Hey Alexandros,

That's generally referred to as an Equal Power Crossfade, as you can see in your results it's to keep the perceived loudness of your crossfade output constant. For correlated signals this is generally better - you could also use a logarithmic function (and it's inverse). In the case of an un-correlated signal a linear crossfade might be more suitable.

Cheers,

Joe

On 16 February 2015 at 21:52, Alexandros Drymonitis <adrcki@gmail.com> wrote:

I was checking Integra Live's modules and a Mixer module uses this technique to mix two signals:

[0 \ <- value between 0 and 1
|
[* pi/2]
|
[t f f]_____
| [sin]
[cos] | <- multiplied by the second signal
|
| <- multiplied by the first signal

The above ASCII patch is a simplified version focusing on the basic technique. Doing the same and sending the result to [env~] stays between 95 and 98 approximately, but graphing it in an array, it exceeds the limits (-1, 1) by 0.4 approximately.
Is this the way to mix two signals, or is there something else I should be looking for?

_______________________________________________
Pd-list@lists.iem.at mailing list
UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list

Follow me on Twitter @diplojocus