[Ffmpeg-devel-irc] ffmpeg.log.20150620

[burek]

[06:25:39 CEST] <orbea> Is there any benefits to compiling ffmpeg with openssl instead of just gnutls?
[09:57:02 CEST] <jarroddzz> hi everyone im trying to record a mp4 from a rtmp stream but it throws an AMF3 Object encapsulated in AMF stream does not start with AMF3_OBJECT!
[09:57:19 CEST] <jarroddzz> i could not find any good info on this error and why it happens any experts know why this happens?
[10:35:32 CEST] <gurpartap1> I am using following command line to stream audio on localhost and then recieve stream from ip camera and combining both streams with mp4 container.
[10:35:41 CEST] <gurpartap1> ffmpeg -fflags genpts -f alsa -i default -ac 1 -acodec libmp3lame -vn -sn -f rtp rtp://127.0.0.1:1235 &
[10:36:20 CEST] <gurpartap1> but there is little lag in audio sometimes...please review my commandline is right or wrong...
[10:37:00 CEST] <gurpartap1> This command is to mux two streams ...ffmpeg  -fflags genpts -i stream.sdp -rtsp_transport tcp -i rtsp://192.168.0.127 -tune zerolatency  -vcodec libx264  -acodec libmp3lame -s 600x480 -y -f mp4 2streams.mp4
[10:45:00 CEST] <gurpartap1> Please help somebody...
[10:56:19 CEST] <gurpartap1> what is the purpose of -fflags genpts .can somebody explain ?
[10:58:21 CEST] <anoop_r> i used this ffmpeg -i input.mp4 -c:v copy -c:a libfdk_aac -vbr 3 output.mp4 to encode but mediainfo showing it as Bit rate mode : Constant. why
[11:01:04 CEST] <anoop_r> anyone help
[11:18:04 CEST] <anoop_r>  i used this ffmpeg -i input.mp4 -c:v copy -c:a libfdk_aac -vbr 3 output.mp4 to encode but mediainfo showing it as Bit rate mode : Constant. why
[11:18:20 CEST] <anoop_r> please please
[12:34:01 CEST] <nyuszika7h> how do I convert mp4 to webm without making the quality horrible?
[12:35:26 CEST] <nyuszika7h> ah, I got it, https://trac.ffmpeg.org/wiki/Encode/VP8
[12:35:31 CEST] <nyuszika7h> "Note that you shouldn't leave out the -b:v option as the default settings will produce mediocre quality output. "
[12:36:42 CEST] <nyuszika7h> ah, and there's this: "Important: If neither -b:v nor -crf are set, the encoder will use a low default bitrate and your result will probably look very bad. Always supply one of these optionsideally both."
[15:43:51 CEST] <x86iac> i have a mp4 that is badly out of sync - if i use -i .mp4 .avi it gets in sync but quality is lost as is with -vcodec copy -acodec copy   - i can convert to any format .. any information to keep quality and sync it up?
[15:44:15 CEST] <x86iac> same with output to mkv
[15:46:40 CEST] <JEEBsv> if your input is already fucked up I can't really expect too much from just remuxing it. -c copy or -c:a/-c:v copy should just remux those streams so I don't see how you manage to lose quality there :P
[15:49:15 CEST] <x86iac> it does and is quite apparent - the result looks like 'old fashioned horizontal tv' :)
[15:49:34 CEST] <JEEBsv> then post your full command line and output on a pastebin and link here
[15:50:29 CEST] <x86iac> just simple   ffmpeg -i a.mp4 a.mkv   or  to a.avi
[15:50:41 CEST] <JEEBsv> well fucking naturally then
[15:50:50 CEST] <x86iac> that is why i am asking
[15:51:11 CEST] <x86iac> but it gets in sync
[15:51:42 CEST] <JEEBsv> you are not specifying a video or audio format, so ffmpeg picks some with a specific list, and then whatever default rate control is for that given format. in many of things that is like 200kbps or so
[15:52:04 CEST] <JEEBsv> tl;dr always specify a -c for all streams
[15:52:19 CEST] <JEEBsv> and if you are re-encoding always set rate control of your choice
[15:52:32 CEST] <x86iac> what would you suggest?
[15:52:39 CEST] <JEEBsv> ffmpeg -i in.mp -map 0 -c copy out.mp4 leads to what kind of result?
[15:52:45 CEST] <JEEBsv> *in.mp4
[15:52:59 CEST] <x86iac> brb
[15:53:00 CEST] <JEEBsv> -map 0 is "select all streams from the first input file"
[15:53:19 CEST] <JEEBsv> -c copy sets all stream's "encoding" to "copy", which is basically remuxing
[16:12:13 CEST] <testboy> hey
[16:13:25 CEST] <testboy> I am downmixing 5.1 Dobly Digital audio to 2 chanel AAC
[16:13:55 CEST] <testboy> but volume is lower
[16:14:41 CEST] <testboy> its ok in vlc with 2ch downmixing
[16:14:58 CEST] <testboy> anyone knows how to make it normal?
[16:15:53 CEST] <JEEBsv> now try guessing which is actually correect
[16:15:55 CEST] <JEEBsv> *correct
[16:16:45 CEST] <testboy> our users dont like the lower one
[16:16:52 CEST] <testboy> so loudest is correct
[16:17:11 CEST] <JEEBsv> then just separately make it louder?
[16:17:29 CEST] <testboy> how?
[16:17:48 CEST] <testboy> is there native way to do it like vlc
[16:18:15 CEST] <JEEBsv> you are being awfully vague
[16:18:25 CEST] <JEEBsv> how am I supposed to know what you mean with "native way"
[16:19:13 CEST] <testboy> sorry im new at ffmpeg
[16:19:44 CEST] <JEEBsv> anyways, you should first of all check your decoding results between liba52 and libavcodec without any mixing look like, and then look at the possible differences in mixing
[16:20:13 CEST] <JEEBsv> because one thing is sure, libavresample (and possibly libswresample) generally do the mixing correctly.
[16:21:18 CEST] <testboy> liba52?
[16:21:34 CEST] <testboy> im using libfdk_aac
[16:21:35 CEST] <JEEBsv> the thing that most VLC binaries use for AC-3 decoding AFAIK
[16:22:18 CEST] <JEEBsv> it probably has a command line decoder or if you trust VLC to not poke it too much you could just output WAV with the original stream mapping from it or so
[16:24:19 CEST] <testboy> so if i understand
[16:24:37 CEST] <testboy> i must use libavresample or libswresample
[16:25:39 CEST] <JEEBsv> that is what ffmpeg uses when you resample etc
[16:26:04 CEST] <JEEBsv> I'm trying to refresh my memory regarding VLC/liba52/possible differences between lavc and liba52 regarding AC3 decoding
[16:26:46 CEST] <JEEBsv> < kierank> it's correct in ffmpeg iirc <- so the dialog normalization stuff I was thinking of seems to be correct in FFmpeg
[16:27:08 CEST] <testboy> this is my test code if it helps https://i.imgur.com/re0oI9J.png
[16:28:12 CEST] <JEEBsv> > fdk_aac > 320k rate for 2 channels
[16:28:31 CEST] <JEEBsv> you're just throwing the bits around :P AAC is no MP3
[16:28:51 CEST] <testboy> :)
[16:28:58 CEST] <JEEBsv> anyways, after some quick research I must say that that kind of usage of av/swresample should be correct
[16:29:57 CEST] <JEEBsv> and personally I wouldn't start randomly modifying the volume of random crap
[16:30:28 CEST] <testboy> i have video links if you want
[16:30:40 CEST] <JEEBsv> you could run EBU R-128 normalization over an episode or whatever if you want, but I really wouldn't do more than that
[16:32:35 CEST] <testboy> what you think about audio filters?
[16:32:47 CEST] <testboy> i found this https://fluent-ffmpeg.github.io/index.html#audio-options
[16:33:30 CEST] <JEEBsv> as I said, at most I would possibly run the EBU normalization algo over it, but anything else is a no-go.
[16:34:32 CEST] <testboy> I am using node.js and fluent-ffmpeg so EBU is not possible
[16:34:55 CEST] <JEEBsv> as far as I can see that piece of shit is just calling ffmpeg and you're giving it parameters
[16:35:17 CEST] <JEEBsv> so unless it specifically doesn't let you some parameters you should be just fine :P
[16:35:59 CEST] <JEEBsv> although you do need to go through it separately first :P
[16:36:05 CEST] <JEEBsv> so that might be a limited factor
[16:36:47 CEST] <JEEBsv> *limiting
[16:40:42 CEST] <testboy> every month 4 million people watching videos encoded from this 'piece of shit'
[16:40:47 CEST] <testboy> whatever
[16:40:50 CEST] <testboy> thank you :)
[16:41:41 CEST] <JEEBsv> well it's not encoded with it, it's just calling ffmpeg :P and with that way more than that amount of people are watching the results of
[16:45:48 CEST] <testboy> oldu o zaman
[16:45:56 CEST] <testboy> zaten bi i_ede yaramad1 dediklerin
[16:46:05 CEST] <testboy> dsfgfd birazda sen translate et
[16:46:11 CEST] <testboy> hadi öptüm can1m
[16:46:26 CEST] <testboy> görü_ürüz
[16:46:36 CEST] <testboy> bb
[17:24:40 CEST] <halfie> hi, how to I enable all those "AV_LOG_DEBUG" messages? I am trying to hack on the mov
[17:24:52 CEST] <c_14> -loglevel debug
[17:26:01 CEST] <halfie> c_14, thanks!
[17:31:15 CEST] <halfie> this is my first day of hacking on the ffmpeg source, I am hacking on the "mov" demuxer to transform audio stream data before passing it to the "aac" decoder.
[17:32:10 CEST] <halfie> I was wondering if "aac" decoder can transform the audio data itself, it there a way to tell the AAC decoder more about the input audio data?
[17:32:22 CEST] <halfie> s/it there/is there
[17:39:15 CEST] <Mavrik> halfie, transform in what way?
[17:45:07 CEST] <BtbN> Sounds like something a bitstream filter would do
[17:46:41 CEST] <halfie> Mavrik, the raw aac audio samples are not available directly, they need to go through some de-obfuscation steps (which I call "transform")
[17:47:14 CEST] <halfie> essentially, my code will detect such samples, do the transform on the audio buffer, and then pass them to the original aac decoder
[17:48:35 CEST] <Mavrik> halfie, that functionality would probably best fit in a bitstream filter
[17:48:41 CEST] <Mavrik> not the muxer as such
[17:50:13 CEST] <halfie> Mavrik, I see. and which part of ffmpeg will activate this bitstream filter (which I will write)? does the flow look like "file --> muxer --> filter --> aac decoder". I haven't done any audio processing ever :)
[17:50:50 CEST] <halfie> https://ffmpeg.org/pipermail/ffmpeg-user/2012-December/011759.html I see
[17:51:11 CEST] <Mavrik> halfie, yes, bitstream filters work on encoded stream data and are inserted between (de)muxer and (de/en)coder
[17:51:57 CEST] <Mavrik> halfie, and I guess your filter might even exist
[17:52:00 CEST] <Mavrik> https://www.ffmpeg.org/ffmpeg-bitstream-filters.html
[17:52:03 CEST] <halfie> cool :), I will stop putting randoms printf statements all over the "mov.c" file now.
[17:52:29 CEST] <halfie> Mavrik, probably not, I just reversed engineered the de-obfuscation part.
[17:52:47 CEST] <Mavrik> ah
[17:52:57 CEST] <Mavrik> aac_adtstoasc filter should be a good sample for you then
[17:53:16 CEST] <halfie> Mavrik, cool! thanks for all the tips.
[19:16:44 CEST] <x86iac> my a.mp4 is badly out of sync but good video quality - with ffmpeg -i a.mp4 a.mkv   i get a good synced a.mkv with poor video quality - but with ffmpeg -i a.mp4 -sameq a.mk4 i get a poorly synced video but good quality -  can someone give me some information on this
[19:17:00 CEST] <x86iac> the commands i was given before didn't do anything to help
[19:17:11 CEST] <JEEBsv> so uhh
[19:17:32 CEST] <JEEBsv> I told you what to test so I'd like specific info on the results
[19:17:51 CEST] <JEEBsv> and just wipe sameq out of your fucking skull, it isn't what many people seem to think it is
[19:20:23 CEST] <x86iac> hey dude knock the bs talk off
[19:20:39 CEST] <x86iac> get your respect together
[19:21:40 CEST] <JEEBsv> if you actually read my lines I have very little bs in there, and if you don't remember, just scroll into your backlog to see what I noted for you to test
[19:22:28 CEST] <JEEBsv> if your backlog doesn't go far enough I can copypasta my lines from there
[19:24:23 CEST] <x86iac> i tested it and same as original mp4  - good video but no sync
[19:24:32 CEST] <JEEBsv> thank you
[19:25:28 CEST] <JEEBsv> is your source constant frame rate btw?
[19:26:02 CEST] <JEEBsv> would make it simpler to just override that and remux than to do a quick re-encode, use the timestamps from that and remux with those
[19:27:04 CEST] <x86iac> would mplayer -v   show that ?  ffmpeg -loglevel verbose?
[19:28:01 CEST] <JEEBsv> well I'd tell you to check the input, but since the input is wonky... you had some re-encodes lying around already, right?
[19:28:08 CEST] <x86iac> how to do the remux with the timestamps
[19:28:09 CEST] <JEEBsv> if you ffmpeg -i welp.mp4 those
[19:29:21 CEST] <JEEBsv> does it tell you that the input video stream has a frame rate (it probably does, but that is not a definite piece of info, but if that input frame rate looks standard we could just try with that)
[19:29:30 CEST] <x86iac>   Stream #0.1(und): Video: h264 (High), yuv420p, 1280x528 [PAR 1:1 DAR 80:33], 849 kb/s, 23.98 fps, 23.98 tbr, 24k tbn, 50 tbc
[19:29:35 CEST] <JEEBsv> ok
[19:30:00 CEST] <JEEBsv> ffmpeg -i original.mp4 -r 24000/1001 -c copy output.mp4
[19:30:01 CEST] <JEEBsv> test this
[19:32:33 CEST] <JEEBsv> that should force the video to have that frame rate
[19:32:54 CEST] <JEEBsv> also, what kind of player are you using to check the a/v synch?
[19:34:54 CEST] <x86iac> mplayer
[19:35:46 CEST] <JEEBsv> I also recommend testing out mpv, that is probably the least retarded of the mplayer family of players
[19:36:25 CEST] <x86iac> is the the -c copy    just -acodec copy and -vcodec copy   toghether?
[19:36:31 CEST] <JEEBsv> yeah
[19:36:49 CEST] <x86iac> it gives good video but out of sync
[19:37:23 CEST] <JEEBsv> of course, you're just copying the streams over
[19:37:33 CEST] <JEEBsv> if that would modify the quality we would be in big trouble ;)
[19:38:15 CEST] <JEEBsv> ok, for a quick test let's try out ffmpeg -i input.mp4 -c:v copy -c:a aac -strict experimental -b:a 192k out.mp4
[19:38:54 CEST] <JEEBsv> (-strict experimental enables "experimental" features, in which the lavc aac encoder still is for whatever reason)
[19:41:59 CEST] <Guest81481> Want to use FFMPEG to capture .m3u8 stream to .mp4 file - using code in http://pastebin.com/9c9tEW9y
[19:43:00 CEST] <BtbN> I wouldn't use -re for a stream
[19:43:11 CEST] <BtbN> Also, i wouldn't use mp4 to capture a stream
[19:43:17 CEST] <BtbN> flv or mkv is better
[19:44:50 CEST] <Guest81481> First time ffmpeg user. Any help appreciated. Why not use mp4 format? What will happen when stream ends?
[19:45:40 CEST] <JEEBsv> doesn't really matter unless you kill off ffmpeg somewhere in the middle
[19:45:52 CEST] <JEEBsv> because when the stream ends it will write the header and the file will be playable
[19:45:57 CEST] <BtbN> There is no good reason to use mp4, but a lot of reasons not to use it
[19:46:11 CEST] <JEEBsv> flv and matroska just let you play a file even if you stopped writing it in the middle
[19:46:22 CEST] <BtbN> If you need it for compatiblity with something, remux it after the recording is done
[19:48:30 CEST] <Guest81481> Since I'm creating a scheduled of batch files to record matches while away, should I do something to end capture?
[19:48:45 CEST] <JEEBsv> if the input ends, ffmpeg stops
[19:49:00 CEST] <JEEBsv> x86iac: did the audio-only re-encode work?
[19:50:16 CEST] <x86iac> the last one with the experimental looks good but out of sync
[19:50:25 CEST] <JEEBsv> ooh-kay...
[19:50:43 CEST] <x86iac> the straight    ffmpeg -i a.mp4 a.mk4 is in sync
[19:51:05 CEST] <JEEBsv> yeah, that re-encodes both.
[19:51:15 CEST] <x86iac> i'll try the -qscale ?
[19:51:41 CEST] <Guest81481> BtbN: thanks will remove -re and what else do I need to change for mkv format?
[19:51:53 CEST] <BtbN> the file extension.
[19:51:54 CEST] <JEEBsv> well we'd be using something better if we really have to re-encode. I'm trying to probe as much as I can without actually having the file to see if it can be salvaged without re-encoding
[19:52:03 CEST] <JEEBsv> because re-encoding as always costs quality
[19:52:21 CEST] <JEEBsv> x86iac: would it be possible for you to share the file in some sane way?
[19:53:05 CEST] <x86iac>  ffmpeg -i s.mp4 -qscale 1 ss.mkv   works perfect :)
[19:53:21 CEST] <x86iac> let me check the size
[19:53:31 CEST] <JEEBsv> yes, but that most probably isn't optimal in any way or form
[19:53:47 CEST] <x86iac> 5 times bigger then the original lol
[19:53:48 CEST] <JEEBsv> also it basically just gives up on the original streams :P
[19:53:50 CEST] <JEEBsv> well yes
[19:54:12 CEST] <x86iac> it is the swordfish movie
[19:54:14 CEST] <JEEBsv> I have no idea what your ffmpeg build defaults to but whether it's mpeg4 or libx264 q:v 1 isn't good news
[19:54:39 CEST] <x86iac> how can i tell what it defaults to?
[19:54:58 CEST] <JEEBsv> see the output part of the terminal output
[19:56:43 CEST] <JEEBsv> anyways, if you just want to fix a file for watching that should be OK :P
[19:57:17 CEST] <JEEBsv> there probably would be a way to fix up the original streams, but if it's just some random pirated thing we can go without the effort
[19:59:00 CEST] <jarr0dsz> hi anyone knows why a AMF3 Object encapsulated in AMF stream does not start with AMF3_OBJECT! happens?
[19:59:07 CEST] <jarr0dsz> im trying to debug this problem but it still persist
[19:59:18 CEST] <jarr0dsz> trying to save rtmp stream with flv or mp4 container always same error
[19:59:28 CEST] <x86iac> the -map 0  should have worked though?
[20:00:03 CEST] <x86iac> i've had this problem with other mp4 but usually just running them in ffplay fixed the sync
[20:00:45 CEST] <JEEBsv> uhh, -map 0 just picks all streams from the first input file
[20:00:58 CEST] <JEEBsv> and ffplay does exactly nothing to the files
[20:01:12 CEST] <JEEBsv> it's a proof of concept crappy thing of a player :)
[20:01:37 CEST] <x86iac> if fixes a lot things that won't play in mplayer ...
[20:01:59 CEST] <JEEBsv> then your use of the word "fixes" is kind of weird, because it fixes exactly nothing in a file
[20:02:10 CEST] <jarr0dsz> x86iac you refer to my message?
[20:02:17 CEST] <JEEBsv> in that case mplayer is just fucking things up
[20:02:18 CEST] <x86iac> jarr0dsz: sorry no
[20:02:24 CEST] <JEEBsv> you should try out mpv as an alternative
[20:02:33 CEST] <x86iac> i will try mpv
[20:03:21 CEST] <x86iac> thanks for all your command line options - this is good to keep testing
[20:03:48 CEST] <jarr0dsz> anyone successfully saved a rtmp stream to mp4 with ffmpeg?
[20:03:49 CEST] <JEEBsv> also I must note that if mpv on your distro is not available or old, you should try the mpv-build repository's scripts
[20:03:56 CEST] <jarr0dsz> i cannot find any much info on the error either
[20:04:18 CEST] <JEEBsv> that builds mpv with some of its dependencies from the latest sources as a static binary
[20:04:30 CEST] <JEEBsv> https://github.com/mpv-player/mpv-build
[20:04:34 CEST] <x86iac> mpv 0.9.2
[20:04:41 CEST] <JEEBsv> ok, that's pretty new :)
[20:05:23 CEST] <x86iac> jarr0dsz: try rtmpdump?
[20:05:37 CEST] <jarr0dsz> can rtmpdump save directly as mp4 then x86iac ?
[20:05:48 CEST] <JEEBsv> nope
[20:05:54 CEST] <JEEBsv> it dumps the flv stream
[20:06:13 CEST] <JEEBsv> also in theory if you provide the correct options to ffmpeg it should be able to read the stream as well
[20:06:23 CEST] <JEEBsv> if not, you could try poking the trac issue tracker
[20:07:15 CEST] <Guest81481> BtbN: Just changing the file extension to mkv gave me errors -"incorrect codec parameters"
[20:07:37 CEST] <JEEBsv> Guest81481: put the full command line and terminal output to a pastebin
[20:07:38 CEST] <JEEBsv> and link here
[20:07:53 CEST] <JEEBsv> (also the actual reason for that error is generally higher up in the log before it)
[20:09:07 CEST] <x86iac> JEEBsv: how do i find what the ffmpeg build defaults to but whether it's mpeg4 or libx264  ?
[20:09:27 CEST] <jarr0dsz> okay my need is to save directly to mp4
[20:09:36 CEST] <jarr0dsz> i can already do flv savings but mp4 is causing me lots of problems
[20:09:50 CEST] <JEEBsv> x86iac: it shows in the terminal output of the command
[20:09:58 CEST] <JEEBsv> you have one part that describes the input stream(s)
[20:10:05 CEST] <JEEBsv> and another that describes the output stream(s)
[20:10:15 CEST] <JEEBsv> the output part lists what encoders are used
[20:10:53 CEST] <Guest81481>  JEEBsv:  http://pastebin.com/9c9tEW9y Objective create batch file to save game stream while away.
[20:11:24 CEST] <JEEBsv> Guest81481: thank for you for not posting the other part which would actually have the information
[20:11:32 CEST] <JEEBsv> you are being so helpful I am about to burst <3
[20:11:41 CEST] <JEEBsv> </sarcasm>
[20:12:54 CEST] <jarr0dsz> noboy knows more about the error AMF3 Object encapsulated in AMF stream does not start with AMF3_OBJECT! ?
[20:13:01 CEST] <jarr0dsz> im also getting tis error on rtmpdump
[20:13:10 CEST] <jarr0dsz> it seems that on every stream im trying to record this error is thrown
[20:13:22 CEST] <jarr0dsz> probably be able to fix that error will let me save in mp4 also
[20:14:39 CEST] <x86iac> the listing of the libs?
[20:15:00 CEST] <JEEBsv> pastebin the terminal output and I will note you the point :P
[20:15:02 CEST] <Guest81481> JEEBsv: http://pastebin.com/BQqg31nK  </lol>
[20:15:21 CEST] <JEEBsv> Guest81481: [matroska @ 0594f420] Error parsing AAC extradata, unable to determine samplerat
[20:15:50 CEST] <JEEBsv> although I'm not fully sure that's the reason
[20:16:02 CEST] <JEEBsv> try leaving out the bitstream filter?
[20:17:21 CEST] <Guest81481> JEEBsv: if I only change it back to .mp4 file - works great no errors (noobie here - first time ffmpeg user)
[20:17:52 CEST] <JEEBsv> well great then, enjoy your life
[20:18:49 CEST] <x86iac> it seems to depend on what the output file extension given is
[20:19:20 CEST] <JEEBsv> x86iac: yes since certain defaults depend on what output format you're using
[20:19:24 CEST] <JEEBsv> (as in, container format)
[20:19:54 CEST] <JEEBsv> I wouldn't be surprised if AVI had mpeg4 (which is the MPEG-4 Part 2 encoder) and mp4/mkv had libx264 if available
[20:20:01 CEST] <JEEBsv> but it really depends on how your ffmpeg was packaged
[20:20:29 CEST] <Guest81481> Just a noobie trying to learn. The mp4 format is fine for me- just trying to optimize the cli code and learn at the same time.
[20:20:30 CEST] <x86iac> right there is no suitable output format for avi
[20:20:55 CEST] <x86iac> lol i typed mvi   avi is mpeg4
[20:20:57 CEST] <JEEBsv> Guest81481: well mp4 as such is just fine if the input streams fit it
[20:21:31 CEST] <JEEBsv> x86iac: so did that mp4 file work with mpv?
[20:22:18 CEST] <x86iac> give me a minute
[20:28:12 CEST] <Guest81481> thanks for the help - GTG pick my son from work. Much appreciated.
[20:33:51 CEST] <x86iac> it can't find my zlib installation .. i don't have a zlib.pc ?? i'll work on it
[20:35:31 CEST] <x86iac> qscale 8 is pretty good too
[20:35:43 CEST] <x86iac> only 50% bigger - not 5x
[20:36:01 CEST] <JEEBsv> if you are using libx264, use the -crf option instead of qscale/q:v
[20:36:22 CEST] <JEEBsv> crf is the closest to variable bit rate "constant quality" encoding we have
[20:36:36 CEST] <JEEBsv> start with around crf 23 and go down if it looks bad, go up if it looks good
[20:36:53 CEST] <JEEBsv> then you will find the highest crf value that still looks good to you
[20:37:30 CEST] <halfie> "ffmpeg -i input.mp4 -vn -acodec vorbis -ac 2 -ab 96k -f ogg -strict -2 -bsf:a testing -loglevel debug OUTPUT" ... my new bitstream filter ("called testing") is never called it seems. what goes?
[20:38:02 CEST] <JEEBsv> see -v debug
[20:38:10 CEST] <JEEBsv> it will output a lot but possibly should help
[20:38:28 CEST] <JEEBsv> also is the filter supposed to work before or after encoding?
[20:39:21 CEST] <halfie> JEEBsv, I am trying to make it work before the decoding. I assumed that it will always be that way, heh.
[20:39:35 CEST] <JEEBsv> halfie: then try putting it before -i
[20:40:16 CEST] <halfie> JEEBsv, woot! that did the trick, thanks!
[20:40:21 CEST] <JEEBsv> :)
[20:40:56 CEST] <halfie> option parsing code in ffmpeg must be super complex, and crazy! it can do so many things.
[20:43:08 CEST] <chungy> It's a little unusual to most programs. Order matters and duplicate arguments exist :P
[20:44:58 CEST] <halfie> maybe I spoke too soon! "ffmpeg -bsf:a CHOMP -i test.mp4 -vn -acodec vorbis -ac 2 -ab 96k -f ogg -strict -2 -loglevel debug OUTPUT" results in "Option bsf:a (A comma-separated list of bitstream filters) cannot be applied to input file test.mp4 -- you are trying to apply an input option to an output file or vice versa. Move this option before the file it belongs to."
[20:46:23 CEST] <halfie> maybe I need to move more options towards the left of "-i"
[20:48:27 CEST] <JEEBsv> not really
[20:48:30 CEST] <JEEBsv> hmm
[20:49:04 CEST] <JEEBsv> can't say much more than try to read the output of the verbose output (-v debug , no idea if -loglevel debug does the same)
[20:58:52 CEST] <x86iac> the mpv needs newer ffmpeg libs then i have - i'll try it on my new machine
[20:59:19 CEST] <x86iac> JEEBsv: thanks very much - without your guidance ... :)
[20:59:39 CEST] <x86iac> i've settled on -qscale 8
[21:00:32 CEST] <x86iac> how come -sameq works but not with the direct ffmpeg -i .mp4 .mkv
[21:00:42 CEST] <x86iac> which gives the sync
[21:01:21 CEST] <chungy> don't use -sameq
[21:01:46 CEST] <JEEBsv> uhh, sameq never works and I think the option was already removed effectively :P
[21:01:55 CEST] <JEEBsv> it was meant for a very limited amount of debugging
[21:02:04 CEST] <JEEBsv> it was supposed to copy quantizers from input to output
[21:02:12 CEST] <JEEBsv> which just doesn't make sense in any usual use case
[21:02:25 CEST] <chungy> yaeh it's removed in the latest versions for a while now
[21:03:07 CEST] <chungy> "ffmpeg -i INPUT -c copy OUTPUT" if you don't want to lose any quality (the streams are copied)
[21:03:25 CEST] <JEEBsv> chungy: he had either player issues or input issues, which re-encoding of the video side of things seem to have fixed
[21:03:47 CEST] <JEEBsv> (or his PC is just too slow to play the clip, can't say much more detailed because I have no way of accessing the file :P)
[21:03:59 CEST] <chungy> probably player/computer issue, yea
[21:04:18 CEST] <chungy> I have seen ffmpeg fix some malformed containers though just by copying the streams to a new file :P
[21:04:26 CEST] <JEEBsv> sure
[21:04:31 CEST] <chungy> although when things are malformed, all guarantees are gone anyway.
[21:04:35 CEST] <JEEBsv> ye
[21:06:04 CEST] <chungy> Also, put in "-map 0" if you want all streams to be copied. by default ffmpeg will only do the first video stream and the first audio stream. If your file is more complicated, you could end up missing some.
[21:08:51 CEST] <x86iac> the ffmpeg -i .mp4 -qscale 8 .mkv   fixes it good
[21:09:58 CEST] <halfie> JEEBsv, I can get "chomp" filter to apply just before the data is written to the output file. however, no luck in getting it to work on input data.
[21:10:17 CEST] <JEEBsv> halfie: since you actually are poking insides of libav* I recommend you poke #ffmpeg-devel
[21:10:18 CEST] <x86iac> only increases the file size by 50%
[21:11:06 CEST] <chungy> x86iac: try -c copy instead.
[21:11:21 CEST] <JEEBsv> x86iac: well you are just setting quantizer to 8 with whatever encoder you're using. it's much less adaptive or "intelligent" than f.ex. crf with libx264, but if you're using the mpeg4 encoder then qscale/q:v is the least bad you can go with
[21:11:41 CEST] <JEEBsv> well, since it's a crappy pirated file feel free to use whatever you want for your own viewing pleasure :)
[21:13:22 CEST] <halfie> JEEBsv, doing so now.
[21:14:54 CEST] <x86iac> chungy: the -c copy gives good video but no sync
[21:15:45 CEST] <JEEBsv> yeah, which either means that whatever is wrong is getting fixed with the re-encode, or your system isn't fast enough to play the original file for whatever reason (or the player you're using is borked regarding these file(s))
[21:15:58 CEST] <JEEBsv> to be honest I'd either put it on the side of the player or your hardware
[21:16:21 CEST] <x86iac> yea it is a 750MHz machine
[21:16:25 CEST] <JEEBsv> oh
[21:16:29 CEST] <JEEBsv> ok
[21:16:31 CEST] <JEEBsv> yeah
[21:16:39 CEST] <x86iac> but no 'slow computer' message
[21:16:48 CEST] <JEEBsv> then you will want mpeg4 part 2 (the mpeg4 encoder)
[21:16:51 CEST] <chungy> if you play it through mpv (or mplayer, even), it'll give a message if your computer is too slow to play it
[21:16:59 CEST] <JEEBsv> or libx264 with the -tune fastdecode option
[21:17:07 CEST] <JEEBsv> and most AVC files will either play slow or be unplayable
[21:17:10 CEST] <JEEBsv> my condolences :/
[21:17:16 CEST] <chungy> the baseline H.264 profile probably would work on that computer too. The default is high.
[21:17:20 CEST] <chungy> -profile:v baseline
[21:17:29 CEST] <JEEBsv> chungy: uhh rather just -tune fastdecode
[21:17:40 CEST] <JEEBsv> which is made for exactly this scenario
[21:17:51 CEST] <JEEBsv> (originally for stuff like xbox)
[21:17:59 CEST] <JEEBsv> (which funny enough was ~733MHz or so)
[21:18:19 CEST] <chungy> ah ok
[21:18:34 CEST] <chungy> xbox had a Pentium 3 IIRC
[21:18:39 CEST] <JEEBsv> celeron
[21:18:47 CEST] <chungy> not even Pentium?
[21:18:48 CEST] <JEEBsv> a bit castrated version of the p3, in other words
[21:19:02 CEST] <chungy> that's.... sadder than I already thought it was D:
[21:19:25 CEST] <JEEBsv> x86iac: in other words the input file is probably alright, but your hardware cannot handle it. whether or not mplayer gives out its famous message
[21:19:49 CEST] <JEEBsv> which explains why re-encoding with mpeg-4 part 2 fixes it
[21:20:46 CEST] <chungy> Should probably look into getting a computer newer than 15 years old. :P
[21:22:08 CEST] <x86iac> i have a newer hp - and will install the newer ffmpeg and try the mpv
[21:22:48 CEST] <x86iac> not much i can do with this machine - it is for python mostly
[21:24:27 CEST] <x86iac> python for (youtube-dl)
[21:25:06 CEST] <JEEBsv> the really sad thing is that those old machines aren't even economical on your electric bill in the end :/
[21:25:06 CEST] <x86iac> is there a lot of overhead for mp4 cause that is the only file type i have problems with
[21:25:17 CEST] <JEEBsv> no, the container most probably has nothing to do with it
[21:25:28 CEST] <JEEBsv> it is just most likely to contain AVC
[21:25:57 CEST] <JEEBsv> and AVC with CABAC and friends enabled is slower than older formats (but on the other hand you get immensely better compression)
[21:27:17 CEST] <JEEBsv> of course if you disable CABAC and optimize for fast decode (as -tune fastdecode does with libx264) then it can be relatively on the level of MPEG-4 Part 2 although I haven't really tested since I've long lost any will to play things on my xbox
[21:28:23 CEST] <JEEBsv> and now with the current-gen HEVC format the faster alternative to CABAC was even removed to simplify the format
[21:29:06 CEST] <x86iac> converting to mkv is certainly faster then to avi
[21:29:32 CEST] <JEEBsv> I will guess that mkv defaults to libx264?
[21:29:41 CEST] <JEEBsv> that's a much more optimized encoder
[21:30:19 CEST] <JEEBsv> but as I said, you probably want -tune fastdecode with that to be able to play the output :D
[21:30:37 CEST] <x86iac> is this the output line ?    Stream #0.0(und): Video: mpeg4, yuv420p, 1280x528 [PAR 1:1 DAR 80:33], q=2-31, 200 kb/s, 1k tbn, 23.98 tbc
[21:30:54 CEST] <JEEBsv> if that is not for the input, then yes
[21:30:59 CEST] <chungy> yeah container formats aren't going to be any hinderence to decoding speed. that's the codecs contained within them.
[21:31:34 CEST] <JEEBsv> and ok, it seems like either your ffmpeg has no libx264 linked in or is old enough to not have it default to it for mp4 and mkv
[21:34:08 CEST] <x86iac> libx264.so.114 => /usr/local/lib/libx264.so.114 (0xb6e70000)
[21:34:22 CEST] <JEEBsv> well then it's the latter case
[21:34:41 CEST] <JEEBsv> if that is what ffmpeg is linked against
[21:35:49 CEST] <x86iac> yes
[21:36:28 CEST] <JEEBsv> yeah, then you should be able to test out libx264 with -c:v libx264 -tune fastdecode instead of your -qscale 8
[21:36:43 CEST] <JEEBsv> that should create AVC without in-loop filtering and using CAVLC
[21:37:27 CEST] <x86iac> what about the -crx ?
[21:37:37 CEST] <JEEBsv> -crf you mean? 23 is the default
[21:37:42 CEST] <x86iac> yes sorry
[21:37:58 CEST] <JEEBsv> basically x264 defaults to relatively sane stuff, -crf 23 and -preset medium
[21:37:58 CEST] <Mavrik> *sigh*
[21:38:10 CEST] <JEEBsv> unless your ffmpeg is old enough to need manual setting
[21:38:15 CEST] <Mavrik> This Cisco DCM hates me
[21:38:26 CEST] <JEEBsv> (and it wouldn't have the latter option in that case, even [I think])
[21:39:03 CEST] <JEEBsv> basically if you get an error from libx264 saying you're using "broken ffmpeg defaults", then you're having way too old stuff installed ;)
[22:02:46 CEST] <Elias_> kjkml
[22:02:52 CEST] <Elias_> good evening
[22:03:15 CEST] <Elias_> hey there! can anyone help me to make a ffmpeg command like this: https://www.youtube.com/watch?v=kfHDUD5DYCE best, elias
[22:04:22 CEST] <Elias_> i wanna get 16:9 scale
[22:13:30 CEST] <Weishaupt> Hey, I want to change the default audio stream in a mkv file I merged together. How do I go about that?
[22:19:00 CEST] Last message repeated 1 time(s).
[22:21:47 CEST] <Polochon_street> hi!
[23:00:30 CEST] <Elias__> hey there! i have a problem with my ffmpeg command, it works perfectly on my windows pc but not on my macbook, who can help me fix the problem? http://pastebin.com/BJ4Sqjxf thanks in advance! best, elias
[23:02:42 CEST] Last message repeated 1 time(s).
[23:02:42 CEST] <relaxed> Weishaupt: ffmpeg -i video -i audio -map 0:v -map 1:a -c copy out
[23:02:42 CEST] <relaxed> oh, the default that's played first?
[23:02:48 CEST] <Daan_> Elias__ : stop spamming please
[23:03:23 CEST] <Weishaupt> relaxed: Jup.
[23:03:55 CEST] <Elias__> hey there! i have a problem with my ffmpeg command, it works perfectly on my windows pc but not on my macbook, who can help me fix the problem? http://pastebin.com/BJ4Sqjxf thanks in advance! best, elias
[23:05:29 CEST] <Weishaupt> relaxed: It is marked (default) in ffprobe
[23:06:23 CEST] <Elias__> @Weishaupt could you help me please?
[23:06:49 CEST] <Weishaupt> Elias__: I'm a newbie to ffmpeg myself. Sorry.
[23:07:02 CEST] <Elias__> No problem!
[23:07:30 CEST] <Elias__> Can someone help me pleasE?
[23:07:35 CEST] <JEEBsv> Elias__: [mp4 @ 0x7fe121012000] Tag [1][0][0][0]/0x00000001 incompatible with output codec id '65548' ([0][0][0][0])
[23:08:03 CEST] <JEEBsv> whatever that is :P
[23:08:13 CEST] <JEEBsv> oh, it's raw pcm
[23:08:27 CEST] <JEEBsv> Stream #0:1: Audio: pcm_s24le ([1][0][0][0] / 0x0001), 44100 Hz, stereo (24 bit), 2116 kb/s
[23:08:55 CEST] <Elias__> @Jeebsv okay thank you!
[23:08:56 CEST] <JEEBsv> dunno what that all-zeroes thing is :P
[23:09:49 CEST] <Daan_> @Elias__ i have the same problem
[23:15:35 CEST] <relaxed> Weishaupt: you may have to use mkvmerge, ask in #mkvtoolnix
[23:16:35 CEST] <relaxed> er, #matroska
[23:21:14 CEST] <JEEBsv> relaxed: you actually can set it through the command line now
[23:21:30 CEST] <JEEBsv> https://ffmpeg.org/pipermail/ffmpeg-devel/2015-January/167371.html
[23:22:03 CEST] <JEEBsv> -metadata:s:v:0 disposition:default=1 -metadata:s:a:0 disposition:default=1
[23:22:05 CEST] <JEEBsv> something like this
[23:23:41 CEST] <Guest14699> JEEBsv: Nice, I'll try that then.
[23:24:21 CEST] <Guest14699> exit
[00:00:00 CEST] --- Sun Jun 21 2015