MFCCs + Decoding

[Isobel W]

Am I right in thinking we have 39 MFCCs per 25ms frame (12 mfccs, 13 deltas, 13 delta-deltas ?). Are we expected to explain the reasons why we used MFCCs in our report

I’m also a little confused about what our language model and acoustic model is in the context of this sentence about viterbi decoding :

Combine probabilities from the acoustic and language models to determine the overall most probable word sequence

[Simon]

Yes, there are a total of 39 elements in the feature vectors. 12 of them are the MFCCs. You have 12+13+13=38 though.

The language model computes P(W) where W is the word sequence of one utterance to be recognised. For the digit recogniser, can you locate and inspect the language model that is being used?

The acoustic model computes P(O|W) where O is the observation sequence. How does O relate to the MFCCs?

How are P(O|W) and P(W) combined to calculate P(W|O), and why do we need to do that?

[Author]

Posts