The WhisperTrigger voice detection system is offered with digital or analogue microphone options, designated WhisperTrigger D and WhisperTrigger A, respectively.
The WhisperTriggers, designed as hard silicon IP, are suited for triggering voice recognition algorithms thanks to their low detection latency, enabling faster wake-up of the voice processor. A Noise Detected as Speech Rate (NDSR) of 8% of power ratio threshold level at 10dB, helps keep power consumption low by reducing the number of unnecessary in-error SoC powerups.
WhisperTrigger offers a saving of up to 30% in terms of DSP activity, due to its efficient detection algorithm. Its power consumption is a typical 15μW, added to codec stand-by mode consumption for voice detection. Detection latency is as low as 8ms depending on the type of phoneme.