I use the Sonos plugin TTS to announce when doors and windows are opened. Recently only certain doors and windows will announce when triggered. I send a phrase using the Say command like “Someone Opened the Front Door” which it says but if I change the phrase to “Someone Opened the Back Door” it won’t say anything. Nothing has changed on my Vera Plus or the scenes I use. Anyone has an idea what could have changed. I’m using Sonos ver 1.4, using 1.5 I lost all TTS.
I am just guessing here but what you are describing is a good indicator that responsive voice has stopped working for some time and that the announcements you are getting are the ones which were cached. (they were obtained from the responsive voice api a while ago and stored on your local machine and called back when needed)
But caching wasn’t a feature until 1.5…
Edit: And it looks like the day has come, folks… the URL that we’ve been using for RV is now returning “Forbidden”. Get ready to switch over to Microsoft Azure (still in development)
1 Like
</start of PSA>
Just as the infamous “Cousin Oliver” doomed the Brady Bunch TV show, it seems whenever I explore a new Vera plug-in feature – such as Sonos TTS and the ResponsiveVoice API – it either breaks, becomes payware, gets deprecated, or suddenly orphaned.
I blame my being such a latecomer to advanced Vera usage and the somewhat steep (for me) learning curves that accompany high-tech projects like Reactor, Lua, and Text-to-Speech. (Or am I really cursed, like poor Cousin Oliver?)
Happily, @rigpapa has once again stepped up to the plate to help us all avoid obsolescence, with his willingness to curate and steadily improve the SONOS 1.x (and soon-to-be 2.x) app on the community’s behalf. I look forward to hearing mechanized voices from my speakers one day!
Speaking of “the plate”, may I heartily encourage users of Patrick’s exemplary plug-ins to visit his Donations Portal to toss him a few shekels for his tireless efforts?? It’s only fair, but moreover helps guarantee he’ll be able to find the time to keep coding!
</end of PSA>
1 Like
Y’all are too kind!
RIP Responsive Voice. It looks like Microsoft Azure, supported only by Sonos 2.0 in development, is going to be the new recommended engine.
Sonos 2.0 is mostly functional. I’m just working on the last of GUI/JavaScript updates to the UI7 APIs (it was still using the old pre-UI7 calls). So anyone brave enough to install the Github “develop” branch can get access to the Microsoft Azure TTS engine immediately. It’s actually much cleaner than any of the prior engines, IMO. Really nice clear speech, and lots of voices and languages.
BUT eyes open if you go this route: version 2.0 is also a conversion of the plugin from per-device to parent-child. This makes it much more efficient in memory/resources. It is supposed to adopt your current Sonos devices in situ, so the Vera device numbers will not change, thus preserving scenes, Lua, Reactor configs, PLEG, etc. “Supposed to.” I’ve tested it many times on my own system, and it works, but it’s one data point, on my system my way, and I suppose any possible oddity out there is possible that could trip it up. Once done, restore from backup is the only undo. Backups are your friend. Also, this 2.0 development code base has never run on anything less than 7.30, so that may also be a consideration. Finally, I have not exhaustively tested every possible action across the long list of services that these things support–literally dozens of actions. I don’t think most people use a fraction of the actions supported, and most of the important stuff is exercised by the UI (which works), but still, if you’re a Sonos+Vera power user/scripter you may run into some hiccups (which of course, I will troubleshoot and fix ASAP).
Just a brainfart suggestion…
I used Azure just before my last switch and it was working quite well for a long time. I think it is a good choice but their API at one point changed from bing to azure and required for me to adjust the old plugin from lolodomo. As I got tired of all this cloud dependance, I decided to kill it and move to a local solution, exploring maryTTS and some others and finally settled on… MacOS. So I am presently no longer relying on any plugin but on pure lua code in a scene calling a sonos-http-API Server running on a macOS desktop. No more cloud… Maybe you could integrate such an option in your plugin?
I’m committed to remaining on 7.29 for another few days (waiting out the initial wave of 7.30/7.31 upgrade lemmings to leap…)…
…so I might be brave and pull a “Leroy Jenkins” with Sonos 2.0 dev tomorrow, just to see what happens. If I do, will certainly report back … if I live, that is!
Well, I’m glad that I have an answer to my question. How do I go about loading what I guess is beta Sonos 2.0 from Github? Do I have to do something with Microsoft Azure before I download the Sonos plugin.
@rigpapa, are you sure that Responsive Voice is dead ? I did some googling and can’t find any confirmation. Also on their own website no sign that they changed things or it is not free anymore ?
I don notice although it does not work anymore on the Sonos plugin
My sense is that the free subscription is going away, moving to paid. I only signed up yesterday, and that’s the feeling I got from RV’s site, not that they were going away or stopping the service itself.
The URL we’ve been using now returns “Forbidden”.
As it was, their service changed some time ago. It used to be that they published an API for converting text to a speech audio file. Now they produce a runtime-loadable API for speech on web sites, but the conversion API documentation all disappeared. The old API still worked for a long time. Now it does not.
1 Like