4 ms·
Has anyone tried to build an open-source voice assistant using the $100 Respeaker far-field mike array dev board, https://respeaker.io https://respeaker.io & ht
by walterbell 8y ago
Has anyone tried to build an open-source voice assistant using the $100 Respeaker far-field mike array dev board, https://respeaker.io https://respeaker.io & http://wiki.seeedstudio.com/ReSpeaker_Core_v2.0/ http://wiki.seeedstudio.com/ReSpeaker_Core_v2.0/?
> ReSpeaker Core v2.0 is designed for voice interface applications. It is based on the Rockchip RK3229, a quad-core ARM Cortex A7, running up to 1.5GHz, with 1GB RAM. The board features a six microphone array with speech algorithms including DoA (Direction of Arrival), BF (Beam-Forming), AEC (Acoustic Echo Cancellation), etc. ReSpeaker Core v2.0 runs a GNU/Linux operating system ... Our BF and AEC algorithms are provided by Alango, a professional audio DSP company used in the automotive industry.
Voice assistants are magical when they do-what-you-mean. If there was an active OSS community for voice assistants, it would enable experimentation with privacy-preserving architecture and customization for niche use cases and language.
- j88439h84 8y ago- https://github.com/NaomiProject/Naomi https://github.com/NaomiProject/Naomi - http://mycroft.ai http://mycroft.ai - http://snips.ai http://snips.ai
- walterbell 8y agoThat reminds me that Google has DIY kits at Target, but they are designed as Google service extensions. https://aiyprojects.withgoogle.com https://aiyprojects.withgoogle.com
- utopcell 8y agoHardware seems to be by far the easiest component of the equation. Google home minis are practically free.
- RodericDay 8y agothat's because you're paying to install spyware in your house
- remir 8y agoMozilla was doing some work with speech recognition. Don't know if that project is still active. But yeah, these things needs lots of data to work well. That would be a massive project.
- StudentStuff 8y agoMozilla DeepSpeech is pretty active considering the holidays: https://github.com/mozilla/DeepSpeech/pulse https://github.com/mozilla/DeepSpeech/pulse Looks like they'll be having their v0.4 release sometime this month?