Skip to content

Tool for synthezing audio files for multi-microphone UCA, neural network & benchmark - master thesis

Notifications You must be signed in to change notification settings

Satiskylle/UCA_microphone_audio_generation_tools_CNN_and_Benchmark

Repository files navigation

Uniform Circular Array wav-files synthesis library in python.

Table of Contents

Introduction

Complete set consist of 4 scripts described below. For more information, please, analyze function's documentation inside of the scripts.

Scripts

uca_matrix_files_synthese.py

Script used to generate .wav files for N microphones, based on number of microphones, matrix radius and audiowave arrival angle. Audiowave arrival angle is set from 0-359, where 0 is northern microphone

neural_network_dataset_prepare.py

Script used to prepare dataset for neural_network.py, using structures generated via uca_matrix_files_synthese.py.

neural_network.py

Uses neural_network_dataset_prepare.py datasets as input values into neural network. It's core for ML solution to solve DOA-estimation problem.

benchmark.py

Used for benchmarking usability of NN (ML) in comparison with GCC-PHAT and MUSIC algorithm.

Summary

Readme currently would not be updated until project go into stable phase. Last Edit: 06.02.2022

About

Tool for synthezing audio files for multi-microphone UCA, neural network & benchmark - master thesis

Topics

Resources

Stars

Watchers

Forks

Packages

No packages published

Languages