SpiroMask Mini Project

Table of Contents

1. Overview
2. Process of making
3. Project Structure
4. Requirements
5. Installation and Setup
6. Results
7. Referrals

Overview

I made this mini project with the help of the library librosa which is used for audio processing in python mainly as I belive it will be eventually be integrated with an ml model.

Process of making

1. First of all I attached the mic on a mask such that it is as near to the mouth as possible like shown below.(placement might differ in a depending on the mic size and the mask shape but due to the constrained time I only had this mask available).

2. Recorded audio using Audacity after connecting the mic to laptop.

3. Did all the required tasks on the audio recorded.

Project Structure

The repository is organized as follows:

• mini-project.ipynb: Is the main jupyter notebook where all the code is written
• breathing.wav: Recorded breathing audio with the help of audacity
• breathing_filtered_librosa: Processed breathing audio.
• requirements.txt: For downloading required libraries.
• README.md: This README file.
• images:contain images for the README file.

Requirements

It was built using Python 3.12.0

Installation and Setup

1. Clone the repository to your local machine.

   git clone https://github.com/Jay042003/SpiroMask-Mini-Project.git

2. Install requirements.txt by the command given below.

   pip install -r requirements.txt

Results:

1. Sampling Rate found is: 44100 Hz(It was also observed while exporting the audio file)
2. Low pass filter of 5kHz cut-off frequency was used to remove higher frequencies, cut-off frequency decided by frequency domain analysis of the signal
3. It can be said that frequency range of breathing information is between 0-5kHz even in that mainly in range 0 to 2-2.5kHz, decided by observing Power spectral density of the audio signal(both processed and raw).

Referrals

Documentations

• Librosa Documentation
• Scipy

Youtube videos

• Audacity