0
$\begingroup$

I have not worked with MFCCs before, but I am faced with the observation that the 5th MFCC of a signal is actually of interest for me in my speech research. In order to understand why that one is important for me, I need to make sure that I understand what it represents.

If I know the number of Mel frequency spaced filters applied to the signal, and the sample rate, can I assume that the 5th coefficient is related to the 5th mel spaced filtered regions?

If not, how do I learn what in the frequency domain is represented by the 5th MFCC?

$\endgroup$
0
$\begingroup$

In my understanding you need a little more information. You need to know the minimum and maximum frequency and number of coefficients in the MFCC representation.

If the coefficients are linearly spaced in the mel scale, $m(f)$

The first coefficient is $m(f_1)$ the last coefficient is $m(f_N)$, and the $i-$th coefficient is

$$m(f_i) = m(f_1) + \frac{i-1}{N-1} (m(f_N) - m(f_1))$$

Then you replace apply the definition of interest for $m(f)$, e.g.

$$m(f) = \frac{1000}{log(2)}log\left(1 + \frac{f}{1000}\right) $$

And solve for $f_i$.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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