Home My Page Projects MPTK: The Matching Pursuit ToolKit
Summary Activity Forums Tracker Lists Docs News SCM Files

Forum: help

Monitor Forum | Start New Thread Start New Thread
RE: Creating new dictionaries [ Reply ]
By: suresh kumar on 2013-06-03 06:20
[forum:116343]
Hi Boris,
As you mentioned <windowopt=4.085> how does this 0.805 is related to sampling frq of 16k.

RE: Creating new dictionaries [ Reply ]
By: Nicolas Vinuesa on 2012-10-22 09:55
[forum:110177]
Dear Boris,
I have a new problem. In the gammatone function i have to implement the bandwith depends of the central frequency by a ERB function (ERB=24.7 + 0.108*fc). So my function would be something like (for order 4 n=4): g(t)=a*t‎‎^(n-1)*e^(-2*pi*1.019*ERB)*cos(2*pi*fc*t + phase).
I can't see how to make the bandwith b dependant of fc in the dictionary.
Thanks a lot!

RE: Creating new dictionaries [ Reply ]
By: Nicolas Vinuesa on 2012-10-15 09:13
[forum:110166]
Thanks a lot Boris!, i had no idea.
After i finish and check the results i will see if it works for me.

RE: Creating new dictionaries [ Reply ]
By: Boris Mailhe on 2012-10-12 17:16
[forum:110160]
Hi Nicolas,

There is already a gamma window implemented in MPTK so you might not need to code anything. You should be able to get a gammatone block by building a Gabor block with the following tags:

<windowtype=gamma>
<windowopt=4.085> (for example)

The optional parameter contains both the window order (here 4) and bandwidth/10000 (here 0.085 means the bandwidth is 850). For now it seems to assume that the signal is sampled at 16000Hz.

Hope it helps,
Boris

Creating new dictionaries [ Reply ]
By: Nicolas Vinuesa on 2012-10-12 15:08
[forum:110158]
Dear all,
I am new at MPTK so i am still in the process of understanding how it works, but my idea is to implement a decomposition over a gammatone functions basis.
So my question is if it is possible to create a new dictionary for a totally different kind of function (as the gammatone).
Thanks for your time