LING 15 Lecture Notes - Lecture 17: Phoneme, Speech Synthesis, Total Synthesis
Document Summary
5/30: written qs available on gs all of finals week, assignment #5 due monday @ 9am. Technology & language: modeling language behavior: speak commands instead of typing/mouse; computer could respond in language instead of whatever, automating language analysis, other benefits: search applications, research, error detection. Text-to-speech: text to speech: process of converting words and/or text to auditory output, 2 basic approaches (not only ones): whole-word, phonemes, both approaches require stored data; program looks up what sounds to make, whole word. More time spent sorting thru words to get to right one (not efficient) Words with multiple forms (walk, walked, walks) need diff recording for each word. String of words = must look up each word separately and play them in sequence. Problems: need multiple versions of each word ( hello world vs. Record individual phonemes ( h" or ck" sound); 40 phonemes = 40 recordings.