[sdiy] Waveform mixing - normalization?

Timothy Daugard daugard at cox.net
Tue May 10 17:40:45 CEST 2016


On 5/9/2016 6:14 PM, neil harper wrote:
>> I interpreted the original poster's use of "normalising" to mean "making
>> all the waveforms sound like they're the same volume to a human" which
>> is not the formal definition of normalising, but which I thought was the
>> intent of the original comments.  Are you suggesting he means "keep the
>> same total volume when two or more waveforms are selected and mixed" (by
>> attenuating each waveform when mixing more than one?)
>
> Hey guys, sorry maybe I used the wrong term. I mean consistent 
> perceived volume from a waveform mixing section. So lets say there's 3 
> different waves (triangle, sine, square). If you're running the 
> triangle alone, it sounds exactly the same volume as 'triangle with a 
> bit of square', which sounds exactly the same volume as 'full 
> triangle, full sine, full square'.

SIne = one frequency
Square, triangle = infinite (multiple) frequencies

Volume perception would be a function of coloration of following 
circuits and hearing frequency response of the listener. About the best 
that could be done would be equal power - at the fundamental frequency - 
of the signals.

That would be relatively easy to do with a trimmer on each waveform output.

Of course this also ignores the Munson-Fletcher loudness curve effects. 
(Volume changes change harmonic response.)

Timothy Daugard



More information about the Synth-diy mailing list