Speech recognition

I’ll get straight to the point. I have a nuerological disease similar to MS. One thing it does is mess with my speech. Long story short my speech always sounds like I’ve had about 20 beers. I also have an amazon echo dot and an iPhone. Alexa understands me when I’m 3 feet away looking right at her. Siri understands me 9/10 times but I suspect that is because I have the phone right up at my mouth. One of my problems is the trigger word. Alexa is 3 syllables so I changed to echo (2). Now my problem seems to be she starts responding before I am done talking. There isn’t much documentation on voice components for HA. Can anyone please point me towards any that may help? Perhaps diagflow? Much appreciated.

I had my echo dot set to the wakeword echo for ages, only to find that it picks up Alexa a thousand times easier.

Unfortunately that doesn’t help you one bit.

The only thing that springs to mind for me at the moment is Snips. I haven’t read up on it much yet but because it runs on your device (rather than some server a million miles away) I would guess you can tinker with the settings to make it a bit more patient and customize it around your voice patterns.

It might not, but certainly worth a read until anyone else comes along with a better researched suggestion…

Thanks for the quick response. However I have hassio so sudo is pretty much void for me as I understand.

Snips is also available as addon.
The official does not work, because the image will give you a “unsupported bus type” error.
Official is under build in addons in Hass IO

I have created a fix, you can add my repo: https://github.com/Romkabouter/hassio-addons
Be sure to add the Snips addon, not the split-up services :wink:

1 Like