I was wondering if I can use tts to have announcements played in my google devices for various events, for example: "Water was detected in the bedroom floor".
But, I'd like it to be "smooth", it should at least:
Not interrupt whatever is playing (instead, pause and continue after announcement is complete).
Play some sort of sounds effect before the announcement (like in the airports) so your brain gets "triggered" and focus on it.
Executed locally so you don't miss it because internet is down.
Preferably I would like to use the same voice already set in the google device but if it is possible.
I was reading through the posts and unless I got this wrong, it sounds like sonos can do just what I want (announce without stopping the music) the problem is... I already have google hubs and minis through the whole house, even inside the bathrooms.
There is a built-in app called “ChromeCast Integration (beta)” that allows you to play TTS on Google devices. I use it regularly.
That said, I haven’t tested it while playing music, so not sure if it pause and continue.
It does sometimes start with a sound effect, but that doesn’t seem to be the case every time. You could always store a .mp3 file on your hub and have it play that before the announcement.
It will use the TTS voice you have setup in Hubitat.
When you first speak through a speaker it opens a cast connection on the speaker and that's what the sound is from. If you speak again while that connection is open it won't make another sound. There is an option called "stop after tts is complete", if you have that on it should close the connection automatically so that the sound will play again the next time.
I performed some tests and no, it won't resume playing anything.
I looked around it it appears that it can only be done using the "broadcast command" which is not supported by this app. Looks like people are doing it by hosting a separated server (relay) which is more than I can handle at the moment.
In my app - I call it "GroovyMachine" - I use TTS a lot for announcements.
But...
It's not possible to broadcast to the speakers, so you have to interrupt whatever is playing...
What I do: For normal announcements I let only the idle speakers speak; But for urgent announcements I set the volume on 100 and use all speakers.
You can play any sound you like with the help of playTrack.
I've never tried to send TTS while internet is down... But as every sentence is "translated" to an mp3 by an Amazon service I doubt that it will work. (Maybe when a sentence is still in the cache?)
BTW: You choose the TTS voice in the Hubitat settings (IMHO it uses Amazon's voices), so it's not 100% the same as in Google Home.
There is a “Play” command. If you are using Rule Machine, you should be able to send the command after your announcement. You may want to do that only if it was playing however... There might be a way to check... not sure though...
UPDATE: Scratch that. I tried it and the play command doesn’t seem to work. It doesn’t resume playing what was playing before.
What I "heard" on another thread is that first time it goes to an Amazon web service to generate the audio file. Next time it is processing the same sentence it will use the local file instead.
So in theory, if you generate all the files while you have internet connection then... you would be able to go offline. But the question is: who delete these files and when?
Oh man... I don't think there is way to accomplish this without the assistant relay.
I wanted it to be able to work 100% offline but since tts also uses the cloud I think I will go with virtual switches and let google handle the alert using routines.
I'm hopping the routines will be able to pause and continue whatever is playing. I will have to test it to confirm.
If you want to be 100 % off line you may want to look at Ikea SYMFONISK
WiFi bookshelf speaker, black
$99.00
(Sorry, couldn't link) The TTS to the speaker is local. I don't have one but several people here seem to like them.
- I use TTS a lot for announcements.
