Speech Activity Detection on Multichannels of Meeting Recordings
The Purdue SAD system was originally designed to identify speech regions in multichannel meeting recordings with the goal of focusing transcription effort on regions containing speech. In the NIST RT-05S evaluation, this system was evaluated in the ihm condition of the speech activity detection task. The goal for this task condition is to separate the voice of the speaker on each channel from silence and crosstalk. Our system consists of several steps and does not require a training set. It starts with a simple silence detection algorithm that utilizes pitch and energy to roughly separate silence from speech and crosstalk. A global Bayesian Information Criterion (BIC) is integrated with a Viterbi segmentation algorithm that divides the concatenated stream of local speech and crosstalk into homogeneous portions, which allows an energy based clustering process to then separate local speech and crosstalk. The second step makes use of the obtained segment information to iteratively train a Gaussian mixture model for each speech activity category and decode the whole sequence over an ergodic network to refine the segmentation. The final step first uses a cross-correlation analysis to eliminate crosstalk, and then applies a batch of post-processing operations to adjust the segments to the evaluation scenario. In this paper, we describe our system and discuss various issues related to its evaluation.
KeywordsFalse Alarm Gaussian Mixture Model Speech Activity Speech Segment Speech Frame
Unable to display preview. Download preview PDF.
- 1.Garofolo, J.S., Laprun, C.D., Michel, M., Stanford, V.M., Tabassi, E.: The NIST meeting room pilot corpus. In: Proceedings of the 4th International Conference on Language Resources and Evaluation, Lisbon, Portugal (2004)Google Scholar
- 2.Janin, A., Baron, D., Edwards, J., Ellis, D., Gelbart, D., Morgan, N., Peskin, B., Phau, T., Shriberg, E., Stolcke, A., Wooters, C.: The ICSI meeting corpus. In: Proceedings of International Conference on Acoustics, Speech, and Signal Processing, vol. 1, pp. 364–367 (2003)Google Scholar
- 3.Burger, S., MacLaren, V., Yu, H.: The ISL meeting corpus: The impact of meeting type on speech style. In: Proceedings of International Conference on Spoken Language Processing, pp. 302–304 (2002)Google Scholar
- 4.Chen, L., Rose, R.T., Parrill, F., Han, X., Tu, J., Huang, Z., Harper, M., Quek, F., McNeill, D., Tuttle, R., Huang, T.: Vace multimodal meeting corpus. In: Proceedings of MLMI 2005 Workshop (2005)Google Scholar
- 5.Chen, L., Maia, E., Liu, Y., Harper, M.P.: Evaluating factors impacting forced alignment in a multimodal corpus. In: Proceedings of the 4th International Conference on Language Resources and Evaluation, Lisbon, Portugal (2004)Google Scholar
- 6.Liu, D., Kubala, F.: A cross-channel modeling approach for automatic segmentation of conversational telephone speech. In: Proceedings of IEEE ASRU Workshop, pp. 333–338 (2003)Google Scholar
- 7.Pfau, T., Ellis, D.P., Stolcke, A.: Multispeaker speech activity detection for the ICSI meeting recorder. In: Proceedings of IEEE ASRU Workshop, pp. 107–110 (2001)Google Scholar
- 8.Wrigley, S.N., Brown, G.J., Wan, V., Renals, S.: Speech and crosstalk detection in multichannel audio. IEEE Transactions on Speech and Audio Processing, 84–91 (2005)Google Scholar
- 9.Chen, S.S., Gopalakrishnan, P.: Speaker, environment and channel change detection and clustering via the Bayesian information criterion. Technical report, IBM (1998)Google Scholar
- 11.Fiscus, J.: Spring 2005 (RT-05S) Rich Transaction Meeting Recognition Evaluation Plan (2005), http://nist.gov/speech/tests/rt/rt2005/spring/