[Ffmpeg-devel-irc] ffmpeg.log.20180819

burek burek021 at gmail.com
Mon Aug 20 03:05:01 EEST 2018


[00:08:14 CEST] <anden> well, just me overestimating my ability to solve this issue when my brain is really too tired for it. thanks for the help guys. i'll continue after i had some sleep
[00:09:36 CEST] <Cracki> mystery remains
[01:26:01 CEST] <qxt> what is the most supported format in browsers for video and audio? Thinking about using ffmpeg to transcode video on nextcloud servers for streaming.
[01:26:29 CEST] <qxt> kinda silly that I have to download a video to watch it imo
[01:26:39 CEST] <qxt> webm?
[01:41:30 CEST] <furq> qxt: h264/aac in mp4 is universal
[01:41:39 CEST] <furq> webm isn't supported by apple and probably won't be
[01:43:47 CEST] <qxt> thx furq.
[13:48:17 CEST] <kerio> thank fuck for the iphone
[13:48:23 CEST] <kerio> or safari on macos would be left behind
[13:48:25 CEST] <kerio> :3
[19:50:11 CEST] <ePirat> How do I correctly do 2-pass processing in ffmpeg?
[19:50:46 CEST] <JEEB> depends on what you're doing. various video encoders take in -pass 1/2/x
[19:50:56 CEST] <ePirat> ffmpeg -i .\file.wmv -pass 1 -af loudnorm=I=-13:TP=-1 -c:v prores -c:a pcm_s16le -f mov NUL
[19:50:58 CEST] <ePirat> is what I did
[19:51:19 CEST] <ePirat> and then a second pass after this
[19:51:34 CEST] <JEEB> so which thing are you trying to 2pass?
[19:51:38 CEST] <JEEB> the filter or the video encoder?
[19:51:42 CEST] <ePirat> the filter
[19:51:45 CEST] <JEEB> ok
[19:51:52 CEST] <JEEB> then it's specific to the filter
[19:52:19 CEST] <JEEB> and you wouldn't have to do anything for the video
[19:52:56 CEST] <JEEB> https://www.ffmpeg.org/ffmpeg-all.html#loudnorm
[19:53:19 CEST] <JEEB> I don't see anything regarding it saving 1st pass data. just stats which are printed, even
[19:54:34 CEST] <ePirat> JEEB, well the description has "double pass (files) modes" so I assumed it would "just work" with the 2-pass stuff&
[19:54:36 CEST] <ePirat> thats a bit confusing
[19:55:23 CEST] <JEEB> the "pass" option is an AVOption I think, and I don't think avfilters get those passed on from the cli, so the options have to be given to the lavfi filter's parameters specifically
[19:55:33 CEST] <JEEB> and the -pass 1/2/x options become irrelevant
[19:55:46 CEST] <ePirat> so I need to do the first pass, with print_format and then for the second provide data for measured_* stuff out of the json?
[19:55:50 CEST] <JEEB> and yea it says it supports both modes
[19:55:54 CEST] <JEEB> whatever that means
[19:56:04 CEST] <JEEB> since the document only mentions that stats thing, which goes to stdout?
[19:56:30 CEST] <JEEB> (also no examples mentioned in docs)
[19:57:34 CEST] <JEEB> yea the code has no mention of 2pass or anything, just the stats
[19:57:35 CEST] <JEEB> http://git.videolan.org/?p=ffmpeg.git;a=blob;f=libavfilter/af_loudnorm.c;h=314b25fa39f5fd3b1b92a5132eee0aedf3d0c651;hb=HEAD#l805
[19:58:08 CEST] <JEEB> so I guess it's a case of "parse that crap and set values accordingly"
[19:58:26 CEST] <ePirat> great :p
[19:59:07 CEST] <ePirat> JEEB, thanks for your help. should I submit a patch to improve the documentation about that?
[19:59:17 CEST] <JEEB> if you could be so kind that'd be nice
[19:59:24 CEST] <JEEB> if you notice what I mentioned actually works
[20:32:46 CEST] <ePirat> JEEB, can I somehow make ffmpeg just process the filter? I just tried to copy codec but I get: Filtergraph 'loudnorm=I=-13:TP=-1:print_format=summary' was defined for audio output stream 0:1 but codec copy was selected.
[20:33:34 CEST] <JEEB> ffmpeg -i blah -map 0:a -af blah -f null -
[20:33:38 CEST] <JEEB> is how I usually do it
[20:52:29 CEST] <ePirat> JEEB, it seems to work when providing it with the needed information.
[20:53:40 CEST] <JEEB> alright
[22:48:52 CEST] <the_fluffybread> hello
[22:49:01 CEST] <the_fluffybread> could i tell ffmpeg to output data according to an sdp file
[22:49:14 CEST] <the_fluffybread> rtp, specifically
[23:25:04 CEST] <Cracki> what did you google for?
[23:33:24 CEST] <the_fluffybread> "ffmpeg output to sdp" returns a bunch of stackoverflow question about recieving input with an sdp file
[23:50:28 CEST] <c_14> the_fluffybread: yes
[23:50:37 CEST] <c_14> use the rtp output with -write_sdp_file or something
[23:50:52 CEST] <c_14> -sdp_file
[23:51:29 CEST] <the_fluffybread> yeah but that prints sdp
[23:51:32 CEST] <the_fluffybread> i'm making a voip app
[23:51:56 CEST] <the_fluffybread> and it would be great if ffmpeg could automatically adapt to the sdp file it receives from the remote host
[23:52:14 CEST] <c_14> nah
[23:52:17 CEST] <c_14> It can't do that
[23:52:25 CEST] <c_14> You can do it with the libraries, but not with the cmd app
[00:00:00 CEST] --- Mon Aug 20 2018


More information about the Ffmpeg-devel-irc mailing list