it. This leads me to believe that to get a good running version of x264 you need to do additional patches before you can get it to work 'good'. Is this correct?
[16:26] <lmf40> Devedse: well, I use the current revision and it works fine. I think you should look into what patches he did and than decide if you need them.
[16:26] <Devedse> I can't really find much information about that sadly
[16:27] <relaxed> He's usually here and in #x264 most of the time. Read the patches.
[16:27] <Devedse> + spending half a day on compiling something that will most likely run just as good when taking a build from the internet is a bit too much for me
[16:27] <Devedse> Didn't know of x264, I'll join that too, thanks :)
[16:28] <relaxed> why do you want to use x264 directly and not ffmpeg with libx264?
[16:28] <lmf40> relaxed: good question
[16:29] <Devedse> I'm using MEGUI with Avisynth to run some scripts
[16:30] <Devedse> and from what I read ffmpeg was not directly compatible with avisynth
[16:30] <Devedse> Since avisynth uses VFW (correct me if i'm wrong)
[16:30] <Devedse> <--- still new in the encoding business
[16:31] <relaxed> ffmpeg has avisynth support on Windows, plau it has its own filters.
[16:31] <relaxed> plus*
[16:31] <Devedse> ok, I have not spent time on researching that further
[16:31] <Devedse> MeGUI just has a lovely GUI ;)
[16:32] <Devedse> Command line is great sometimes but if I have to create my own scripts to run ffmpeg to run scripts I find MeGUI with "one click encode" work easier
[16:37] <burek> Devedse, command line is always awesome
[16:37] <burek> if you know what you are doing (which you should) :)
[16:39] <burek> also, no need to spend "half a day" compiling things, there is a ton of tutorials on our wiki which you could use and build your ffmpeg in no more than 15 minutes
[16:39] <burek> 30 if you are a beginner
[16:39] <Devedse> Well with command line you can't really automate things often
[16:39] <Devedse> with that i mean command line alone
[16:39] <burek> what do you mean?
[16:39] <lmf40> Devedse: WHAT?
[16:39] <burek> can you provide an example?
[16:40] <lmf40> i am wondering how you can automate without a command line
[16:40] <lmf40> :-)
[16:40] <burek> :)
[16:40] <Devedse> For example I wanted to merge 2 vids together, Pick the video stream from video 1 and then pick every other strream from video 2 (with the -map commands)
[16:40] <Devedse> I couldn't find a way to select "-mapeverythingbesidesvideostream"
[16:40] <burek> :)
[16:41] <burek> did you try -map -0:v
[16:41] <Devedse> :)
[16:41] <Devedse> no
[16:41] <Devedse> :P
[16:41] <burek> or should i say: ffmpeg -i input -map 0 -map -0:v
[16:41] <burek> try that
[16:41] <Devedse> burek won't that map 2 video streams to 1 file?
[16:41] <burek> that says "map all the streams from the input 0, then map out all the video streams from the input 0"
[16:42] <Devedse> oh -0
[16:42] <Devedse> ah
[16:42] <burek> (note the minus on the second map)
[16:42] <Devedse> ok I was wrong
[16:42] <Devedse> :>
[16:42] <burek> :)
[16:43] <burek> you would probably end up using something like this: ffmpeg -i input1 -i input2 -map 0 -map -0:v -map 1 -c copy output
[16:43] <burek> or  "-map 1:v" instead
[16:44] <Devedse> yes, I did not find minus thing
[16:45] <relaxed> you can cheat and use -vn, but I prefer -map
[16:45] <Devedse> -vn is deprecated
[16:45] <Devedse> nasty shit
[16:46] <relaxed> it's still listed in -h
[16:47] <Devedse> Oh well, for now I'm happy with MeGUI, I kinda like how it works :x
[16:56] <Computer_hobby> I'm trying to generate a video for testing using the mandelbrot filter.  I can see it when i overlay the filter onto an input video, but I just get a black video if I try to use it on its own.  Anything obvious I'm missing in my command?
[16:56] <Computer_hobby> ffmpeg -y -f lavfi -t 2 -an -i mandelbrot mandelbrot.mp4
[16:58] <ubitux> that command works fine here
[16:59] <ubitux> and output playes fine
[16:59] <ubitux> -e
[17:00] <Computer_hobby> huh, strange.  I get a 2 second long black video
[17:11] <burek> Computer_hobby, what version of ffmpeg
[17:15] <Computer_hobby> ffmpeg version 1.1.git
[17:20] <durandal_1707> Computer_hobby: what you use to play mp4?
[17:20] <Computer_hobby> quicktime on OSX
[17:21] <durandal_1707> that supports only baseline h264
[17:21] <durandal_1707> thus you need to add -pix_fmt yuv420p
[17:21] <durandal_1707> when encoding to mp4
[17:21] <Computer_hobby> oh yes, looks like vlc can play it
[17:25] <Computer_hobby> Thanks durandal_1707, that did it.
[19:55] <mgeary> hi folks
[19:56] <mgeary> i've got some audio-only FLV files that i am converting to ogg and mp3 files. Is there a way i can tell ffmpeg to ignore/not look for a video track? Just trying to keep noise down. I tried passing -vn, but that doesn't appear to have changed anything
[19:56] <durandal_1707> whatadewitt: what version of ffmpey you use?
[20:00] <whatadewitt> durandal_1707 - 1.2
[20:01] <whatadewitt> i
[20:01] <whatadewitt> er, i'm using ffmpeg -ss 20 -i 'lucky.mp4' -vf scale=480:-1,format=rgb8,format=rgb24 -t 5 -r 10 output.gif, i found it online, i don't understand the -vf args
[20:02] <whatadewitt> if i could understand that it would be fantastic
[20:17] <Computer_hobby> the command you're running would spit out an image file based on the output being named 'output.gif'
[20:22] <whatadewitt> computer_hobby: thanks, i know what it's doing, i just dont understand those filters, or really, how the filter process works
[20:22] <whatadewitt> i know doing that gives me a gif from 20sec - 25sec of the video
[20:22] <whatadewitt> and i know these settings give me the best quality
[20:22] <whatadewitt> but i dont understand why
[20:23] <saste> whatadewitt, when all else fails, read the fine docs
[20:24] <Computer_hobby> yeah I just started playing around with ffmpeg last week for work so I'm new at it too.  I read http://ffmpeg.org/trac/ffmpeg/wiki/FilteringGuide and picked out some filters in http://www.ffmpeg.org/ffmpeg-filters.html and just started playing around with it too pick it up
[20:24] <whatadewitt> saste - i'm trying to understand them, but this section seems either thin or too confusing... http://ffmpeg.org/ffmpeg.html#filter_005foption
[20:24] <whatadewitt> computer_hobby: thanks! I'll chec this out
[20:24] <whatadewitt> check it out, too!!
[20:25] <Computer_hobby> heres a file i made where i was trying out different filters https://gist.github.com/ratbeard/5578236
[20:26] <whatadewitt> thanks!
[20:29] <Computer_hobby> -vf and -filter_complex are two ways to declare filters, the latter is a little more verbose but makes more sense to me.  [0]negate,vflip[b] would take the first -i argument source file, then negates and flips it and names the output of that "b", which you could later use [b] as an input to another filter, like '[b][1]overlay' which would combine it with the second -i arg
[20:29] <whatadewitt> sounds tricky :S im just trying to create simple gifs
[20:30] <whatadewitt> based off of some video files i have
[20:30] <Computer_hobby> hehe okay then thats probably more than you need
[20:59] <durandal_1707> whatadewitt: 1.2 have very poor gif encoder
[21:00] <durandal_1707> latest master is better (at least with compression)
[21:00] <whatadewitt> i just used homebrew to install, figured that would install the most recent
[21:00] <durandal_1707> for much better look you need very good dither
[21:01] <whatadewitt> can you expand?
[21:01] <durandal_1707> the palette generated is far from optimal, thus too much dithering is used
[21:02] <durandal_1707> so do not expect too much of it ....
[21:37] <luc4> Hi! When muxing an h264 annex b into an mp4, is it correct to repeat all the sps and pps into the frames I put in the mp4 with libavformat?
[21:38] <Paranoialmaniac> luc4: don't put sps and pps on any frames
[21:39] <luc4> Paranoialmaniac: ah ok thanks, it was kinda weird...
[21:39] <Paranoialmaniac> sps and pps are sent from avcC
[21:39] <luc4> Paranoialmaniac: I've seen codes around doing that so I was in doubt.
[21:40] <Paranoialmaniac> 14496-15 forbids putting any parameter sets into frames
[21:41] <Paranoialmaniac> well, the future amendment defines avc3 and avc4 format which allow putting parameter sets on frames
[21:41] <luc4> Paranoialmaniac: ah yes yes sorry, found it... it states it clearly. Thanks.
[21:49] <whatadewitt> i'm trying to install the newest version of ffmpeg and i'm getting this error when trying to make libaacplus "You must have ggglibtool installed to compile libaacplus."
[21:55] <durandal_1707> why you compiling it, libfdk-aac is better
[21:58] <llogan> install libtool if you must use libaacplus instead of fdk-aac
[21:59] <whatadewitt> i was following this: https://ffmpeg.org/trac/ffmpeg/wiki/MacOSXCompilationGuide
[21:59] <whatadewitt> installing via homebrew apparently doesnt give me the most recent version
[21:59] <whatadewitt> or i may be misunderstanding
[22:03] <whatadewitt> w00ps my client crashed
[22:07] <llogan> you didn't miss anything
[22:08] <llogan> apparently no other osx users are here now
[22:08] <swordfischer> Hello, I'm trying to screencast a game, with a webcam overlay using ffmpeg. I've mixed and matched a couple of scripts I found on google and finally got something working, though my problem is that the webcam overlay some times freezes (-vf "movie=/dev/video0:f=video4linux2 ..") - would it be possible to somehow 'reset' the webcam so the overlay will work again, or do I really have to interrupt my stream? - I'm on Linux (Gentoo 3.7.10), ffmpeg 0.10
[22:08] <llogan> but that guide already has you installing fdk-aac, so i see no reason for libaacplus,
[22:09] <swordfischer> just a sec :)
[22:10] <swordfischer> http://pastebin.com/uGeGersp < my ffmpeg script
[22:11] <llogan> please show an actual command and the complete console output
[22:11] <llogan> also, 0.10 is considered old here
[22:13] <swordfischer> http://pastebin.com/3EKM1aqe - Line 19 contains the full command
[22:14] <swordfischer> I also removed my twitch key, for security reasons
[22:14] <llogan> ah, 10.7. that's better at least.
[22:14] <llogan> 0.10.7
[22:16] <swordfischer> the filter for the webcam is just something I found on a forum somewhere, so I'm not entirely sure what it does though
[22:16] <llogan> since I don't know the cause is i'm obliged to say "try ffmpeg git head".
[22:17] <swordfischer> I'd guess I could try that
[22:17] <swordfischer> not aware of a way to 'reset' the connection to the webcam of sorts?
[22:17] <llogan> no, but using recent from git can ensure that you are not encountering an already fixed bug
[22:18] <swordfischer> makes sense. I'll poke around with that, thanks :)
[22:30] <llogan> swordfischer: also, you can ditch movie filter
[22:30] <llogan> http://pastebin.com/5SrsXKQv
[22:30] <llogan> (untested example)
[22:31] <swordfischer> same thing happened with git version of ffmpeg, so I guess it's down to my filter or the hardware
[22:32] <Fjorgynn> night
[22:38] <llogan> swordfischer: does it work as expected if you output to a file isntead of justin.tv?
[22:40] <swordfischer> well your command has some errors I believe.. the -f video4linux2 - is that supposed to be -vf?
[22:40] <llogan> no
[22:42] <llogan> swordfischer: works for me
[22:42] <swordfischer> http://pastebin.com/XUfyz4Ai - the output
[22:47] <llogan> try replacing 0:v with 1:0 and 1:v with 2:0
[22:49] <llogan> or use 1:v and 2:v (i tested with no audio stream so i only had two inputs not 3)
[22:50] <swordfischer> works with 1:0 and 2:0 - I'll just look into another error I had due to the f fmpeg upgrade though
[22:54] <llogan> hmmm...my webcam stopped too
[22:55] <swordfischer> oh, thats *great*, to some extend :P
[22:57] <llogan> at about 7 or 8 seconds
[23:01] <swordfischer> yeah, mine stops randomly.. this time at 1s
[23:47] <luc4> Hi! When writing avcc stream for ffmpeg to place into mp4, do I have to use a 4 byte field for the size of the frame?
[00:00] --- Wed May 15 2013


More information about the Ffmpeg-devel-irc mailing list