Watching the CEDIA video, I get the idea that Hubitat has a native app with the ability to turn a written phrase into a spoken announcement via a smart speaker.
Is this accurate (or did I misunderstand the video)?
If accurate, how, specifically, could an Ikea (Sonos type) speaker connect? I'd prefer the Ikea over the Sonos both because it is my understanding the Ikea don't have a microphone in them, only a speaker (please correct me if that is wrong) and they are less expensive. Do these connect to Hubitat over zigbee, or is a wifi connection required?
If wifi is required, can I block the speaker device from accessing the internet so that it communicates with the Hubitat entirely locally, and will that still allow the speaker to speak announcements?
Yes, this is possible. There is a built-in Sonos integration that works over your LAN: Sonos Integration | Hubitat Documentation. The Ikea Symfonisk speakers are Sonos devices and will work with this.
The text-to-speech (TTS) feature requires internet access on the hub to generate the audio file. However, playing it over the speaker is local. If the text does not change, you will also not need internet access after the first play, because it is cached on the hub. (But you'll continue to need it if the text can change, like reading a time or temperature.) You can also avoid the need for Internet entirely by playing a local file, whether that's TTS output you already have or just any compatible sound file.
Thank you. Terrific explanation and much appreciated.
I don't know if they have stickies here, but if they do, your response should be a stickie so it can also help others.
Again, thank you.
Sonos is what they used in that video. But if you have Alexa devices, there are a couple of different ways to have all your Alexa devices announce a phrase as well.
If the number of things you want Alexa to announce are static (don't include variables) and relatively few in number, you can use virtual switches with contact capability (community driver) to trigger Alexa routines. If you want to announce a variety of messages that are built at run time with variables and values, look into Echo Speaks.
@bertabcd1234 has described the built-in way to do (i.e. Hubitat uses Amazon Polly to generate a sound file that is played on a speaker locally). As he indicated, this method still requires the hub to have internet access for each new sound file that needs to be generated.
There is a community integration (linked below) put together by @jtmpush18, along with @erktrek, that is fully local. It does require a Raspberry Pi (or equivalent) that is connected to a speakers either directly or via Bluetooth. It can also use Sonos speakers that are on the LAN.
This approach does not require any internet access.
@aaiyar : Thank you for that clarification. I did not realize Hubitat goes through Amazon. I wonder if that transaction identifies the source of the voice file to Amazon? I prefer to not have anything to do with Amazon and their technology if I can avoid it. Need to figure out if I can learn RPI or learn how to set up another always on Linux device...avoiding all that was part of the appeal of Hubitat vs HA...
