Master Thesis - Modeling and identification of non-linearities in distorted digital speaker systems using deep learning

Job Title

Master Thesis - Modeling and identification of non-linearities in distorted digital speaker systems using deep learning

Job Description

Category

Audio, FW, SW, Analytics

Scope

2 students; master thesis 30 hp

Background

Modern digital speaker systems have a complex signal path consisting of multiple audio processing or transducing components, such as digital signal processors (DSP), digital to analog converters (DAC), audio amplifiers and speaker drivers. Each of these steps may introduce undesired non-linearities into the audio signal which are manifested as a distortion between the original audio reference signal and the output audio signal.

These kinds of distortions have an especially detrimental effect in acoustic echo cancellation (AEC) reducing the quality of two-way communication devices. Their presence translates into spurious sounds (artifacts) that cannot be canceled and are perceived as annoying residual echo. A big improvement can be achieved when the reference signal to the AEC corresponds as closely as possible to the actual output signal -- with all the non-linearities included.

For this reason, there is interest to model the total signal path of a digital speaker system, so that different parameters defining the non-linear behavior can be identified and as a result more accurate reference audio signal can be determined. Scientific literature reports several different approaches to such modeling, and the recent evolution of machine learning algorithms (e.g., deep learning, neural networks) has demonstrated unprecedented potential to model these phenomena with high precision.

The scope of this thesis project includes exploration of these possibilities, and the goal is to reach high precision modeling of non-linearities and investigation of the applicability of using these results to enhance AEC and potentially audio quality.

Goal

The main goals of the project are:

· Study how non-linearity of distorted digital speaker system can be modelled and verify applicability of different models

· Evaluate automated identification algorithms for these models

· Use real digital speaker system as reference and capture the signals of interest

· Benchmark the models using the aforementioned measurements.

Who are you?

You have an interest in audio technology, engineering, machine learning, signal processing and/or embedded firmware development, and you do a Master Program in a related field, such as Computer Science, Electronics Engineering, Acoustics Engineering, Mathematics, or Physical Sciences.

The thesis project should be carried out by two students, together, but we accept single applicants also. Please, indicate your possible co-applicant in your application.

OK, I am interested! What do I do now?

You are valuable to us – how nice that you are interested in one of our proposals! There are a few things for you to keep in mind when applying.

· Applications are accepted in both Swedish and English, and you apply via the proposal advert.

· The announced thesis is open only to students affiliated with a Swedish University/College either directly or via an exchange program.

· When the thesis proposal states that it includes two students working together, we would like you to apply in pairs. In these cases, send one application each but make sure to clearly state in your application who your co-applicant is. If you have any questions regarding this, please do not hesitate to contact us.

· Please attach your CV, cover letter and University/College grade summary.

Who to contact for any questions regarding the position!

Duja El-Khamisi duja.elkhamisi@axis.com

Type of Employment

Temporary Employment (Fixed Term)

Posting End Date

2024-12-11

Certain roles at Axis require background checks, which means applicable verifications will be done in these recruitments. Notice will be provided before we take any action.

About Axis Communications

We enable a smarter, safer world by creating innovative solutions for improving security and business performance. As a network technology company and industry leader, we offer solutions in video surveillance, access control, intercom, and audio systems, enhanced by intelligent analytics applications.

With around 4500 committed employees in over 50 countries, we collaborate with partners worldwide. Together, we thrive in our friendly, open, and collaborative culture and inspire each other to think beyond the expected. United by our commitment to inclusion, diversity, and sustainability, we consistently seek to develop our skills and way of working.

Let´s create a smarter, safer world

For more information about Axis, please visit our website www.axis.com.

Listen to Get To Know Axis – Podcast