Master Theses - Details

Every acoustic event is influenced by its physical environment, such as reflections and echoes from the surrounding geometry or the addition of  ambient noise. Every environment has its own acoustic signature. Identifying this signature using simple recordings (e.g., speech recordings) can be useful in many fields, such as forensics, integration into smart-home systems, and indoor navigation. Current approaches to room identification are either inaccurate or require large amounts of training data to function reliably. This is not practical in many real-world contexts. Current research topics include identifying acoustic environments using fewer examples or based on ambient noise.

The following topics are available for a thesis:

• Identification of acoustic environments based on metric learning.
• Creation of a dataset from noise measurements in an anechoic chamber and room impulse responses.
• Identification of acoustic environments based on ambient noise.

If you are interested, further details can be discussed in a personal meeting. The exact scope of the thesis will then be determined in close consultation with the candidate and in accordance with their interests.