Welcome to the world of Scilab Speech
function list--Speech Synthesis
1．find_pmarks
2．plot_pmarks
3．detect_vuv
4 ．tdpsola
1．find_pmarks
----calculates and returns the pitch marks (placed at peaks in the short-time energy function) for the input speech signal.
Calling Sequence:
pitch_marks = find_pmarks(wavefile, fs_in);
Parameters:
wavefile: wave file. Vector or matrix.
fs_in: the frequency of the input speech signal. Integer.
pitch_marks: the pitch marks. Vector or matrix.
Description:
The fs_in is the frequency of the input data, you should get its value first.
Examples:
[x, fs]=wavread('d:\test.wav');
pitch_marks=find_pmarks(x, fs);
back to top 
2．plot_pmarks
----draw some lines in the place where the pitch frequency is.
Calling Sequence:
plot_pmarks(wavefile, pm);
Parameters:
wavefile: wave file. Vector or matrix.
pm: the value that returned by the "find_pmarks" function which is described above. Description:
This function plot pitch marks of a speech signal. The parameter pm is the returned value of the "find_pmarks" function.
Examples:
[x, fs]=wavread('d:\test.wav');
pitch_marks=find_pmarks(x, fs);
plot_pmarks(x, pitch_marks);
back to top
3．detect_vuv
----return voiced/unvoiced decisions considering pitch marks.
Calling Sequence:
[vuv]=detect_vuv(wavefile, fs, pm);
Parameters:
wavefile: wave file. Vector or matrix.
fs: the frequency of the input speech signal. Integer.
pm: the value that returned by the "find_pmarks" function which is described above.
vuv: Matrix. Description:
This is the voice/unvoiced decision function.
Examples:
[x, fs]=wavread('d:\test.wav');
pitch_marks=find_pmarks(x, fs);
[vuv]=detect_vuv(x, fs, pitch_marks);
back to top
4 ．tdpsola
----adjust the prosodics.
Calling Sequence:
y = tdpsola (wavefile, fs, pscale, tscale, pm, vuv);
parameters:
wavefile: wave file. Vector or matrix.
fs: the frequency of the input speech signal. Integer.
pscale: the pitch scale ratio. Integer.
tscale: time scale ratio. Integer.
pm: the value that returned by the "find_pmarks" function which is described above.
vuv: the value that returned by the "detect_vuv" function which is described above.
y: signal. Vector or matrix. Description:
This function adjust the prosodics so that it can change the basic elements of the human speech.
Examples:
[x, fs]=wavread('d:\test.wav');
pitch_marks=find_pmarks(x, fs);
[vuv]=detect_vuv(x, fs, pitch_marks);
y = tdpsola (x, fs, pscale, tscale, pm, vuv);
back to top |
