srusyn - JSRU parallel-formant speech synthesizer with live output


srusyn (-I) (-p parfil) (-g glotfil) (-o outname) (-s) (-t) (controlfile)


srunsyn20 is a software implementation of a parallel-formant speech synthesizer. It is a "C"-language implementation of the FORTRAN software synthesizer of JSRU report 1016. It is based on the original SFS 'soft' synthesizer but modified to take parameter tables as input and output binary or live data. The default action is to generate and play the completed waveform using the SFS replay functions. A switch '-s' allows for simultaneous synthesis and replay on SPARC systems.

Options and their meanings are:

-I Identify program name and version number, then exit.

-p paramfile Use the parameter file paramfile instead of the default parameter file. More information on the parameter file may be found in parfil(5).

-g glottalfile Read the glottal pulse shape and glottal area from the file glottalfile instead of the parameter file. The format for a glottal file is an ASCII file containing two lists of integers separated by spaces. The first list has 72 entries in the range -127 to +127 and represents the glottal pulse shape. The second list has 18 entries in the range 0 to 15 and represents the glottal area.

-o outname Selects binary output data format and filename.

-s Selects SPARC audio output. This is performed simultaneously with conversion and so may not keep up on loaded or slow systems.

-t Selects raw hexadecimal input in format produced by text-to-speech system.


SFSBASE/data/parfil.11.ex - example parameter file.


dbsoft 1.0 - Mark Bishop and Paul Davies
2.1 - Mark Huckvale

soft 1.0s - Mark Huckvale
1.1s - Andrew Breen

srusyn20 1.0 - Mark Huckvale


soft, nsoft, JSRU report 1016 and "dbsoft" manual (Bishop & Davies)


The output sampling frequency (srate) cannot be changed (it is set in include file at compilation time).
Fri Jul 09 14:54:36 2004