[FFmpeg-user] sound volume normalisation

Bouke (VideoToolShed) bouke at videotoolshed.com
Tue Jun 25 20:20:38 CEST 2013


----- Original Message ----- 
From: "jacky" <renaux.jacky at orange.fr>
To: "FFmpeg user questions and RTFMs" <ffmpeg-user at ffmpeg.org>
Sent: Tuesday, June 25, 2013 7:56 PM
Subject: [FFmpeg-user] sound volume normalisation


> Hi all gourous
>
> I do request your knowledge and advices
>
> I am getting flv files encoded from ffmpeg
> and then I had to normalize sound levels
> since few weeks i had flv (H264+mp3) and it was fine
> and a fast process
>
> 1.   demux video , demux audio
> 2.   run lame to find out the sound level
> 3.   remux saved video + audio with volume the corrected volume level
>
> now I am getting flv (h264+aac)
> and doing the new cascading almost ffmpeg runs
>
> 1.   demux video
> 2.   demux aac audio
> 3.   demux and convert aac to mp3
> 4.   run lame on mp3 to find volume level
> 5.   remux saved video + saved aac with corrected volume value
>
> the second process takes a lot longer
> do I make a mistake , is there any advices to speed up
> is there a ffmpeg filter option which compete with lame ?

I'm not a Guru, but nevertheless:
Ffmpeg nowadays has the R128 norm build in. This is a great way to find the 
human subjective loudness, instead of finding peaks.
So you can analyze first, then transcode / rewrap with a volume change. 
(Keep in mind that 1 LU roughly corresponds with 1 dB)
It still will be a two pass process, but without demuxing first.
And since it's only sound to walk trough / re-encode, it should be lightning 
fast.
(Granted, if you need to do the entire YouTube content it might take a day 
or so...)

hth,

Bouke

VideoToolShed
van Oldenbarneveltstraat 33
6512 AS  NIJMEGEN, the Netherlands
+31 24 3553311


> regards
>
> jacky
>
>
> _______________________________________________
> ffmpeg-user mailing list
> ffmpeg-user at ffmpeg.org
> http://ffmpeg.org/mailman/listinfo/ffmpeg-user 



More information about the ffmpeg-user mailing list