Mel Filter Bank

[qureshi000]

I am trying to make Mel Filter Bank that I got from research paper. This Mel filter bank is for 20 filters but some of the connections are not clear. Kindly guide me about it. The pic of filter given in research paper is attached.

 

[GerdW]

Hi qureshi,

---

@qureshi000 wrote:

I am trying to make Mel Filter Bank that I got from research paper.

---

Oh boy, such a block diagram image in a research paper!?

---

@qureshi000 wrote:

This Mel filter bank is for 20 filters but some of the connections are not clear. Kindly guide me about it.

---

What is "not clear" to you?

Do you know the math behind your filter bank? (Then it should be easy to "fix" to bad wiring style of that "scientist"…)

[qureshi000]

Yes, this block diagram is in research paper. I do not know math behind Mel Filter so that I can fix bad wiring myself. The wiring connections which I do not understand are encircled with red pen in the pic attached.

Mel filter is applied for 20 filters coming from the Fast Fourier Transform signal. 20 filters are assumed by us. Assuming no of filter is our choice. A Mel scale to the power spectrum to extract the frequency bands. The aim of the mel scale is to mimic the non-linear human ear perception of sound.. It is time-frequency distributor. Figure 6 describes that Mel filter bank for 20 filters. In that we create subVI for Mel filter bank. This is info given about Mel Filter in the research paper.

[GerdW]

Hi qureshi,

---

@qureshi000 wrote:

Yes, this block diagram is in research paper.

---

Ask the author(s) of that paper for a better explanation and a cleaned version of their VI!

Or read all the papers referenced by that "research"…

---

@qureshi000 wrote:

The wiring connections which I do not understand are encircled with red pen in the pic attached.

---

• The left circle is wired to some code, hidden behind the left FOR loop.
• There are 3 circles before the last loop: the upper wire is connected to the lower input tunnel (I guess). The circle in the center shows the wire from the output tunnel of the loop before to the encircled input tunnel (I guess).
• Guess for the two circles to the right: there is a "exp(2*x*x) / (- (difference)^2 )" computation.

All my guesses can be checked against the math/formulas for your Mel filter…

[AnalogKid2DigitalMan]

No Mel Blank here:

https://www.ripublication.com/ijaer18/ijaerv13n10_117.pdf

That's All Folks!

-AK2DM

[jjohn1]

The "M" constant may be connected in some fashion to the N terminal on the first for loop (which has a coercion dot on it).  It seems like this would be redundant with the autoindexing array input.

[qureshi000]

@AnalogKid2DigitalMan
Yes, this is the research paper. The wiring connections which I do not understand are encircled with red pen in the pic attached.