Voice UI capabilities are becoming more and more important to OEMs. In the home, numerous appliances from smart speakers to set top boxes to microwave ovens all have built-in voice assistant capabilities. The trend is even continuing to the automobile with new natural language processing (NLP) and voice assistant services being adopted by leading brands.
Voice UI processing typically falls into two categories: audio pre-processing and speech recognition/keyword detection.
- Five very long instruction word (VLIW)-slot architecture capable of issuing two 128-bit loads per cycle
- 2X MAC capability versus HiFi 4 DSP for pre- and post-processing, includes:
- Support for eight 32x32-bit or 16 16x16-bit MACs per cycle
- Optional eight single-precision floating-point MACs per cycle
- 4X MAC capability versus HiFi 4 DSP for NN processing, includes:
- 32 16x8 or 16x4 MACs per cycle
- Optional 16 half-precision floating-point MACs per cycle
- New HiFi NN library offers a highly optimized set of library functions commonly used in NN processing (especially speech), which can easily be integrated into popular machine learning frameworks
- Software compatibility with the complete HiFi product line, totaling over 300 HiFi-optimized audio and voice codecs and audio enhancement software packages