Version 7 (modified by sach01, 14 years ago) (diff) |
---|
Support for explicit prosody specifications
MARY TTS support explicit prosody specification using 'prosody' element as described in W3C Speech Synthesis Markup Language (SSML) recommendations; and the different attributes in 'prosody' element like 'rate', 'pitch' and 'contour' are used as specifications to modify predicted phone durations and pitch contour before passing them to synthesizer.
Example:
<?xml version="1.0" encoding="UTF-8" ?> <maryxml version="0.4" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns=" http://mary.dfki.de/2002/MaryXML" xml:lang="en-US"> <p> <prosody rate="fast" pitch="+10%" contour="(10%,low)(50%,-20%)(80%,+10%)(100%,+5st)"> Welcome to the world of speech synthesis! </prosody> </p> </maryxml>
1. Adjusting Speech Rate
Example 1:
<?xml version="1.0" encoding="UTF-8" ?> <maryxml version="0.4" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns=" http://mary.dfki.de/2002/MaryXML" xml:lang="en-US"> <p> <prosody rate="200%"> I am now speaking twice as fast. </prosody> </p> </maryxml>
Example 2:
<?xml version="1.0" encoding="UTF-8" ?> <maryxml version="0.4" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns=" http://mary.dfki.de/2002/MaryXML" xml:lang="en-US"> <p> <prosody rate="70%"> I am now speaking 30% more slowly. </prosody> </p> </maryxml>
2. Adjusting Pitch level
Example 1:
<?xml version="1.0" encoding="UTF-8" ?> <maryxml version="0.4" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns=" http://mary.dfki.de/2002/MaryXML" xml:lang="en-US"> <p> <prosody pitch="low"> I am now speaking with 3/4 pitch. </prosody> </p> </maryxml>
Example 2:
<?xml version="1.0" encoding="UTF-8" ?> <maryxml version="0.4" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns=" http://mary.dfki.de/2002/MaryXML" xml:lang="en-US"> <p> <prosody pitch="-5st"> This is five semitones lower. </prosody> </p> </maryxml>
3. Shaping intonation contour
Example 1:
<?xml version="1.0" encoding="UTF-8" ?> <maryxml version="0.4" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns=" http://mary.dfki.de/2002/MaryXML" xml:lang="en-US"> <p> <prosody contour="(0%,x-low)(30%,low)(50%,medium)(70%,high)(100%,x-high)"> I am talking with rising intonation. </prosody> </p> </maryxml>