The Non-Cloud and Cloud devices are deferent, but build in the same way.
But if you want the new voices, always use Cloud TTS.
As for your questions, yes, the idea is to create a TTS (Cloud or Non-Cloud) for each output device, that way you can configure volume and other settings specificly for each device.
What I want/need is to adjust the volume of one speaker for different flows.
Example:
â For the weather forecast announcement I need 15 %
â For the announcement of the radio station name I need 20 %.
â For the announcement in case of a burglary I need 75 %.
Indeed, i also have experimented A LOTT! during developement, but i also use this very much within my house, announcements are a real part of our house now-a-days.
And i use WAVENET with SSML, still, i only use about 1% or 2% of the 1 milion charâs a month.
Using SSML increases the usage/count a lott, since the complete text is counted as characters.
But with the experience in the last 6 months, i cannot imagine that anyone would reach the 1 million characters, and certainly not the 4 million non-wavenet voices.
You can search YouTube and it will trigger the Device trigger with the YouTube id and url of the first result.
But I will add functionality for playing lists and create an API so other apps and HomeyScript can consume it/search YouTube.
So, it will replace the functions of the current YouTube app that doesnât work anymore because of the Google restriction. The (free) api from YouTube is very limiting, thatâs why the app stopped working.
But since my Google Service App allows people to use there own google account/keys, it will keep well within the YouTube limits, because you only use it yourself.
Personally, I use it a lot through speech. when I say for instance: YouTube fireplace, it will turn on the TV with a (random) fireplace YouTube clip.
I can say anything much faster then typing it with my remote.
Regarding the possibility to adjust the volume unfortunately I have to relativize my statement a bit.
The problem I notice is this:
In order to be able to adjust the volume level up and down, the volume level of the trigger flow must be set relatively high.
My flow for this looks like that:
WhenâŠ
â URL triggered
ThenâŠ
â Play URL at volume 50%
If I want to play an announcement at about 15% volume (original Sonos speaker/app setting), I need to set the volume on the Google TTS âStart saying at volumeâ card to about -6.
But the problem I notice is, that the speech starts silent and then fades to the desired volume.
I add a pause of 50 ms, but didnât work expected.
The volume of one flow wasnât adjustable at all. And the announcement starts silent and than went louder. The same behavior as before.
So I rebuild the URL trigger flow, but without success.
Than I changed the mean volume from 50 to 30% and now it seems to work perfect again. Donât know whyâŠ
@Arie_J_Godschalk Hi Arie. Iâm eager to get this working as the old Google TTS voice really sounds like a moron. Now I think I did everything rightâŠbut still donât get it working.
I got until copy pasting the JSON (using notepad is that ok?) into the app settings via the developer page.
For some reason I donât get to select a default voice, the drop down stays empty.
Then of course when trying to run a flow according to your example it fails and asks whether I have setup the JSON and default voice.
Any tips on where to search what I did wrong?
Hmm. Apparently i had not enabled the TTS API, now I did and i can choose a default voice in google services via Homey developer.
But i still get the error message when trying to execute the flows. It still questions whether i have setup JSON and choose a voice