Acoustic Parameter Extraction from Occupied Rooms Utilizing Blind Source Separation
Room acoustic parameters such as reverberation time (RT) can be extracted from passively received speech signals by some ‘blind’ methods, which mitigates the need for good controlled excitation signals or prior information of the room geometry. However, noise will degrade such methods greatly. In this paper a new framework is proposed to extend these methods for room parameter extraction from noise-free cases to more realistic noise environment, such as occupied rooms, where noises are generated by occupants. In this proposed framework, blind source separation (BSS) is combined with an adaptive noise canceller (ANC) to remove the noise from the passively received reverberant speech signal. Room acoustic parameters can then be extracted from the output of the ANC with existing ‘blind’ methods. As a demonstration we will utilize this framework combined with a maximum-likelihood (ML) based method to estimate the RT of a simulated occupied room. Simulation results show that the proposed framework provides a good estimate of the RT in such a simulated occupied room.
KeywordsSpeech Signal Blind Source Separation Reverberation Time Adaptive Noise Canceller Blind Source Separation Algorithm
Unable to display preview. Download preview PDF.
- 1.Kuttruff, H.: Room Acoustics, 4th edn., pp. 315–333. Spon Press (2000)Google Scholar
- 3.ISO 3382: Acoustics-measurement of the reverberation time of rooms with reference to other acoustical parameters. International Organization for Standardization (1965)Google Scholar
- 4.Cox, T.J., Li, F., Darlington, P.: Extracting room reverberation time from speech using artificial neural networks. J. Audio. Eng. Soc. 49, 219–230 (2001)Google Scholar
- 7.Greenberg, J.E.: Modified LMS algorithm for speech processing with an adaptive noise canceller. IEEE Trans. Signal Processing 6, 338–351 (1998)Google Scholar
- 8.Ratnam, R., Jones, D.L., O’Brien Jr., W.D.: Fast algorithms for blind estimation of reverberation time. J. Acoust. Soc. Amer. 11, 537–540 (2004)Google Scholar