RTF Estimation Using Riemannian Geometry for Speech Enhancement in the Presence of Interferences

Yuval Sitton: LinkedIn, GitHub • Or Ronai: LinkedIn, GitHub
Amitay Bar: LinkedIn, GitHub • Ronen Talmon: LinkedIn, GitHub

Background

In this project, we leverage the Riemannian geometry of correlation matrices to address the problem of signal enhancement. In particular, we focus on signal enhancement in the presence of interference sources, which remains a challenging problem. We develop and analyze new algorithms based on the Riemannian approach rather than the commonly-used Euclidean one.

Files In The Repository

File name	Purpsoe
EstimateSignalRiemannian.m	Code runs all the simulations and plots the graphs
functions/*.m	Functions used in the simulations for code clarity
simulations/*.m	The simulations presented in the paper
plots/*.m	Plot functions for simulations
sounds/examples/	Speech examples comparing the approaches
rir_generator_install.sh	Installing the RIR generator

• We note that the speech examples were tested for the scenario examined in the paper.

Installation

1. Clone the repository.
2. Install the RIR generator using the rir_generator_install.sh script.
3. Alternatively, install it using the instructions appear in AudioLabs - RIR Generator.
4. Install PESQ Matlab Driver.
5. Download the TIMIT Acoustic-Phonetic Continuous Speech Corpus.

Prerequisites

The code was tested using Matlab 2022a on Windows 10.

Run The Simulations

To run the simulations and create the graphs that appear in the paper, run the EstiamteSignalRiemannian.m file.

• We note that the simulations of the graphs not presented in the paper are also included here.

Sources

• TIMIT Acoustic-Phonetic Continuous Speech Corpus.
• Image method for efficiently simulating small-room acoustics.
• PESQ Matlab Driver.
• STOI Implementation.