These methods basically can give higher time-frequency resolution than the traditional FFT-based method, a more accurate pitch estimate, and have shown to be beneficial for speech enhancement, speech recognition, speaker recognition, and monaural speech separation.
The recently released REverberant Voice Enhancement and Recognition Benchmark (REVERB) challenge includes a reverberant automatic speech recognition (ASR) task.
The REverberant Voice Enhancement and Recognition Benchmark (REVERB) challenge is an Audio and Acoustic Signal Processing (AASP) challenge sponsored by the IEEE Signal Processing Society in 2013, and has recently been released for studying reverberant speech enhancement and recognition techniques [1].
We are confident that the challenge's data and achievements will fuel future research on reverberant speech enhancement and recognition.
Voice activity detection (VAD) is an essential module in almost every audio signal processing application, including coding, enhancement, and recognition.
We used the training dataset provided by the "REVERB challenge" (reverberant voice enhancement and recognition benchmark) [40].
The presence of pigment information is verified by our previous experiments [32 34] where information fusion of VL and NIR images led to a significant enhancement of recognition performance.
Table 1 summarizes information about which task(s) each participant addressed as well as the characteristics of the enhancement and recognition system(s) proposed in each submission.
This paper outlined the achievements of the REVERB challenge, a community-wide campaign that evaluated speech enhancement and recognition technologies in reverberant environments.
In order to provide a common evaluation framework for developing and testing of algorithms in the fields of dereverberation as well as reverberation-robust ASR, the REverberant Voice Enhancement and Recognition Benchmark (REVERB) challenge [8] has been launched and REVERB contributions showed significant improvements for speech enhancement (cf., e.g., [9]) and ASR (cf., e.g., [10, 11]).
