1
$\begingroup$

I try to combine my 7 band FIR filters into 1 FIR in order to process an audio signal. I use convolution to combine their coefficients but the effective is not the same as I apply 7 different FIR filter. I also have the gain problem

H = conv(conv(conv(conv(conv(conv(h1,h2),h3),h4),h5),h6),h7)

My questions are Is convolution a correct way to do combine multiple FIR filters into one filter If not, then what technique should I look into? If yes, then what did I do wrong?

Regards

Improve this question
$\endgroup$
2
  • 1
    $\begingroup$ Are you combining the 7 different band filters in series or in parallel? $\endgroup$
    – hotpaw2
    Commented Jan 25, 2015 at 5:28
  • $\begingroup$ In parallel, they are filters for equalizers. $\endgroup$
    – user7226
    Commented Jan 25, 2015 at 5:28

1 Answer 1

Reset to default
4
$\begingroup$

What you tried to do in your example is combine the filters in series rather than parallel.
That is, if your signal is $$y = ((x*h_1)*h_2)\ldots*h_7$$ then, according to the associativity property of the Convolution, $$y=x*\underset{H}{\underbrace{(h_1*h_2\ldots*h_7)}}$$

However, if you wish to perform a parallel filtering then your signal is $$y=\sum_{i=1}^7 x*h_i$$ and acording to the distributivity property of the Convolution, $$y=x*\underset{H}{\underbrace{\sum_{i=1}^7 h_i}}$$

Improve this answer
$\endgroup$
1
  • $\begingroup$ thank you very much. I never thought it would be just a sum of all coefficients. $\endgroup$
    – user7226
    Commented Jan 25, 2015 at 7:58

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.