Facebook Pixel

Design of Voice Activity Detection Algorithms - Digital Signal Processing (DSP) case

Retune DSP

Company description

Retune DSP (www.retune-dsp.com) is a team of highly specialized sound and signal processing specialists. We develop and license best-in-class multi-microphone audio and voice enhancement technologies for mobile devices and consumer electronics. With some of the most advanced sound minds in the industry, we deliver:

•Voice quality and noise reduction technologies

•Smart audio solutions

•Product development and support on embedded platforms.

The company is headquartered in Kgs. Lyngby, Denmark with sales and support offices in United States and China

Project description

For students interested in working with Digital Signal Processing (DSP) and/or Machine Learning (ML), we are offering Master Thesis opportunities. Depending on your expertise and interests, we can host projects in design, test and integration of advanced DSP algorithms for noise reduction, beamforming, echo cancellation, and other audio enhancement solutions.

Possible project example:

We would like to improve our Voice Activity Detection (VAD) algorithms. In the initial step of building personal assistant type (Alexa, Google, Siri, etc.) hands-free voice command products, it is essential to monitor the audio on a continuous basis and accurately detect human voice with very low energy consumption. The project will therefore investigate a number of VAD algorithms and characterize their performance and complexity, incl. algorithms involving neural networks with supervised training.

Some additional project ideas could be:

• Implement and optimize advanced DSP/ML algorithms on embedded platforms
• Small-vocabulary speech recognition suitable for embedded devices

Student description

Interest in and experience with signal processing or machine learning, embedded systems, algorithms and software development. We also expect you to have programing skills in Python/Matlab and C/C++.

We expect you to work with us at our headquarters in Kgs. Lyngby, Denmark.

Also, Retune DSP offers an opportunity to work as a student assistant (paid) or internship position before and after the thesis project. After successful collaboration, there is a possibility to have a long-term contract at Retune DSP.

Design of Voice Activity Detection Algorithms - Digital Signal Processing (DSP) case  | Match My Thesis
Nov 30, 2018


Voice Activity Detection


Digital Signal Processing


Machine Learning