Digital Assistants
Cirrus Logic offers a comprehensive portfolio of ultra low power smart codecs, high-performance SoundClear® algorithms to enhance voice interaction with digital assistants. Multi-mic voice capture suppresses noise and reverberation that degrade digital assistants accuracy and high performance AECs enable the user to reliably barge-in during loud playback.
Highlighted components are offered by Cirrus Logic.
System Features

Multi-Mic Voice Capture
Noise reduction and echo cancellation regardless of the surrounding environment.

Barge In Capabilities
Hands-free control even when your device is playing multi-channel audio.

Hi-Fi Playback
Integrated multichannel 24-bit hi-fi audio hub enable our smart codecs to deliver high quality audio for any application.
Recommended
Voice Processor
-
CS48LV41F
Multi-Mic, Far-Field Voice Processor Featuring Voice Control and Communications up to 6+ Meters
SoundClear compatible
Smart Codecs
-
CS47L15
SmartHIFI Smart Codec With Advanced Audio Features for Mid-Tier Smartphones
SoundClear compatible -
CS47L24
Ultra Compact, Low Power Smart Codec with Dual Core DSP
SoundClear compatible -
CS47L35
Smart Codec with Low Power Audio DSP
SoundClear compatible -
CS47L90
Low Power, Highly Integrated Audio Codec Featuring SmartHIFI for Mobile Playback
SoundClear compatible -
WM8281
Low Power Smart Codec with Active Noise Cancellation and Echo Cancellation
SoundClear compatible
Portable Codecs
-
CS42L51
Low Power Stereo Codec with Headphone Amplifier
Request to Buy Dev Kit
Cirrus Logic CRD1569-1 Voice Capture Development Kit for Amazon AVS.
Digi-Key (Part #598-2471-KIT-ND) »
Highlighted components are offered by Cirrus Logic.
System Features

Two-Mic Voice Capture Reference Board
Features the Cirrus Logic CS47L24 dual core, low power smart codec and two CS7250B digital MEMS microphones, plus SoundClear® algorithms for voice capture, noise suppression, and echo cancellation.

Raspberry Pi 3 Board
Runs Cirrus Logic Control Console with AVS client, Sensory’s TrulyHandsfree™ wake word engine tuned to “Alexa”, SoundClear® tuning and diagnostics. Provides LAN, WLAN and hotspot connectivity for headless operation.

Control Console
Intuitive GUI replaces the need for Raspbian line commands saving hours of setup and start up time. Enables both headless (no display, keyboard or mouse), Wi-Fi and hotspot operation to emulate real product connectivity. Enables remote control, diagnostics and tuning through any internet browser.

Travel Case and Demo/Eval Platform
Durable case with custom anti-static foam inserts to hold and protect all kit components while traveling or in storage. Lid folds flat for use as eval/demo platform holding all components in place and avoiding speaker vibration.

International Power Supply
Full 2.5A @5V to power entire system plus RPi3 peripherals including international travel adapter set. Alternatively, 5V rechargeable battery supply (not included) can be used to emulate battery powered products.

8GB SD Card
High performance Micro SD Card preloaded with OS and all required firmware.

Compact, Passive Speaker Box
Enables full kit functionality without additional components. Simulates the audio experience of AVS enabled products with small integrated speakers. Can be replaced by actual transducer selected for product. Alternatively, just plug any amplified speaker into voice capture boards stereo jack.

Ribbon Cable
Allows voice capture and RPi3 boards to be separated instead of stacked when required to fit in true form factor prototype and demonstration platforms.
Recommended
Smart Codecs
-
CS47L24
Ultra Compact, Low Power Smart Codec with Dual Core DSP
SoundClear compatible
Technical Documentation
-
Voice Capture Dev Kit for AVS User Guide
v1.1 : 792 Kb -
Voice Capture Dev Kit for AVS Quick Start Guide
v1.2 : 923 Kb -
Voice Capture Dev Kit for AVS Product Bulletin
DS1189 : 489 Kb -
CS47L24 Product Data Sheet
Rev 4.1 : 3.4 Mb -
Voice Capture Dev Kit for AVS Schematics
v2.00 : 579 Kb
Highlighted components are offered by Cirrus Logic.
System Features
- Increases accuracy and reliability of “Alexa” wake-word triggering and command recognition, at 6+ meters far-field distances and in noisy environments
- Supports 6+ meters far-field operation with 2, 3, or 4 microphones
- FlexArrayTM Technology supports arbitrary mic array geometries, including circular, square, triangular,linear or 3-D
- AEC is scalable to mono, stereo, 3-channel, and 5-channel (surround movie or music) use cases
- Includes local high accuracy large model Amazon voice trigger or Sensory voice trigger
- Supports far-field voice call mode and in-call voice triggering
- Suitable for wall-powered devices or portable battery-operated products with low mW consumption
- 6 mics built-in, ready to be configured for testing circular, square, triangular, and line arrays in 4-mic, 3-mic or 2-mic designs
- AVS compliant LED indicators and mic-mute button
- Connectivity includes Wi-Fi 802.11, Bluetooth 4.1, stereo line out, USB device port, as well as a USB host port for connection to external USB soundcard
- Also includes expansion bus for connecting GPIOs, ASPs, SPI, I2C, and PDM mics
- Web-based firmware control console for simplified interactive operation, mic array selection, acoustic tuning and diagnostics

Recommended
Voice Processor
-
CS48LV41F
Multi-Mic, Far-Field Voice Processor Featuring Voice Control and Communications up to 6+ Meters
SoundClear compatible
Portable Codecs
-
WM8998
24-Bit Hi-Fi Audio Hub with 3 A/D Converters, 7 D/A Converters and up to 6 analog or 3 digital microphone inputs
Audio Amplifiers
-
CS35L41
11 V Boosted Mono Class D Amplifier with DSP Speaker Protection and Audio Enhancement
Technical Documentation
-
EveryWord Far-Field AVS Development Kit Technology Bulletin
Rev 1.0 : 779 Kb