November 29, 2004
SPIRIT Corp introduced its new Multi-PASS embedded conferencing software designed for server-based and serverless peer-to-peer PC VoIP communication and conferencing.
SPIRIT Corp‘s Multi-PASS is a conferencing solution providing both server-based and serverless peer-to-peer and multipoint conferencing, based on SPIRIT’s Prosodic Active Speaker Selection (PASS) algorithm.
Prosody is the study of patterns of stress and intonation in language. By combining prosodic analysis with with dynamic utterance modeling, SPIRIT says it has made speaker activity estimation more accurate, enabling each speaker to get the best chance or time to speak, to be heard, and to hear all the meaningful speech of others.
The software enables any number of conferences and unlimited number of participants per conference, without degradation of speech quality and user experience, preserving clarity as in a face-to-face conversation, according to the company. The analysis is performed on each speaker’s side and does not load server CPU.
Controlled by the PASS software info, the conferencing engine optimally passes voice data from active speakers to the other participants. Broad bandwidth is not required, and due to smart traffic channeling, voice traffic and channel usage are optimized to reduce CPU load.
The company’s objective is to provide better-than-PSTN quality at low cost, thus enabling it to serve the exectutive conferencing market, as well as all other conferencing needs, all via VoIP.
SPIRIT communication software works within hardware solutions from clients including Agere, Atmel, Furuno, JRC, NEC, Nortel Networks, Panasonic, Philips, Samsung, Siemens, Tadiran, Texas Instruments, and Toshiba among 100+ other OEMs. Texas Instruments is shipping SPIRIT’s software ROM-ed into TI C54CST processor.