Background information

Speech recognition

• Hidden Markov Model Toolkit manual: HTK is the most widely used package for teaching acoustic models. The manual gives a good introduction to acoustic modelling and basic Hidden Markov Models in speech recognition. The manual does not give much information about language modeling and state-of-the-art decoding (integration of information from language model and acoustic model) methods.
• Frequently asked questions in newsgroup comp.speech

Language modeling

• A bit of progress in language modeling. A comprehensive paper about language modeling from speech recognition point of view.
• Luonnollisten kielten tilastollisen käsittelyn kurssin kotisivu.

Data mining with text data

• Websom demo

Signal Processing

• Signal processing and speech analysis