I currently have a leak sensor setup to send a notification to my Sonos speaker. The trouble is that when doing this it clears the current playlist which I don't like.
Is there a small zigbee/z-wave battery powered speaker I could use just for notications? I was hoping I could somehow use the speaker in my Nest smoke alarm but that's a dead end!
I'm using eufy speakers in two rooms. It requires a community driver called Media Render to work. $25 each, works as a DLNA speaker for TTS and plays mp3 and wav files. They also have a hardware microphone off switch. So the mic stays off through loss of power and reboots.
I really hope that Amazon adds some back end functionality that will allow more direct TTS integrations. The community solutions here are ingenious and the devs should get a ton of credit for their hard work - but it is work they shouldn't have had to do!
This is a integration "problem" that is actually caused by Sonos themselves. The current integration is using the undocumented but highly used uPNP integration for sending commands and getting status from Sonos players. There is a namespace called audioClip which is available to all current generation Sonos Players that does NOT interrupt currently playing items (playlists/favorits/streaming) and is the preferred method of sending short audio clips (TTS) to players. The audioClip function is available through the official Sonos API today. Not the bad part is the official API is cloud based but it is very fast. However because the namespace exists it wouldn't be too much trouble to do uPNP scanning and hacking to find the appropriate local call and then use it.
Recommendation is to petition Sonos to document and make the local API officially available for integration.
I use Google minis, got one free and one for 5 bucks. They work great but are not battery powered although you could use a battery pack to drive the minis
Any non-cloud speakers though? I too am looking for a completely local solution. I have Sonos, but am only using them for music. I use Samsung R1’s for TTS.
What I need is for one of the genius coders on here to make this PoE IP intercom work. $50 after shipping.
Yeah but you still need something to do the actual TTS synthesis. Most services/systems use Google or VoiceRSS or they rely on the MS Voices or Apple Voices for the speech synthesis.
Having a speaker you can send a mp3 file to or data stream to is only part of the equation.
Yes, DLNA is a communication protocol of sorts for AV equipment. A few of my TV's and my audio receiver are on my LAN and are picked up as DLNA devices by the Media Renderer app. I don't use them for notifications as they're typically either off or being used for other more important work. I bought a couple Eufy smart speakers (they're Echo knock-offs) turned off the mic and only use them as DLNA end points. I had to setup a server, I use an Intel NUC with Windows, a rPi would be cheaper, but I had the NUC and it wasn't doing anything. The server is used to host the sound files I play. From now through the Holidays it plays 3 seconds of sleigh bells when the front door opens. The rest of the year simple ding sound suffices.
What provides the voice synthesis. The conversion of the text to speech? Hubitat may have something built in like pico or using android voice synthesis locally then transferring the mp3 to the speaker. Or more likely because of resources using google translate or google TTS to perform the synthesis.
Now that you mention that, I believe you're right. I think I read somewhere that Hubitat uses Amazon for the actual TTS. So there is a cloud dependency for TTS. If you wanted to avoid that dependency you'd have to make your own mp3 and play those. But my point stands you don't need to manually add a service as it's already in there.