What is VoiceSpot?

VoiceSpot is the most accurate, smallest memory and MIPS profile wake word engine available on the market today. Customer defined voice trigger words or phrases are available.


Why did Retune develop VoiceSpot?

Engineers designing Voice Control solutions need a reliable, cost-effective wake word solution, and that’s exactly what VoiceSpot is. They also need great beamforming and acoustic echo cancellation to deliver a complete solution, and that’s where VoiceSeeker comes in. Together, these two solutions enable our customers to move quickly through the software integration process with a single software supplier to get to a winning, more cost effective solution.


How Well Does It Work?

VoiceSpot enables the identification of a trigger word or phrase in challenging and noisy environments, with single microphone performance at levels comparable to the 7-microphone Echo Dot @ 2m. With a low false reject rate, low false accept rate, and robust to interfering noise, it’s easy to love VoiceSpot.

Features:

  • Very small memory footprint

  • Low computational complexity allows ALWAYS ON functionality

  • Very low latency, up to 100mS faster response time than competing solutions

  • Custom Voice Triggers for target words / phrases and/or demographics are supported

  • Multiple Trigger instances are supported

  • Performs well in noise and reverberation, even with a single microphone

  • Low False Reject Rate (FRR)

  • Low False Accept Rate (FAR)

Available on these processors:

  • Qualcomm: QCC5121, QCC5125, 68100

  • Dialog Semiconductor: DA14195, DA14196

  • Analog Devices: ADADN8080

  • ARM: Cortex A-Series

For further information on VoiceSpot, please reach out to us at info@retune-dsp.com.