2021-10-14 04:09 AM
hi, am modifying the ASC python code for keyword spotting. Am using dataset which is 1 sec long wav files whereas ASC dataset is 30sec long wav files.
while doing spectrogram-feature extraction am getting value 0.
what is the frame length and hop length I need to give for 1sec long audio?
Also, whether I can use the same log mel feature extraction module for this?
Please help.
Python Script link-
https://colab.research.google.com/drive/1lBca6J1kUTNG8MmaM_TTPV8M-5IXdy7Y#scrollTo=1pZBm4W3khO8
2021-11-08 12:03 AM
Hi @ShivaRagu
I am not the owner of that script, but the purpose is to have 1024ms frame to be in the power of 2 (to ease computation). In the script, the frame length should be adapted to 1s to fit the dataset length. Where you able to solve you issue ?
To give better visibility on the answered topics, please click on Accept as Solution on the reply which solved your issue or answered your question.