12.12.2012 Views

Festival Speech Synthesis System: - Speech Resource Pages

Festival Speech Synthesis System: - Speech Resource Pages

Festival Speech Synthesis System: - Speech Resource Pages

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

( [ g u ] "'" EI = g )<br />

( [ g u ] EI = g )<br />

( [ g ] = g )<br />

( [ h u e ] = u e )<br />

( [ h i e ] = i e )<br />

( [ h ] = )<br />

( [ j ] = x )<br />

( [ k ] = k )<br />

( [ l l ] # = l )<br />

( [ l l ] = ll )<br />

( [ l ] = l )<br />

( [ m ] = m )<br />

( [ ~ n ] = ny )<br />

( [ n ] = n )<br />

( [ p ] = p )<br />

( [ q u ] = k )<br />

( [ r r ] = rr )<br />

( # [ r ] = rr )<br />

( LNS [ r ] = rr )<br />

( [ r ] = r )<br />

( [ s ] BDGLMN = th )<br />

( [ s ] = s )<br />

( # [ s ] C = e s )<br />

( [ t ] = t )<br />

( [ w ] = u )<br />

( [ x ] = k s )<br />

( AEO [ y ] = i )<br />

( # [ y ] # = i )<br />

( [ y ] = ll )<br />

( [ z ] = th )<br />

))<br />

We could simply set our lexicon to use the above letter to sound system with the following command<br />

(lex.set.lts.ruleset 'spanish)<br />

But this would not deal with upper case letters. Instead of writing new rules for upper case letters we can define that a<br />

Lisp function be called when looking up a word and intercept the lookup with our own function. First we state that<br />

unknown words should call a function, and then define the function we wish called. The actual link to ensure our<br />

function will be called is done below at lexicon selection time<br />

(define (spanish_lts word features)<br />

"(spanish_lts WORD FEATURES)<br />

Using letter to sound rules build a spanish pronunciation of WORD."<br />

(list word<br />

nil<br />

(lex.syllabify.phstress (lts.apply (downcase word) 'spanish))))<br />

(lex.set.lts.method spanish_lts)<br />

In the function we downcase the word and apply the LTS rule to it. Next we syllabify it and return the created lexical<br />

entry.<br />

[ < ] [ > ] [ > ] [Top] [Contents] [Index] [ ? ]<br />

24.2.3 Phrasing<br />

Without detailed labelled databases we cannot build statistical models of phrase breaks, but we can simply build a<br />

phrase break model based on punctuation. The following is a CART tree to predict simple breaks, from punctuation.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!