Experimental results for baseline speech recognition performance using input acquired from a linear microphone array

Abstract

In this paper, baseline speech recognition performance is determined both for a single remote microphone and for a signal derived from a delay-and-sum beamformer using an eight-microphone linear array. An HMM-based, connected-speech, 38-word vocabulary (alphabet, digits, 'space', 'period'), talker-independent speech recognition system is used for testing performance. Normal performance, with no language model, i.e., raw word-level performance, is currently about 81% for a set of talkers not in the training set and about 91% for training set data. The system has been trained and tested using a close-talking bead-mounted microphone. Since a meaningful comparison requires using the same speech, the existing speech database was appropriately pre-filtered, played out through a transducer (speaker) in the room environment, picked-up by the microphone array, and re-stored as a digital file. The resulting file was post-processed and used as input to the recognizer; the recognition performance indicates the effect of the input device. The baseline experiment showed that both a single remote microphone and the beamformed signal reduced performance by 12% in a room with no other talkers. For the array tested, the error is generally attributable to reverberation off the floor and ceiling. / / ! / J J / -J 7 J J t / / / s' / j J" J / J

The Huge Microphone Array (HMA)

“…For audio and videoconferencing applications, reverberations and competing sources cause errors in determining source locations as well as poor audio quality. For speech-recognition applications, these phenomena adversely a ect the performance 10,11,12,13,7]. The received signal may degrade over time and change with any movement of a talker since a modi cation of position or orientation alters the propagation paths to the acoustic sensors.…”

“…Flanagan and coworkers have studied the use of microphone-array technology at AT&T Bell Laboratories (and now at the Center for Computer Aids for Industrial Productivity (CAIP), Rutgers University) for more than ten years 1, 2, 3, 4]. Since 1987, research into microphone-array technology particularly for data acquisition for speech-recognition has been underway in the Laboratory for Engineering Man/Machine Systems (LEMS) at Brown University 5,6,7]. Acoustic measurements have been accumulated, new algorithms formulated, and several all-digital research systems have been built over the years.…”