Nicla Voice is the latest board from the Arduino PRO family with support for always-on speech recognition thanks to the Syntiant NDP120 “Neural Decision Processor” with a neural network accelerator, a HiFi 3 audio DSP, and a Cortex-M0+ microcontroller core, and the board also includes a Nordic Semi nRF52832 MCU for Bluetooth LE connectivity.
Arduino previously launched the Nicla Sense with Bosch SensorTech’s motion and environmental sensors, followed by the Nicla Vision for machine vision applications, and now the company is adding audio and voice support for TinyML and IoT applications with the Nicla Voice.
Nicla Voice specifications:
- Microprocessor – Syntiant NDP120 Neural Decision Processor (NDP) with one Syntiant Core 2 ultra-low-power deep neural network inference engine, 1x HiFi 3 Audio DSP, 1x Arm Cortex-M0 core up to 48 MHz, 48KB SRAM
- Wireless MCU – Nordic Semiconductor nRF52832 Arm Cortex-M4 microcontroller @ 64 MHz with 512KB Flash, 64KB RAM, Bluetooth 4.2/5.0 LE connectivity. Note: via U-blox ANNA-B112 module just like in the Nicla Sense board
- Storage – 16MB SPI flash
- USB – Micro USB port for power, programming, and debugging
- Sensors
- IM69D130 microphone
- 6-Axis IMU (BMI270)
- 3-axis magnetometer (BMM150)
- External microphone connector (ZIF)
- I/O expansion
- 1x I2C ESLOV connector
- Headers with castellated pins with 1x serial port, 1x SPI, 2x ADC, programmable I/O voltage from 1.8-3.3V
- Power Supply
- 5V via micro USB port
- Battery connector for 3.7V LiPo battery with integrated battery charger and fuel gauge (BQ25120AYFPR)
- Dimensions – 22.86 x 22.86 mm
- Weight – 2 grams
- Temperature Range – 0°C to +70°C
The Nicla Voice is compatible with Arduino Portenta and MKR boards and can be programmed with the Arduino IDE, Arduino CLI, or Arduino Web Editor. You’ll find instructions to get started and more technical details on the documentation page.
The board can be used for work safety and preventive maintenance with the board listening to unusual sounds in combination with its motion sensors. With support for multiple wake words and the ability to recognize voice commands, the Nicla Voice can be leveraged in harsh environments where workers need to wear protective gear and cannot easily interact with keyboards or switches. Arduino also expects the smart audio board to be used for building automation, notably in smart doorbells, to enable faster identification of critical and time-sensitive events such as glass-breaking or intruders detection
The Syntiant NDP120 is the main highlight of this board and it is designed for low-power TinyML audio applications and battery-powered devices:
The Syntiant NDP120 is a special purpose processor for deep learning and is ideal for always-on applications in battery-powered devices. The NDP120 applies neural processing to run multiple applications simultaneously with minimal power consumption.
Built using the Syntiant Core 2 programmable deep learning architecture, NDP120 is designed to natively run multiple Deep Neural Networks (DNN) on a variety of architectures, such as CNNs, RNNs and fully connected networks.
The Nicla Voice is not available yet, but Arduino will sell it for $82 US or 69 Euros on the company’s store where you can register your interest to be informed once the board becomes available. A few more details may also be available on the product page on the Arduino Pro website.
Via Hackster.io
Jean-Luc started CNX Software in 2010 as a part-time endeavor, before quitting his job as a software engineering manager, and starting to write daily news, and reviews full time later in 2011.
Support CNX Software! Donate via cryptocurrencies, become a Patron on Patreon, or purchase goods on Amazon or Aliexpress
Its sort of oh great and then you get to
$82 US or 69 Euros on the company’s storeand maybe not.“The board can be used for work safety and preventive maintenance with the board listening to unusual sounds in combination with its motion sensors … The smart audio board to be used for building automation … to enable faster identification of critical and time-sensitive events such as glass-breaking or intruders detection.”
Fantastic.
We can now have systems that recognize footsteps from afar and alert us of approaching intruders, thus eliminating the need for tall fences. Systems that can alert us to noises from malfunctioning machinery in factories, ships, trains or airplanes. Systems that mind the tell-tale noises from the electric grid, or mind enemy troop movements in the battlefield. Systems that can compliment building maintenance and security — or listen for specific noises — even from the weather — no matter how overwhelming the surroundings sounds might be.
A $25 eufy cam can do the same thing.
Camera picture not as useful if deaf/hard of hearing. aka speach / multipl speakers or direction/type of sound(s).
Ya…for $82 they are smoking to much weed.
and there is little to no support from the factory. Wasted weeks to run their simple audio model car demo that was missing far too much documentation. Received compliments in correcting their buggy tutorials. So far, have not seen them deliver on what they promise. It appears that every nice to have feature is a paid feature and you have to have a ‘business model’ to receive their attention. Many other options out of China for < $ 2 USD (single piece qty) for similar NN devices.