Knowles, has launched a development kit for the Raspberry Pi for prototyping products based on a homegrown AISonic IA8201 DSP that was introduced not long ago.
Designed to bring voice, audio edge processing, and machine learning (ML) listening capabilities to various systems, the devkit supports wake-on-voice processing for low latency voice UI, noise reduction, context awareness, and accelerated machine learning inferencing. It also supports
“high bandwidth processing of advanced, always-on contextually aware ML use-cases and memory for multiple algorithms,” says the company,
and offers a choice of 2 mic array boards with 2x or 3x mics respectively. There’s also an adapter board with three buttons connecting to the Raspberry Pi, as well as the DSP board itself connected via a flat cable to a microphone array.
Software support for the IA8201 SDK includes Amazon Alexa, Sensory, Retune, and Alango voice services. It also supports TensorFlow Lite Micro SDK for AI/ML applications. Configuration tools, firmware, algorithms, and other product briefs can all be accessed via the company’s Solutions Portal.
The Knowles’ AISonic IA8201 OpenDSP is aimed at lower-power AI processing. It features a Cortex-M4 chip along with 2x Tensilica-based, audio-centric DSP cores: a DeltaMax” Xtensa LX5 128-bit 4-way floating-point SIMD DSP core optimized for frame-based audio processing and DNN acceleration, and, a “HemiDelta” Xtensa LX5 64-bit 2-way floating-point SMID DSP core optimized for low power, always-on processing of sensor inputs.
Other features and specifications of the AISonic IA8201 DSP include:
- 1.44MB RAM (1MB is made available to the user)
- Up to 4x PDM Digital Microphones – 1 stereo input, 4x mono inputs, and 1 stereo output (clock rates up to 6.144 MHz)
- Up to 3x I2S/TDM ports supporting 8x 32-bit audio channels
- SPI, I2C, UART, and GPIOs.
- Up to 175 MHz clock speed
- Voltage supply requirements
- IA8201BC – 1.8V Vdd
- IA8201CQ – 1.8V Vdd and 3.3V Vdd
- Temperature Range: -20°C ~ 85°C
- Packaging:
- eWLB 3mm x 2.6mm x 0.715mm, 0.4 pitch, 42 ball (IA8201BC)
- QFN 6.00mm x 6.00mm x 0.75mm, 0.5 pitch, 40 lead (IA8201CQ)
Usage for both the kit and the DSP includes smart home and building voice hubs applications, home entertainment systems, proximity detection systems, wireless earbuds, and inference engines for many industrial and commercial applications. You can also use it for security systems that detect glass breakage and for smoke alarms too.
The kit appears to be available for order but you may have to reach out to the company to get one as it is not being sold online. Other useful details on both the devkit and the IA8201 DSP can be found on the kit product page and datasheet respectively.