18 Nov
2022
18 Nov
'22
6:28 p.m.
On Fri, Nov 18, 2022 at 02:10:04PM +0100, Florian Paul Schmidt wrote:
This gain can be cancelled by scaling the magnitude of
the fft bin by
1/(2*cos(theta)). This then ensures summing to the original signal.
You'd need quite a high (up to inifite) gain if the phase is near 90
degrees.
There is a much simpler solution:
Create a filter, using an inverse FFT, that doesn't modify the phase
but just has some random gain between -1 and +1 in each frequency bin.
Let F be that filter. It will be symmetric in the time domain, so have
a delay equal to half its lenght. Let D be that delay.
With X your input signal, compute
X1 = D (X) # X delayed
Y = G * F (X) # X filtered and delayed, gain G
L = 0.5 * (X1 + Y)
R = 0.5 * (X1 - Y)
where user paramter G controls the amount of effect.
Summing L and R will always give X1, the delayed input.
--
FA