[BA/MA] Spectral modeling synthesis for moving sound sources

[FanyuMeng]

Spectral modeling synthesis (SMS) is a technique applying deterministic and stochastic decomposition to obtain sound source signal. In the deterministic part tones are tracked by searching peaks, and subtraction between original signal and the peaks as the stochastic remaining part is synthesized by filtered white noise.

This project aims at developing SMS for moving sound source. As widely known there is frequency shift of moving sound source due to Doppler Effect. Instead of tracking the peak continuation in each STFT of windowed snapshot of the recording signal, the developed algorithm searches for the altering frequency trajectory corresponding to Doppler shift. The dependence on source speed is also important because the method in the thesis will be applied in high-speed moving vehicles. With the synthesized source signals, their auralization in virtual environment (CAVE) is possible.

https://gigamove.rz.rwth-aachen.de/d/id/di4pcgSRzsMejr
http://www.akustik.rwth-aachen.de/go/id ... abschl_22/

Supervisor
Fanyu Meng / Frank Wefers
fme@akustik.rwth-aachen.de

Institute of Technical Acoustics
Kopernikusstraße 5
Room 009