Amazon Echo to HA Controllers

[jdk972003]

That's what I am talking about...basically the HA software becomes the software bridge between the Echo and the devices.  vs. the layer being multiple apps all having to be configured, updated, monitored, etc....
 
I understand Amazon needs it to be multiple apps in order to drive the business model ($ potential = 3rd party developers) but Pete's HS way seems way easier for a larger system.
 
Your HA software does the work and you just monitor the single bridge driver.
 
Pete...could you use that to control Sonos through HS?

 

[jkmonroe]

Not Pete, and not using HomeSeer, but yes.  We have done something similar with CQC and can have it control a Sonos.

 

[pete_c]

Pete...could you use that to control Sonos through HS?

 
yes
 
Any defined device / variable can be controlled.  IE: In Homeseer 2 have close to 1000 variables defined for this or that.
 
Many folks utilizing Homeseer today are utilizing are also utilizing Sonos.
 
I am still adjusting to using the Amazon Echo as I am also using MS SAPI (many voice fonts).
 
I am very used to using SAPI and starting to play with the Kinect. 
 
Alexa is / was almost too easy to use compared to MS SAPI stuff. 

 

[mdesmarais]

How does the Echo do with distinguishing a voice if there is music playing that is NOT sourced from the Echo?

 

[pete_c]

How does the Echo do with distinguishing a voice if there is music playing that is NOT sourced from the Echo?

 
Don't know yet.  Primary audio distribution here is via Russound.  Music has been whatever.   I am very used to it.
 
I set up the Amazon Echo such that it is heard and not seen. I utilize the blue tooth microphone and flip back and forth between the two devices.
 
I am going really slow with it (must be my age?).   Over the years have turned down the chatty kathy of my automation TTS and went to using chimes. (1 if or 2 if or 3 if or 4 if - gets a bit convoluted though). 
 
I have stated here on the forum that I do not do cloud or wireless stuff and here I am playing with the Amazon Echo.

 

[wuench]

How does the Echo do with distinguishing a voice if there is music playing that is NOT sourced from the Echo?

Put it this way, I just asked echo to set an alarm standing about 5 feet away with the TV on kinda loud.  It heard me better than I heard it's response...

 

[ano]

Unfortunately I also find that the Echo also seems to understand a bit too good. I have had some instances where it started playing music or telling jokes because it THOUGHT we were speaking to it when we weren't.

 

[pete_c]

Yeah I have a 40 year old parrot that does the same except she doesn't play music.
 
Here have only had one spontaneous unexpected utterance from Alexa since turning her on.

 

[IVB]

My daughter had a friend over who alternated between Alexa and lexi. Her parents came over too. They like Alexa but she prefers Lexi. Her parents thought the echo was awesome.

As of Sunday, her name is now officially Lexi. And they're on the lookout for the prime day pricing.

 

[wkearney99]

The keyword recognition is pretty liberal in what it'll accept for the trigger.  But the use of a-lex-a is deliberate, as three syllables are going to be easier to distinguish as a trigger than just two.  The Echo does indeed do a pretty remarkable job of detection and subsequent transcribing.
 
If you're not sure why it got something wrong, use the Echo app on your phone or tablet and you can play back exactly what it recorded.
 
I've likewise only had it mis-trigger itself once.  Whereas my Xbox gets it wrong all the damned time, enough that I've disabled that feature.  Contrasted to a Windows 10 box with a Kinect setup and it's decidedly hit-or-miss whether it'll catch the "Hey, Cortana" trigger phrase.  Same thing for my wife's Moto X and the ridiculously clunky "Ok, Google Now".  Siri's pickup from an iPad 4 was never good for anything more than a foot or two away from it (and it was horrible at differentiating from background noise).  Amazon is significantly ahead of others in this ability.
 
It's a bit exciting to see the easy acceptance folks are having when trying the Echo.  Hopefully it'll bode well for better automation integration (across many vendors) in the near future.  Lots of things are finally coming together with a whole lot less tedium for setup and regular use.

 

[pete_c]

@Bill,
 
Does it learn from the cards (questions)?  I have been deleting mine when I go to the web page.

 

[ano]

My daughter had a friend over who alternated between Alexa and lexi. Her parents came over too. They like Alexa but she prefers Lexi. Her parents thought the echo was awesome.

As of Sunday, her name is now officially Lexi. And they're on the lookout for the prime day pricing.

You must have a newer model than mine. Mine can only recognize Amazon or Alexa.

 

[IVB]

You must have a newer model than mine. Mine can only recognize Amazon or Alexa.

 
the echo is amazon or alexa. We chose Alexa.
 
My daughters friend is alexa or Lexi.
 
But her friend & her folks came to my house this weekend and played with the Echo, the kid is now Lexi so they can dictate the tunes.

 

[wkearney99]

@Bill, Does it learn from the cards (questions)?  I have been deleting mine when I go to the web page.

 
I don't know.  I'm assuming if you point out an error about a specific event (using the app) that someone would have some way to review it from the audio.  I do know a 3rd party developer, however, cannot obtain the audio stream at all.  The apps get only what the voice recog system determines, not the raw audio stream.

 

[pete_c]

Thank you Bill.
 

I'm assuming if you point out an error about a specific event (using the app) that someone would have some way to review it from the audio.

 
I am only assuming what it is I see and hear. 
 
I do not know at this time if there is a common voice print database out there in internetlandia that is analysing and making comparisions.
 

The apps get only what the voice recog system determines, not the raw audio stream.

 
In order to create a base the voice recog system will need to determine what it's doing based on what it hears which does come from the raw audio stream.
 
Yeah here doing baby steps do review the audio and do comment whether it is a good interpretation or not.  
 
Listening to the audio; the hit rate is very good and my comments are typically acknowledgement of same said.