[opus] Antw: Re: Antw: [EXT] Opus merging streams

Sampo Syreeni decoy at iki.fi
Wed Apr 6 13:21:28 UTC 2022

On 2022-04-06, Andrew Sonzogni wrote:

> So if I resume, it's better to decode, then use the sum of each 
> channel scaled by the square root of the number of channels.

That's if you want to maximize your usage of bit depth. But I'd still 
argue for just doing a sum. I mean, if your application is anything like 
the one I'm envisioning -- the conferencing kind -- would you want 
everybody to be able to speak at a constant volume? To shout over each 
other without the havoc being compressed to noise?

Please, just work in say 24 bits, leave some headroom, and sum directly. 

> Is that right ? Or is there some kind of trickier and better algorithms that 
> I'm not aware of ?

Sounds about right.

> Sorry if my questions seems childish, I don't have that much 
> experience in audio based algorithms and it's kind of fascinating how 
> much it can be tricky to adjust correctly some parameters or even find 
> the correct algorithms for a certain use case.

Nothing childish about it, and also, I'm no guru.

E.g. as I said already, that mix-normalization question of yours is 
*ages* old and *never* to date settled. My preferred solution wouldn't 
even have been implementable a decade back, and pretty much just 
analogue cinematic sound engineers would abide by it.
Sampo Syreeni, aka decoy - decoy at iki.fi, http://decoy.iki.fi/front
+358-40-3751464, 025E D175 ABE5 027C 9494 EEB0 E090 8BA9 0509 85C2

More information about the opus mailing list