search

abhijeet3922 / Speaker-identification-using-GMMs

It uses GMM to train a speaker identification model. The training and testing has been done on subset (34 speakers) from VoxForge data corpus.
https://appliedmachinelearning.wordpress.com/2017/11/14/spoken-speaker-identification-based-on-gaussian-mixture-models-python-implementation/
55 stars 37 forks source link

MFCC Doubt #11

Open pranjalnaman opened 4 years ago

pranjalnaman commented 4 years ago

Hi! The work you have done is wonderful. I need a little help regarding the MFCC function. While taking MFCC, I seem to get a 2D array of n rows and 20 columns. But the idea I got from your blog was that I would be receiving a 1D array of size 20. Could you help?

abhijeet3922 commented 4 years ago

It will be no_of_frames * 20.

Each frame will have 20 mfccs. Speech file is chunkized into 25 msec with 10 msec overlapping frames. So for 1 sec audio, you get 10*20 arry 2D