Tweet
Do you think it is possible for Loxberry to implement such a feature with a plugin ? So we can transfer the voice (text) command to miniserver via mqtt ?
-
-
...at least the application "ffmpeg", which comes pre-installed with the Loxberry, already allows text-to-speech via its "libflite" library. I have used this to stream text to Unifi cams, see here. But I have not yet worked with speech-to-text, aka speech recognition. But ffmpeg also seems to support this, see here. So using the two filters ASR (automatic speech recognition) and subsequently extracting the generated data via ametadata (Manipulate frame metadata) should get you further? Everything combined, likely, with the AnyPlugin...
Alternatively, maybe using Whisper AI? it may run on a Raspberry Pi 4... not sure if it would run on a Raspi 3?Zuletzt geändert von MarkusCosi; In den letzten 3 Wochen. -
Well, it's not only the software part, but also the hardware part which could be really challenging.
I use Alexa with Homebridge and Loxone and I am really happy with that. If you wait for an offer, an echo costs 25 Euro. But of course Amazon is listen to you....
If you do not want to have a cloud solution, I would go for HomeAssistent and include that into Loxone. They have a promising new voice solution without cloud and with good hardware.
🇺🇦 Hilfe für die Menschen der Ukraine: https://www.loxforum.com/forum/proje...Cr-die-ukraine
LoxBerry - Beyond the Limits
Kommentar
Kommentar