-->

SPIRIT Announces New Vocoders

MOSCOW, RUSSIA - SPIRIT, the DSP software developer for TI platform in continental Europe, offers new 1000 bps and 2000 bps proprietary vocoders, closely related to SPIRIT's 1200 bps and 2400 bps vocoder algorithms. Featuring reduced bitrate and good speech quality, these vocoders are best fitted for bitrate-critical applications with limited storage capacity, such as digital answering machines, PDAs, home appliances and more. Both vocoders are based on the classic LPC model, with Mixed Multi-Band Excitation (MMBE) model utilized for excitation signal that ensures correct transmission of all speech signals with tone, noise and mixed spectrum components. The key factors that make 1000/2000 bps vocoder ideal for using in bitrate-critical applications with limited storage capacity are low bitrate, high level of speech quality, noise robustness and automatic built-in frame synchronization of input bit stream. Both vocoders are compliant with TI's eXpressDSP standard. SPIRIT's 1000 bps and 2000 bps vocoders are available for licensing in the form of portable ANSI C code. "In Q1 '03, these vocoder algorithms will be ported to TI's TMS320C54x and TMS320C55x platforms" stated SPIRIT's CEO, Andrew Sviridenko. "We believe that their high level of speech quality combined with simple application interface will make 1000/2000 bps vocoders the best solution for any parties working with TI C54x and C55x platforms". 2400 bps vocoder for iPaq SPIRIT's proprietary 2400 bps vocoder is now available for Compaq iPaq personal digital assistant equipped with Intel's StrongArm processor. 2400 bps vocoder for iPaq is targeted at application integrators who want to incorporate voice communication capabilities of this device for real-time voice communication over IP networks. This also provides value to the end users who will use their handheld organizers as web-phones or digital recorders. "2400 bps vocoder can be used in a vast range of application, such as telephony, digital answering machines, Internet multimedia, call-logging systems, and more. The algorithm is hand-optimized for lowest MIPS and memory consumption, and thoroughly tested for bit-exactness and high robustness to acoustic noises and errors" said SPIRIT's CEO, Andrew Sviridenko.
SpeechTek Covers
Free
for qualified subscribers
Subscribe Now Current Issue Past Issues