for dvd- dynamic range & normalization tools and settings ?

[halsboss]

Hi a noobie seeking guidance on compressing dynamic range (I think) ...

Some video sources (usually .AVI sometimes .MP4 and .MKV) have various audio components such as mpeg2, mp3, ac3, etc. My need is to extract/convert the audio to a DVD compatible format such as .MP2 or .AC3 in 2 channel stereo (surround if nice).

I've been doing that reasonably satisfactorily with FFMPEG like this :-
ffmpeg.exe -i "Experiment.avi" -ab 224 -ar 48000 -ac 2 -acodec mp2 -y "Experiment.mp2"
or
ffmpeg.exe -i "Experiment.avi" -ab 224 -ar 48000 -ac 2 -acodec ac3 -y "Experiment.ac3"


A few (particularly from USA sources) have audio where volume at times is OK, then middlin, then low, then drifts loud again, then cycles. My guess, is that compressing the dynamic range and then normalization would fix these audios before DVD authoring.

I tried to read up, but am caught short on understanding the concepts and options and command-line tools to use (command line so as to embed in my win32/XP batch files).

Topic http://forum.doom9.org/showthread.php?t=56020 talks about normalization to -31 dBFS for DVD, and Dynamic Range Compression ... Couldn't identify new settings for FFMPEG, so any suggestions on the approach and specific tools and settings ?

Thanks

[Pookie]

Probably more effective for 2 channel audio would be to normalize (or even limit) the WAV file before transcoding to AC3/MP2. You can do that with Wavegain, or the very nice Normalize.exe app located at http://savannah.nongnu.org/download/normalize/normalize-0.7.7-win32.zip . Command line dynamic range compressors are fairly rare, although it certainly can be done in Avisynth.

normalize -l 0 --Peak -v YourWAVfile.wav

ffmpeg -i Yourwavfile.wav ....


You should consider using Aften. It is incredibly fast and the AC3 sound quality is superior to Ffmpeg.

[halsboss]

Nice. Thankyou. You also mention
Command line dynamic range compressors are fairly rare, although it certainly can be done in Avisynth.
Wondering how that works - I currently feed the AVISYNTH directly into HC which I thought didn't process audio. Does it mean I run 2 passes of the video - one into HC and then another afterward feeding into another audio product of some kind, ignoring the video component ? Even a link to elsewhere would be appreciated.