Introduction
Photo: Arrangement for making a two-channel recording of two simultaneous speakers. These recordings were used in blind speech separation (BSS) research funded by the Air Force Research Laboratories (AFRL), Human Effectiveness (HE) directorate.
The Advanced Speech and Audio Processing (ASAP) Laboratory at New Mexico State University conducts research in the following areas: speaker recognition, speech enhancement, and speech coding. The speech lab is located in Goddard Annex Room 132 and contains state-of-the-art recording equipment, test and measurement equipment, digital audio workstations, and embedded DSP hardware and software development tools.
Blind Source (Speech) Separation Corpus
The Blind Source Separation (BSS) corpus is a collection of digital recordings of mixtures of speech signals and background noise in various acoustic environments. These real-world recordings can be used to evaluate various BSS and other speech enhancement algorithms using a standard test set. In addition, impulse responses for various acoustic environments are available for use in controlled simulation of different “mixing environments.”
The recordings were made under a grant from the Air Force Research Laboratories and are freely available for download.
[ Anechoic Chamber | Construction Site | Lecture Hall | Stairwell | Study Lounge | University Ave | Acoustic Impulse Responses ]