Audio, Voice, and Speech

• The Converged Audio Voice Speech (cAVS) subsystem consists of a collection of controller, DSP, memory, and link interfaces that provides the audio experience to the platform. This subsystem provides streaming of audio from the host SW to external audio codecs, with the host CPU and/or DSP providing the audio enrichment.
• The optional DSP can be enabled in the audio subsystem to provide low latency HW/FW acceleration for common audio and voice functions such as audio encode/decode, acoustic echo cancellation, noise cancellation, etc
• The cAVS is fully backward compatible with the Intel HD Audio specification, with the controller implements a number of Output Stream DMA engines and Input Stream DMA engines for data transfers, as well as a Command Output DMA engine and a Response Input DMA engine for control transfers.
• The cAVS also supports I2S audio codecs which are not Intel HD Audio standards. The General Purpose DMA engines has the ability to do simple data transfers or control transfers between system memory and the FIFO in the DSP I/O peripheral interfaces directly, however, these transfers are not optimized for power management.