BPI-AI-Voice (Microsemi)

From Banana Pi Wiki
Revision as of 03:33, 29 May 2018 by Sinovoip (talk | contribs) (Hardware Spec)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search



Banana Pi BPI-AI-Voice(Microsemi) module for speech recognition,Now you can use it on Raspberry pi with amazon AVS, same as Microsemi AcuEdge™ Development Kit

BPI-AI-Voice (Microsemi) development kit

  • Tested and approved by Amazon for
— microphone solution for both 180o & 360o audio pick-up
  • Amazon ecosystem
—Reference code implements complete Echo functionality
—Provides developer access to Alexa Skills
  • Provides a fast and easy way method of evaluating Microsemi’s Automatic Speech Recognition Assist audio enhancement solution with Amazon’s AVS
  • Fast Prototyping
—Application development platform
—HW prototyping
  • Training Tool

Key Features

  • Supports 2 or 3 microphones linear or triangular array placement for enhanced audio pick up
  • Noise reduction
  • Smart automatic gain control (AGC)
  • High-quality stereo acoustic echo canceller for barge-in and full duplex audio
  • Fully configurable and scalable solution delivering a two microphone configuration for both 180° and 360° audio pick-up
  • Enhanced beam forming technology to enhance audio pick-up in adverse conditions, in presence of noise and outside sound sources
  • 2-way full duplex voice communication
  • Turnkey designs available as well as all appropriate drivers and documentation, along with worldwide support


Designed for world-class voice front-end applications, Microsemi’s new ZL38063 audio processor features the company’s AcuEdge™ technology. This innovative technology is a set of highly-complex and integrated algorithms that significantly improves automatic speech recognition (ASR) for both embedded and cloud-based ASR solutions. The ZL38063 delivers audio enhancements that perform noise reduction and smart automatic gain control (AGC), enabling speech recognition at distance in noisy, real-world conditions.

Microsemi AI 1.png

Application Examples

  • IoT Sensor
  • Smart Lighting
  • Smart Thermostat
  • Home Gateway/Controller
  • Speaker/Sound Bar
  • TV/Set-Top Box

Hardware Spec

To make a fully assembled smart speaker demo the following additional components are required:

  • Raspberry Pi 3 board or Banana pi board
  • 2A or greater Micro-USB power supply for the Raspberry Pi 3
  • SD memory card for Raspberry Pi/Banana pi
  • JBL Clip speaker
BPI-AI-Voice External Interfaces
Raspberry Pi 3/Banana Pi Header P2:•I²S port;•SPI;•8 GPIO
Audio Header JMMA1:•Digital microphones;•Analog out;•3 GPIO
SPI Flash Devices U2: Optional SPI flash component used to store ZL38063 firmware
USB J3: Optional USB power and debug port
Debug Headers JAIB2/2: Auto tuning headers
Digital Microphone Headers JM1–4: Optional header for off-board microphones
Audio Characteristics
Digital Microphones 4 on-board digital microphones (AKU441): Supports a 2 microphone configuration for 180° and 360° audio pick-up
Analog Output J1: 2 × 2.65 Low cost class D audio amplifier (NCP2820)
Stereo 3.5 mm male-to-male audio cable P1: Audio output
Micro-USB cable J3: Optional USB power and debug port

GPIO Pin define


Building Your Alexa AVS Smart Speaker Prototype

Microsemi's Amazon Alexa development kit is designed to help developers quickly and easily set up prototypes that demonstrate a high quality voice recognition interface. It comes complete with all the building blocks to help developers build their smart speaker prototypes.now ,we just design this AI board ,not test on banana pi SBC ,this project with Amazon and Microsemi just support raspberry pi 3.

Microsemi AI 2.png

Development Kit Technology Segmentation


Supporting the Amazon AVS Ecosystem

Developers can use Microsemi's Amazon Alexa development kit to build customer solutions for Internet of Things (IoT), smart home and industrial applications.

Microsemi AI 3.png

amazon AVS project