VOICES_info.txt

Cotovia is a unit-selection text-to-speech system, i.e., generates the synthetic speech signal as a
concatenation of prerecorded segments. Cotovia determines the sequence of sounds, its intonation and 
duration from the input text. Actually several intonation contours are considered in parallel and for
each one a sequence of speech units is selected. The final intonation contours is selected according to 
the suitability of the sequence of speech units found. Obviously the computational load and, therefore,
the time of execution depends on the number of speech units of the speaker and the number of parallel
intonation contours considered. The number of parallel intonation contours iniatially considered may be
 specified by the optional parameter -N (default 50), although the speech unit search is only performed for 
the M most different contours (parameter -M, default 10). Of course, the quality also depends on the number
 of speech units  and on the number of parallel intonation contours.

At this time two Galician voices are available at sourceforge. These two voices are also suitable for Spanish,
 with little distortion, since the phonemes of Spanish may be considered a subset of Galician phonemes. The
 nicknames of the two speakers are Iago and Sabela. Nevertheless you will find three voices available for
 Cotovia:

* iago : Galician male speaker, duration of the recording: 80 minutes. It was Cotovia's first voice and the
 quality is limited by the reduced recording time.

* sabela-large : Galician female speaker, duration of the recording: 14.5 hours. Use this voice if you want to
 obtain maximum quality speech and you are not too worried about execution time.

* sabela: default speaker. A subset of Sabela's recordings (about 4 hours). A good compromise between quality
 and execution time.

 An additional Spanish speaker, free for noncommercial use, is available through external links (see Cotovia Blog
 for further information):

* david: Spanish male speaker, recording duration about 2.5 hours.


Examples:
_________


cotovia -i input_text.txt -s   (default values:   sabela voice, N = 50, M = 10)
cotovia -i input_text.txt -s -V sabela    (sabela voice, default values: N=50, M=10)
cotovia -i input_text.txt -s -V sabela_large  (sabela-large voice, default values: N=50, M=10)
cotovia -i input_text.txt -s -V sabela_large -N 1 -M 1 (sabela-large voice,N = 1, M = 1)
cotovia -i input_text.txt -s -V david -l es