Re: [gnuspeech-contact] Re: Waveform comparison

[David Hill]

Hi Felipe,

On Mar 6, 2011, at 1:06 PM, Felipe Castro wrote:

That is strange. I wonder if the grey scale data is actually included in the build code that is compiled under GNUStep. If it isn't, the spectrograph bar would not work. There should be 101 tiff images, each one a line the width of the spectrograph bar, covering the density range from 00 (full white) to 100 (full black). The other possibility is a bug in the density indexing that only shows up in the GNUStep version. Or maybe something to do with the makefile not accessing that data, assuming there is a makefile (or lack of a make file)?

The .tif images are there, but their suffix are "tif", not "tiff". Does this could affect something?

No, that is the correct extension for TIFF files. It must be something else. I intend looking into a number of GNUStep problems as a matter of urgency, including getting the Server up and running. I had not realised that had been left un-ported.

For what purpose would you be using Cepstral analysis?

Don't worry, just wondering. Because one technique for voice recognition is applying Mel coefficients and Cepstral transformation, I think. But I'll study this stufffurther later.

Are you interested in speech recognition then? I have worked extensively on speech recognition.

For now, I'm trying to get the details of Monet and its database.

I should get the revised manual out ASAP then!

Warm regards.

david

--------

David Hill

address@hidden

http://savannah.gnu.org/projects/gnuspeech

--------

 The only function of economic forecasting is to make astrology look respectable. (J.K. Galbraith)

--------