Shubhabrata08 / AudioClassificationTFLite

This project aims to classify UrbanSound8K audio and deploy the model to a microcontroller
0 stars 0 forks source link

Generating MFCC with custom sample rate, duration and interval #4

Closed sa-paul closed 5 months ago

sa-paul commented 5 months ago

Installing dependent libraries:

I have installed these libraries using these commands.

image image

Errors while running actual code:

OSError: PortAudio library not found

image

Steps to resolve the errors:

  1. I have done these same steps in my local computer with ubuntu 22.04.
  2. The same error is showing there also
  3. I have searched for the error in google.
  4. I found that pyaudio is to be used instead of sounddevice [ Reference ]
  5. I have implemented the same but got many errors which is library specific.

image

  1. So, I sticked to sounddevice but tried another method to run through this library. [ References ] image

  2. But, got another error.

image

  1. I have modified the code, but the error is same.

    Code:

    image

    Error:

    image

  2. I have searched for the error Error querying device -1 but got no implication to my code. [ References ]

Shubhabrata08 commented 5 months ago

Error querying device denotes the inability to access the sound device/microphone. Make sure all drivers are properly installed. This is expected error in case of colab as it cannot mock an audio device input. This progress is good. You can now try performing the same tasks but now on on Arduino board.

Shubhabrata08 commented 5 months ago

@sa-paul Change title to something meaningful... Like how your additions are supposed to affect the project, what feature does it bring in... Add a description in brief as well to the PR desc

sa-paul commented 5 months ago

Description of Pull Request:

  1. Continuously records audio with recording device with specified interval.
  2. Extracts features, MFCC through librosa library in python
  3. Custom sample rate, duration, interval and hop length