Homey Community Forum

Google Services - App for Translate and new Text-to-Speech for any device with url input

I’m tracking this for some time already, but can someone tell me, what are the advantages of this approach compared to TTS ? Is it faster and more natural, or ? Thank you :wink:

Hey RWin,

Thanks, i’ve added the defaults.
Also included all your other advises, thanks!

Submitting it to Athom now.

Hey Sharkys,

Well, to list the options you have:

  • Pick any voice you want, there are a few hundreds, multiple per language (dutch have a total of 10 you can pick from)
    These voices are much better then the old google translate voices.
  • You can choose WaveNet voice, kinda like A.I. voices, which are way more natural then the old (or new/normal) voices.
  • You can set Pitch (highness of the tone)
  • You can set Speed in percentage, not just slow/normal
  • You can set delay accourding to your device to use conditioncards and wait for speech to finish before the next actioncard.
  • You can add pauses to the end or beginning (handy for Chromecast users)
  • You can use SSML (see link at the topicstart) - This is also a real game changer!
  • You can choose a device profile, which is an audio effect which will be applied.
    Set this to large speaker or heaphone or google mini
  • You can set a volume gain.
1 Like

Thank you, what about the reaction itself, is it faster or slower comparable to TSS ?

Don’t know, haven’t really compaired it. My old TTS was just about as fast as the new TTS, but if this App’s TTS is faster then the one you used in another app, i don’t know, depends on the app.
The URL is recieved pretty fast, and because the lat (most used) 100 speech texts are store in memory, repeated texts are almost instantly.

I have never used the Google TTS app because i don’t have any google devices, just Sonos, Smart TV’s, Phone’s/tables (with own app) throughout the house and my smartwatch (which still talks to me in the old Goolge TTS dutch female horrible voice , havent had time to update it lol)

1 Like

No, you don’t need another device: You get an unique URL for each Unique speech (including text, pitch, speed, pauses, etc). And those URLs are SHARED between TTS devices if there the exact same speech with the exact same settings.
So creating a device for each text would be pointless.

You create a device for each output-device you have, each speaker you want to speak alone oon, and/or create a device for which the trigger will send the url to all speakers.

Because for each TTS device you link it to a speaker/device through a trigger and a action card sending the url.

Ofc you could use it differently when using H.O.O.P. (Hope) - Homey Object Oriented Programming

1 Like

I’m using a Sonos Beam for my primary speech output. With the latest Sonos app (6.0.9) I don’t have the ‘play audioclip’ card anymore. Now I’m experiencing the same issues as @fantross ran into before he started using the play audioclip card; the MP3 keeps repeating and it takes a few seconds before it starts talking. It seems to pause the current item on the Sonos (be it watching TV or listening to radio / Spotify) and switches to the MP3 file. It doesn’t revert back. I’m using the 'play ’ card.

I tried updating my Sonos Say app (was using 2.0.1 with my own http-api docker image), but after installing 3.1.5 it doesn’t find my speakers and can’t switch back to my own http-api.

Any thoughts?

Are you on Dutch? Because, yes you should have them, but they gave it Dutch names, in dutch it is now “Play (url) at volume” image

This is Athom’s “copy” of the Sonos Say Announcement/send URL cards, it goes fast and does not repeat.
Unfortuantly it doesn’t (yet) restore groupings etc after playing.

1 Like

Weird that Sonos Say doesn’t find your speakers, does it list All speakers as an/the only option?
Or nothing at all?

Yes, the All speakers is an option, haven’t tried it yet, it also gives a timeout.

Going to homeyip:5005 results in an error:

{"status":"error","error":"No system has yet been discovered. Please see https://github.com/jishi/node-sonos-http-api/issues/77 if it doesn't resolve itself in a few seconds."}

I guess it has to do with my Sonos being on a different SSID than my Homey. Although it should relay multicast / broadcast and stuff

Did you have your docker in the same network (SSID) as the Sonos or Homey?

I use Asus AiMesh throughout the house, so my speakers are always on the same SSID, but it shouldn’t matter as long as there on the same network/subnet.

It’s all on the same subnet and vlan. Docker is a wired connection. Created a additional 2g only SSID for most IoT devices, but that’s all.

I’ve got the following cards:

image

It seems, compared to your card, I should be using the second. I was using the first one which repeats.

Perhaps restart the App, and what a minute before opening any flows that show the Speaker list.

I have had the issue during developement that, when i opent a list with speakers in a flow to fast after restart, it would generate the exact same messages and NOT resolve afterwarts. I needed to restart the App and wait.

Indeed, and thanks for the image, i guess Athom also renamed the English version to Play () at volume instead of Play Audio clip.

Restarted the app, restarted homey… nothing seems to matter.

Great, that worked. So always be using the one with ‘volume’. Great, good to know. And I agree, it’s really fast.

Any updates from Athom on the grouping / degrouping issues? I’m currently not using it, but just ordered a new ARC and FIVE and would like to use these grouped.

Seems like the Athom Sonos App currently allready regroups very well, if been testing it, and so far its good. It keeps my grouping.

1 Like

6.0.9? There is no stock for the ARC so have to wait.

Yeah, 6.0.9.
So far, my groups stay correct after announcements :slight_smile:

1 Like