Introduction:

Usually it is assumed that sound decays exponentially but as previous research has shown, most of the time the decay is multiple sloped (see fig., energy decay from the Boston Symphony Hall with a double slope [1]). In those cases applying linear regression to calculate room acoustic paramters like EDT or T20 is questionable. A framework has been developed in a previous Master Thesis [2] which allows for calculating decay distributions with a Bayesian framework. Now the work should be extended analyzing decays from famous concert halls and compare them to conventional room acoustis parameters. The question should be answered, if the new decay times can us tell additional information about the quality of a concert hall. Furthermore a famous acoustician (Mr. Toyota) asked the question, if curved energy decays are good or bad. If possible, this questions should be addressed, too.

Tasks:

• Calculate energy decay curves and decay distribtuions
• Improve MEDD algorithm (Maximum Entropy Decay Distritbution)
• Compare results with conventional room acoustic parameters like EDT, T20, T30 etc.
• Answer question: Are curved energy decays good or bad?

More Information:

• Timeframe: October 2018- July 2019
• For more Information please contact Jamilla Balint, balint@tugraz.at

[1] Manfred Schroeder, A new method for measuring reverberation time, Bell Labs 1964

[2] Florian Muralter, Analysis tool for multiexponential decay curves in room acoustics, Master Thesis at TU Graz, 2018