Need help with speech recognition algorithms?

Question

Speech recognition

I want to understand the topic. I can well imagine how decoding takes place from phonemes to the level of words and gammatics. But I am absolutely not guided in the first stages of signal processing: signal -> vector of features -> phonemes.
Maybe someone has done it. It would be interesting to link to articles, as well as tools for decoding a signal into a vector of features and phonemes, and corresponding dictionaries. I looked at cmu-sphinx diagonally, I didn’t figure it out from the raid.
Thank you.

Answer 1

You can start with the HTK Toolkit for example . There is a link to a book in which the theory comes first. The toolkit itself uses hidden Markov models for recognition.