You are here:Home Courses Archived courses Speech Processing (2020-21) Topic map & slides

Topic map & slides

The topic map will help you understand the scope of the course, and how all the topics fit together. There are also copies of the slides used to make the topic videos.

On this page:

Slides

Here are the slides for the topic videos in

Here are the live class slides for

Topic map

This map is a new feature and therefore something of a work in progress. Click on a topic in the map to find related course content. This only links to Simon’s videos and not the PHON content from Rebekka. There is also a printable version.

Theory Application
Speech Signal processing Probabilistic modelling Speech Synthesis Automatic Speech Recognition
Signals Production Perception Front end Waveform generation Feature extraction Pattern matching Hidden Markov Models Connected speech
Concepts Time domain Sound source Pitch Digital signal Describing data Tokenisation & normalisation Waveform concatenation Series expansion Exemplar Generative model of sequences Hierarchy
Periodic signal Harmonics Cochlea Short-term analysis Discrete & continuous variables Pronunciation Features Distance Sub-word unit
Frequency domain Vocal tract resonance & formants Mel scale Spectral envelope Joint, Conditional, Bayes’ formula Prosody Feature engineering Sequence Hidden state sequence N-grams
Models & data structures Filter Resonant tube Filterbank Impulse train Gaussian Finite state transducer Feature vector Sequence of feature vectors Hidden Markov Model
Impulse response Source-filter model Phoneme Pitch period Generative model Decision tree Diphone Grid Lattice Graph
Algorithms & analysis Convolution Fourier analysis Fitting a Gaussian to data Handwritten rules Overlap-add MFCCs Dynamic programming (DTW) Dynamic programming (Viterbi) Composition (“compiling”)
Cepstral analysis Classification Learning decision trees TD-PSOLA Baum-Welch Approximation (pruning)

Themes

Periodic signals in the time domain The vocal tract is a filter Frequency domain and beyond Sound categories Probability Miscellaneous Interpretable methods Sequences of variable length Generative models Finite state networks

The Big Picture

Here is how the course teaches those topics. Click the image to see a vector-graphic PDF version suitable for printing. Before starting a topic, make sure you understand the preceding topics that feed into it.