DVSI offers AMBE+2™ vocoder libraries optimized for x86 processor platforms using Intel Single Instruction Multiple Data (SIMD) technology. Depending on the target platform these libraries are up to 90% faster than DVSI's standard x86 vocoder libraries. This addition to DVSI's high performance vocoder libraries allows x86 processors to run with fewer MIPS and gives customers the ability to increase overall system performance by reducing power requirements and/or increasing channel density.
With almost all modern x86 processors now incorporating SIMD instructions, users reap the benefits of hundreds of additional high performance CPU operations. DVSI’s optimized libraries for SSE and AVX allow designers to take advantage of these benefits and get a 40-90% faster increase in performance depending on the target platform. Both the SSE and AVX libraries incorporate SSE2, SSE3, SSSE3, SSE4.1 instructions and the AVX library additionally includes AVX and AVX2 instructions. The libraries are available for 32 and 64 bit Windows or Linux platforms.
DVSI's patented AMBE+2™ Voice Compression Algorithm can operate at virtually any data rate from 2.0 to 9.6 kbps. With this high degree of flexibility, the user can optimize speech and FEC rates to provide excellent voice quality with superior robustness to bit errors and acoustic background noise. The unparalleled performance of this vocoder technology has resulted in it being chosen as the standard for voice quality in Land Mobile Radio and Mobile Satellite Communications systems around the world.