Hello HA Friends! FutureProofHomes.net team here - and we build super powerful voice assistant hardware & software for Home Assistant.
We’re excited to announce that pre-orders for our new Satellite1 Voice Assistant & Multi-Sensor are now live! Shipping will have devices arriving at your doorstep this April (next month) for US customers
Also, starting in June, we’re expanding our store globally as we’re currently in the process of getting our UL, CE, and RoHS certifications to ensure the highest quality and safety standards worldwide.
It’s been a ton of work to get here, so huge thank you for everyone supporting our work btw — big things are ahead!
I’ve watched the demo video and this looks really nice.
I’ll ask the same question here as I did with the HA VPE…
does this support initiating an interaction from the smart assistant? Or are all interactions only initiated by the user?
I occasionally use my Echo devices to initiate an interaction by it asking me if I want to do an action or have HA perform an action automatically.
for example the biggest one I use is to have the echo ask if I want to turn off the lights if there hasn’t been any motion in the room for a while. Then I can either respond no to keep the lights on, yes to turn them off or no response in which case it turns the lights off automatically since it assumes I’m not there to hear it.
and I assume you can use it as a TTS speaker for announcements?
I definitely like your design better that the HA VPE and especially the additional sensors and expandability.
Thanks @finity! Yea, the Sat1 supports all the possible interactions the Nabu’s voice assistant supports. So, yes, it can give TTS announcements, and initiate the interaction, or ask a question to the room, etc. All things are possible!
I have heard quite some people stating that the HA VPE is not yet very good at picking up voice. If this is indeed due to the selected hardware, is this improved on Satellite1, that would make it possibly a preferred pick over VPE.
As soon as I can source the speakers (all of the recommended ones are on backorder) I’ll let you know. I have a devkit ready to be assembled and plan to put it where one of my existing vpes sit.
Very exciting to see indeed. For me personally, the biggest challenge to overcome is indeed reliability of wake-word detection in real-life scenarios. So I am really interested to see how this fares in real-life scenarios.
I got excited to see this has a 4 microphone HAT as that could potentially mean it has better pick-up I guess? Though I see in the documentation (Introduction - FutureProofHomes Docs) that at the moment the XMOS FW only supports two microphones, and 4-microphone support would come later via a FW update? So would that mean that, at least for now, we could expect similar wake-word performance as the Voice PE?
Anyway, nice to see more interesting HW options popping up.
I wonder if the ability of Google / Amazon voice units’ ability to pickup wake-words and other text so well is indeed the hardware. Perhaps it’s more in the quality of their software processing?
I will not claim to be an expert, but I suspect you are right. I do think they can do better beam forming with more microphones. So it is likely a combination of both. I assume they also do more than the standard pre-trained models to isolate your specific voice.
Also, starting in June, we’re expanding our store globally as we’re currently in the process of getting our UL, CE, and RoHS certifications to ensure the highest quality and safety standards worldwide.
