In this course, the subjects of Speech Recognition and Synthesis will be taught and the applications about the course will be developed. Firstly, speech production and acoustic modeling will be presented. Then, The methods of Speech Recognition and Synthesis will be explained.

Acoustic Theory of Speech Production, Human hearing, acoustics, and phonetics. Signal Representation, Vector Quantization. Speech spectrum analysis (Fourier analysis, cepstral analysis, spectrogram reading). Fundamental frequency analysis (F0 estimation, prosody models). Speech synthesis. Linear Prediction (all-pole model, LPC, PARCOR, LSP analysis). Learning algorithms and application (Viterbi algorithm, Bayes’ Theorem). Speech coding (waveform coding, PCM, LPC). Dynamic time warping and acoustic modeling. Hidden Markov Modeling, expectation-maximization, and search. Language Modeling. Graphical Models. Segment-Based ASR, Finite State Transducers.