tebasuna51 [Question about downmix]

[marcio]

What would be the correct way to downmix multichannel to stereo?

I am using this:
eac3to input.xxx output.wav -down2 -phaseshift -mixlfe -down16 -16 -48000 -keepdialnorm

-down2: stereo target
-phaseshift: sometimes recommended when downmixing to 2 channels. Should I really use it?
-mixlfe: dolby don't recommend it. Should I really use it?
-down16 and -16: what should I use? I want to get 16 bit wave file.
-48000: to make sure wave file is 48000. Should I really use it?
-keepdialnorm: Is this a good idea or not?

http://img833.imageshack.us/img833/6417/eac3to.jpg

Since I have everything working with eac3to, should I also use -sonic or -nero?

Thanks.

[tebasuna51]

If there are options is because don't exist a "correct" way.
I don't know your source and I don't know for what you need a stereo wav file 48000 Hz 16 bits.

-down2: Dolby ProLogic II downmix
-phaseshift: Never with ac3 source. Optional with other source.
-mixlfe: can produce artifacts, or not. Optional.
-down16: to get 16 bit wave file.
-resampleTo48000: to make sure wave file is 48000.
-keepdialnorm: here do nothing, only when extract an ac3 track or apply a delay over an ac3 without recoding. I never recommend -keepdialnorm.

My first attempt is always:

eac3to input.xxx output.wav -down2 -normalize -libav

You can add:
-down16 -resampleTo48000

I use -libav for that. I never test sonic. I'm not sure if -nero apply DRC.
-normalize: needed after a downmix to be sure you obtain the max volume output.

[marcio]

Thanks man.

I could only trust you =]

I will be using this for safety:
eac3to input.xxx output.wav -down2 -normalize -libav

and this for risk:
eac3to input.xxx output.wav -down2 -mixlfe -normalize -sonic

If dialog turn out too soft meaning dynamic range is too high, I will use some compression filter like the one in goldwave.

Thanks again, you are the man =]