Subject: Speech Processing

Topic: Speech Processing Applications

Difficulty: High

(i) The front end does two function, the first of which is to convert raw text containing symbols, such as, number and abbreviations into equivalent words.

(ii) This process is termed as normalization or pre-processing or tokenization.

(iii) The second function is to assign phonetic transcription to each word. The front end converts each phonetic until of the text into units like phrase, clauses and sentence.

(iv) This process of assigning phonetic transcription to words is called text-to-phoneme conversion.

(v) Phonetics transcription and prosody information together will form the symbolic linguistics representation.

(vi) The back-end will also include the computation of the target prosody such as pitch counter, phoneme duration which are then imposed speech output.

(vii) The back-end is the topic of discussion for the current chapter.

(viii) It is crucial component so far as intelligibility and quality of speech is consider.