Paper:
http://web.media.mit.edu/~mehoque/Publications/13.Hoque-etal...I found the choices of software particularly interesting:
"Prosody Analysis: ... To extract prosodic features, we developed an application
programming interface (API) using the low level signal processing algorithms included in the Praat [praat.org], an open source speech processing toolkit."
"Speech Recognition: For real time speech recognition, we used the Nuance speech recognition software development kit (SDK). ..."
