Title :
Room geometry estimation from a single channel acoustic impulse response
Author :
Moore, Alastair H. ; Brookes, Mike ; Naylor, Patrick A.
Author_Institution :
Dept. of Electr. & Electron. Eng., Centre for Law Enforcement Audio Res., London, UK
Abstract :
For a 2D rectangular room of unknown dimensions and with unknown source and microphone positions, the times of arrival of reflections can be described in terms of image source positions. Adopting a microphone-centred coordinates system, it is shown that to satisfy certain combinations of arrival times imposes constraints on the possible room geometry: a second-order reflection from adjacent walls determines the source-microphone distance; a second-order reflection from opposite walls in a given dimension determines the source displacement in that dimension as a function of the source-receiver distance. Given a subset of time differences of arrival, the extent to which the geometry can be determined is related to these constraints. The geometry estimation is further posed as a least squares optimisation problem whose results verify the analytical results.
Keywords :
acoustic signal processing; least squares approximations; microphones; transient response; channel acoustic impulse response; image source positions; least squares optimisation; microphone positions; microphone-centred coordinates system; room geometry estimation; source-microphone distance; source-receiver distance; Acoustic measurements; Acoustics; Geometry; Microphones; Optimization; Speech; Vectors; TOA; acoustic impulse response; geometry estimation; room identification; time of arrival;
Conference_Titel :
Signal Processing Conference (EUSIPCO), 2013 Proceedings of the 21st European
Conference_Location :
Marrakech