Domoticz

I am moving all my stuff from a mac server to a pi. But i am missing 2 mayor components. Speech to text (with a normal voice not a computer stephen hawkings one) and speech recognition, which works kinda nice on the mac.

I found

A real-time full-duplex speech recognition server, based on the Kaldi toolkit and the GStreamer framework and implemented in Python.
https://github.com/alumae/kaldi-gstreamer-server

Due to limited bandwidth and privacy issues (not comfortable with cloud based stuff) I want it local.

So I was wondering what do you use for speech recognition or Text To Speech (tts)?

I have 3 Amazon Echo's around my apartment, I use echobridge to integrate them with Domoticz.

Example usage: Alexa turn on one, or Alexa turn on two.

It can also be Alexa turn on bed lamp.

I just find calling my devices numbers is more reliable for voice recognition.

https://github.com/armzilla/amazon-echo-ha-bridge

I want something that can listen to commands as osx dictation does (offline)
Not full blown siri stuff.

I want to say: "Computer boil kettle"

Like I can now...

NO WAY i'm sending all my conversation to a stalker company.

/beginrant

If these companies (facebook, amazon, microsoft, google, ..) were a person, would you socially interact with them knowing they tell anyone with money what you conversed with them?

-No, you would avoid such person as being non trust worthy

And getting stalked in almost every place you go...?

-No, you would call the cops.





My policy is every smart home product that uses a internet (cloud based) solution and has to connect to the internet (and has no local api), gets returned.
No use with my flacky internet connection...

"Ow I'm sorry sir there is no internet, we can't verify that you are you, cuz we can't login, so no, I can not turn the light on,..."

fcuk that. also that you can not use it anymore in x years when the manufacturer (or their server) is gone... SO not going to buy something expensive that is "lock-in" with a vendor, not having access to my own data, and I having to agree its selling my private data to any third party with money... and who stalks me on other websites to get to know my preferences... NO WAY...

/endrant

Echo's don't send any audio to the cloud unless you say the activation word and if an echo hears that word the top lights up bright blue very, very visibly. Not to mention there is a Mic mute button on top of you really want. I don't have any privacy concerns myself, obviously I'd prefer they work locally but the Echo voice recognition technology is just so brilliant I am prepared to use the cloud to have such fantastic long-range voice recognition functionality.

Range is not an issue, internet connection is, i pay for every byte send/received, and its unreliable it can be down for hours.
(yeah and the privicy stuff... nevermind that)

The problem is, on my mac, i just download all the (high quality) voices and language packs... and it works... no fancy stuff, but it speaks
belgium-dutch nicely, with lack of emotion yeah, but still... I can speak in dutch to it and it runs, scriptable, easy...

Now I go to the pi, which is linux based, and there is either cloud based speech recognition or tts, or really low quality local tts like espeak.. or very complex speech recognition software with no easy setup guide throwing terms I do not know... and do not get.

So that I was wondering... because if there is no offline alternatives on linux for tts having some nice voices (english is ok) and no offline
speech recognition (english is ok) which is not that hardcore to setup (mean no source editing, or manual creating training sets)... then i will
move all my stuff back to my mac...

which has the problem that I run out of usb ports and i could not find "ANY" working USB3 hub since el capitan...
Total crap... file server with external hdds running on usb2... cuz no usb3 hub works... well i found one, which would NOT power up after power failure, which happens here quite a lot...and have to press a button, then you have a file server with no files to server... apple

perhaps you can use your mac as a slave for your RP? [ if Domoticz is running on your mac? [ i hate Apples ] [ Pi3? or buy are Cubieboard!!] ]
Send the voice command with the slave to the master ..??

What is the name of this mac product?

I am looking for a Ducth tts for a long time...
When i can let my wife speak to Domoticz in Dutch...
Whe go horse riding

is called OS X you get it if you buy a mac. (or install a hackintosh)


on 4:48, male voice
https://www.youtube.com/watch?v=ueqXo1plz2U

but i rather like the female voice, I got the belgium one to answer the front door.... (well answer, pff ...
I also wrote a test, that when you press the frontdoor bel, you get a whole conversation between the male voice and the dutch female voice...), I just airplay it to the little speaker (with buzz, and static) attacked to a pizero at the frontdoor...

For saying commands,
The internal mic of my mac reaches about 4 meters to say commands...

So not know about accents... cuz i speak with and without and I tried it only without.

is it @ this moment working with Domoticz?

I keep summiting apost and then it summits my previous post... 3 time today....

Yeah it is working. You can look at the wiki page of os x

https://www.domoticz.com/wiki/Mac_OS_X
https://www.domoticz.com/wiki/Mac_OS_X# ... ith_Z-Wave

I probably can add a section how to interface voice commands with domoticz using applescript...

and with the new macOS Sierra (osx name is changed to macOS) you probably can interface domoticz with siri using applescript..

I give up, searching the web resulting in nothing usefull so far unless a lot of work, without knowing if it gonna work...
not sure what to do with the pi3 now...