Manual for the Automatic Prosodic Aligner (ProsAlign)
This is a manual for the automatic prosodic aligner (ProsAlign) program. This program may be used to label automatically prosodic events in a speech file. The input is a structured file with continuous values for F0, voicing, and RMS-amplitude (usually an ESPS/waves *.f0-file). The output will be a structured label file with the type of prosodic event and its position in time.
The program is currently in development and not all the criteria are implemented. Nevertheless it is useable and the author is interested in getting some feedback about the programs performance.
The program runs (currently solely) on felsenpinguin (Linux machine) and to use it simply type
in a shell. The infile is given without its usual extension .f0 and is usually an ESPS/waves F0-file generated by get_f0 (version 1.14). The output will be a label file in ESPS/waves-format and with the extension .tonesx_gtobi.
When you want to check the automatically generated label file there is a program called showproslab which opens simultaneously the waveform (with assumed extension *.d), the F0-file and the *.words and *.tonesx_gtobi label files. The prosalign program prints a message after the succesful processing of a file that offers the command line how to use the showproslab program.
The view created by showproslab is illustrated in the following picture:
The taskflow within the program is illustrated in the following picture:
There is also a man-page for the program which is accessible by the usual way:
If you need more information you may send an Email to ProsAlign-maintainer
© Norbert Braunschweiler