The 2004 BBN 1xRT recognition systems for English broadcast news and conversational telephone speech

This paper describes the BBN real-time recognition systems used in the 2004 Rich Transcription (RT) benchmark test for the English Conversational Telephone Speech (CTS) and Broadcast News (BN) tasks. We describe the system architecture, along with the algorithms we used in order to reduce computation with minimal impact on recognition accuracy. Particular choices in the design of the final system are analyzed to show the trade-offs between speed and accuracy. We also present recently developed new architecture for the real-time systems, which outperforms the systems we submitted for the RT04 benchmark tests for both domains.