Subject: Speech Processing

Topic: Speech Processing Applications

Difficulty: Low

The system operates in two modes:

a) Training mode; b) Synthesis mode

(i) The training mode record different words. Phonemes of appropriate duration are cut from the recorded words and stored in the form of a lookup table addressed using a six-bit code.

(ii) Another lookup is stored containing the phonetic sequence for each word in vocabulary.

(iii) The word to be synthesized is used to find the phonetic sequence from the store.

(iv) The phonemes from the phonetic sequences will be retrieved from the memory addressed using a six-bit code.

(v) The recorded phonemes are accessed from the memory and concatenated in a proper sequence is given in the lookup table for a phonetic sequences.

(vi) The concatenated word is then filtered and is output from the system.

(vii) It was observed that the synthesized words sounded almost natural.