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/
54 stars 37 forks source link

create new audio file #8

Open ghost opened 5 years ago

ghost commented 5 years ago

I want to create new train files and test files. But I always get this error. "WARNING:root:frame length (1200) is greater than FFT size (512), frame will be truncated. Increase NFFT to avoid."

The audio files are 3 seconds. The times and sizes are quite small. How do I fix this error?

Thank you,

abhijeet3922 commented 5 years ago

Can you provide me more info ?

  1. Sampling rate
  2. Frame length in seconds that you are passing to mfcc()
  3. What is the shape of features you are getting from this 3 seconds audio ?
ghost commented 5 years ago

Iam sorryi I am late. I solved the problem. I change this value. (nfft=1536 ) and solved problem.

ghost commented 5 years ago

thank you for response.