Anyone knows something about DSP here?
I am trying to measure the similiarities between the power spectrum of two audio signals, but I don't really know how to do it.
I basically want to end up with a plot with peaks where the spectrum is the same.
I would like to understand how to do it mathematically, and then use scilab or matlab to do the computation.
So, any help or reference would be hightly appreciated.
>>8384084
Do you currently have them in the time domain?
You don't care about phase at all or do you?
U just need dat FFT https://www.mathworks.com/help/signal/ref/spectrogram.html
>>8384265
Hey,Sorry I just came back.
Yes, I exported them to scilab with the command
[x,S,bits] =wavread("D:\Trabajos\Musica\Proyecto delay\basic delay sin.wav");
So, the signal would be on the variable "x".
Let me explain a bit further.
One is the original signal (an audio track)
The other one is the original signal with a simple FIR delay applied.
The transfer function of the delay (basically a comb filter) is H(z) = 1+gz^(-M), where M = delay time * sampling frequency
I want to analyze bot signals to prove in practice the theoretical results of the analysis of the filter, that is:
-It affects the frequency domain, as a comb filter
-The group delay also has the form of a comb, introducing phase distortion
Thanks.
>>8384084
you can read more about dsp here
https://encyclopediadramatica.se/DarksydePhil
>>8384084
This is not trivial. You can start by reading up on target factor analysis and principal component analysis. And try not to get bogged down in the small controversies regarding the utility of them.
>>8384292
>This is not trivial.
I am a bit in a hurry....but I guess I will have to enter deep into the topic if there is no other way to do it....
>>8384330
Okay I will, thanks.
Should I read and understand this tutorial too?
http://es.mathworks.com/help/signal/examples/measuring-signal-similarities.html
It seems useful. But I will have to adapt everything to scilab, since is what I am using for this project.
>>8384342
what you found is better actually
I don't really use MATLAB so the spectrogram function was just the first catchall approach I could find
In fact the page you found shows you how to directly generate a "Power Spectrum" for each signal. Here is a more specific link https://www.mathworks.com/help/signal/ug/power-spectral-density-estimates-using-fft.html
This is a more direct approach than what I suggested, but spectrograms are fun too.
>>8384342
sorry, i just realized you're using scilab
I think what you need is "pspect"