Introduction to Speech Synthesis

1
Introduction to Speech Synthesis

• Aims
• Prepare for talk
• Terminology
• Overview of Festival
• Evaluation
• Past and future directions

2
Typical Architecture
3
Festival commands

• festivalgt (set! utt1 (Utterance Text "Hello
  world"))
• ltUtterance 1d08a0gt
• festivalgt
• festivalgt (utt.synth utt1)
• ltUtterance 1d08a0gt
• festivalgt
• festivalgt (utt.play utt1)
• ltUtterance 1d08a0gt
• festivalgt
• festivalgt (SayText "Good morning, welcome to
  Festival")
• ltUtterance 1d8fd0gt
• festivalgt

4
Grammar

• After tokenisation
• Parse
• POS tagging (HMM-based n-gram)
• Disambiguation
• I live in Reading
• I can fish
• Homographs e.g. 1996

5
Dictionary/Lexicon

• Morphological analysis?
• Phonemic info
• Prosodic info
• ( "present" v ((( p r e ) 0) (( z _at_ n t ) 1)) )
• ( "monument" n ((( m o ) 1) (( n y u ) 0) (( m _at_
  n t ) 0)) )
• ( "lives" n ((( l ai v z ) 1)) )
• ( "lives" v ((( l i v z ) 1)) )
• What if no entry?
• E.g. proper nouns
• Letter-to-sound rules
• Post-lexical rules
• médecin

6
Letter-to-sound rules

• Pre-processing lexicon into suitable training set
  
• Defining the set of allowable pairing of letters
  to phones.
• Constructing the probabilities of each
  letter/phone pair.
• Aligning letters to an equal set of
  phones/_epsilons_.
• Extracting the data by letter suitable for
  training.
• Building CART models for predicting phone from
  letters (and context).
• Building additional lexical stress assignment
  model (if necessary).

7
Prosody

• Intonation phrasing
• Duration average, Klatt, CART
• Intensity
• Derived from rules?
• Learnt statistically
• ANNs
• CART models

8
Speech signal generation

• Articulatory
• Formant source filter
• Concatenative
• What units?
• Words, syllables, phones, diphones
• Unit selection
• Post-processing

9
Diphone concatenation
di?gw??
10
Evaluation

• Intelligibility
• Naturalness
• Perceptual tests
• Psychoacoustics

11
Future trends

• Best synthesis units?
• Speech signal modification
• Voice conversion
• Variability style, mood, ...
• Better models