[HOME] [EDUCATION] [PUBLICATIONS] [PATENTS] [RESEARCH] [SOFTWARE] [INTERNSHIPS] [COURSES] [PROJECTS] [PASTIMES] 

TIME DELAY ESTIMATION USING EXCITATION SOURCE INFORMATION IN SPEECH

Speaker Localization using excitation source information in speech Vikas C. Raykar, B.Yegnanarayana, S. R. Mahadeva Prasanna, and Ramani Duraiswami, IEEE Transactions on Speech and Audio Processing, Volume 13, Issue 5, Part 2, pp. 751-761, Sep. 2005. 

We propose a novel method to estimate the time-delay between the signals received by a pair of microphones in a noisy reverberant room, using the excitation source information in speech. The time-delay is computed by locating the peak in the cross-correlation of the Hilbert envelope of the Linear Prediction Residuals. Results show that our method gives better performance than the GCC-PHAT,  GCC-ML and Brandstein's pitch based methods.

[ Matlab code ] [ Source localization demo ] [ Demo Setup] [Face Detection Demo] [ Poster ]

Related  publications

Tracking a moving speaker using excitation source information Vikas C. Raykar, Ramani Duraiswami, B.Yegnanarayana, and S. R. Mahadeva Prasanna, In Proceedings of the 8th Eur. Conf. Speech Communication Technology  (Eurospeech 2003),  Geneva, September 2003,  pp. 69-72.