After over two years in the making, Pine64 has now launched the PineVoice (previously PineVox) smart speaker based on the Bouffalo Lab BL606P RISC-V wireless microcontroller with WiFi, Bluetooth, and Zigbee radio interfaces.
The device also features two microphones, a speaker, a USB 2.0 OTG port, volume buttons, a mute button with LED, a start/stop button, and four RGB LEDs. The PineVoice smart speaker is mainly designed to work with Home Assistant.
PineVoice specifications:
- SoC – Bouffalo Lab BL606P
- CPU
- D0 core – T-Head C906 64-bit RISC-V (RV64IMAFCV) CPU @ 480 MHz
- M0 core – T-Head E907 32-bit RISC-V (RV32IMAFCP) CPU @ 320 MHz
- Memory – 544KB RAM (788KB SRAM listed on the Pine64 store), 16MB embedded PSRAM
- Storage – 128KB ROM, 4KB eFuse
- Wireless
- 2.4 GHz 802.11 b/g/n 1T1R WiFi 4
- Bluetooth 5.x Dual-mode (Classis + BLE)
- 802.15.4 radio (Zigbee)
- CPU
- Storage – 128Mbit (16MB) SPI NOR flash
- Audio
- 2x digital microphones
- Speaker
- USB – USB 2.0 OTG port
- Misc
- Volume +/-, Mute, and Start/Stop buttons
- Mute LED, 4x RGB LEDs
- Power Supply – 5V, 2A via USB-C port
- Dimensions – 65 x 65 x 66 mm
The default firmware supports Wyoming Satellite protocol, which, according to the project itself, is now deprecated, and the ESPHome protocol is now preferred. The documentation includes more technical details with PDF schematics for the main and bottom boards, and software resources with links to various repositories for the SDK and firmware development. However, I can’t find the default firmware’s source code for now since there’s nothing specific to the PineVoice.
Pine64 explains that the PineVoice is still in early-stage development and may experience performance issues, notably for wake word detection. It should probably be considered as a smart speaker devkit, rather than a consumer device for your Home Assistant setup. This may change in the future if the firmware improves.
With that in mind, if you’d like to go ahead, the PineVoice can be purchased for $49.99 on the Pine64 store. It ships with a USB-A to USB-C cable and a Quick Start Guide. Warranty is valid for 30 days.
Via Liliputing
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. We also use affiliate links in articles to earn commissions if you make a purchase after clicking on those links.
Does it no have a dedicated DSP for advanced audio processing the voice captured from the microphone so that it can do echo cancellation (AEC) and noise suppression, etc. similar to the XMOS XU316 dedicated audio processor chip used in Home Assistant Voice Preview Edition hardware?
Far-field voice capture from across a noisy room is usually bad without some kind of advanced audio processing the voice captured from the microphone so that it can do echo cancellation (AEC) and noise suppression, etc.
No external DSP, but I can see an AEC circuit with passive components in the PDF schematics. The BL606P also features a built-in audio codec.
As Jean-Luc says, no specific DSP hardware.
But, the whole high-performance C906 core is dedicated for audio processing.
That’s also where local wake word detection is. There is plenty of computing power for any DSP audio processing, but currently there is only wake word detection.
Also, there are some T-Head instructions which should be suitable for DSP.
Wonder how this compares to the Home Assistant Voice – larger speaker but weaker audio processing?
I want to buy one
“802.15.4 radio (Zigbee)” so no Thread?
I am very sorry, I was busy so I am late with the documentation.
SDK was published as well, and you can find it there:
https://codeberg.org/pine64/pinevoice_smartspeaker_sdk
https://github.com/pine64/pinevoice_smartspeaker_sdk