Or go back to Vera Alerts for Android (Or you can probably do something from Tasker) for free TTS on the Android device that works without a Internet connection.
Thanks to this, I’ve installed MaryTTS (http://mary.dfki.de/) on my Raspberry PI (that was only running the Hue Server for the Amazon ECHO Workaround). This is for those that don’t have a 24/7 Mac server running. Hopefully it helps 
Here’s what I did (Full Steps).
- SSH into Raspberry PI
$ cd Downloads
$ wget https://github.com/marytts/marytts/releases/download/v5.1/marytts-5.1.zip
Unzip the application to the /usr/bin directory
$ sudo unzip marytts-5.1.zip -d /opt
At this point it will not run unless the you have already installed Java 1.7 you can determine the current version of Java by executing:
$ java -version
Install the required version of Java (also add openjdk-7-jdk if you intend to do any java development):
$ sudo apt-get install openjdk-7-jre
After installing the new java runtime (jre) it will still not be the default. To set the new jre to your default use:
$ sudo update-alternatives --config java
Selection Path Priority Status
- 0 /usr/lib/jvm/java-6-openjdk-amd64/jre/bin/java 1061 auto mode
1 /usr/lib/jvm/java-6-openjdk-amd64/jre/bin/java 1061 manual mode
2 /usr/lib/jvm/java-7-openjdk-amd64/jre/bin/java 1051 manual mode
Press enter to keep the current choice[*], or type selection number: 2
Having selected option 2 the java version should return something similar to:
$ java -version
java version “1.7.0_65”
OpenJDK Runtime Environment (IcedTea 2.5.1) (7u65-2.5.1-5~deb7u1)
OpenJDK 64-Bit Server VM (build 24.65-b04, mixed mode)
The runtime package delivers the scripts necessary to run the MaryTTS Server, which can be used via a browser of the client to synthesize speech. The server can be launched with:
$ /opt/marytts-5.1/bin/marytts-server.sh
This can then be used either through a browser or via the MaryTTS Client. The browser address will be:
http://(RaspPI_IP):59125
The MaryTTS Client, which is a Java GUI can be launched with:
$ /opt/marytts-5.1/bin/marytts-client.server
Launch the Server
- Edit L_SonosTTS.LUA (Download Apps → Develop Apps → Luup files)
- edit the SAY_EXECUTE variable, replace the 5th line with the Google URL to:
http://(RaspPI_IP):59125/process?INPUT_TYPE=TEXT&OUTPUT_TYPE=AUDIO&FAKE=%s&INPUT_TEXT=%s&LOCALE=en_US&AUDIO=WAVE_FILE
- Please replace the (RaspPI_IP) with your device IP and make sure to put them inside the quotes.
- Upload, replace, restart LUUP.
- Test and enjoy! Congrats, you’re done. No more relying on other TTS servers.
[quote=“lolodomo, post:20, topic:188206”]I read about Microsoft Translator API this afternoon and I am now convinced that adding this TTS engine is doable.
For the user, that would be more difficult to setup because an account has to be created on a particular platform and a new application has to be declared by the user to get the identifiers that will be required by the plugin.
For the plugin itself, the new thing is that a session token has to be obtained and renewed when necessary.
So I will add this new engine to the plugin. Should be available in few days.
Any idea if the Microsoft speech is as good as Google speech ?
Note that the account is free if you translate less than 20 000 characters per month. That’s largely enough for few TTS messages per day.[/quote]
Speaking on behalf of Microsoft (I work there), i’m sure it’s MUCH better!!! ;D
Truthfully, i have no idea… and work no where near that side of the company. However, if you have any issues, let me know and i’ll try to find a contact for you if possible!
I will consider MaryTTS too.
Note that MaryTTS produces WAV files so this could be a problem with the Vera memory.
The quality of french voices looks not very good.
I just wanted to report back that running OS X in a virtual box with the tts server installed does indeed work. there are many tutorials on how to do this. here are some links for anyone else interested.
[url=http://www.macbreaker.com/2014/05/os-x-mavericks-in-virtualbox-with-niresh.html]http://www.macbreaker.com/2014/05/os-x-mavericks-in-virtualbox-with-niresh.html[/url]
[url=https://wolfpaulus.com/jounal/mac/ttsserver/]https://wolfpaulus.com/jounal/mac/ttsserver/[/url]
1 -Maybe buy a cheap Android tablet
2-Install Imperihome ,(tts engine is working)
3-Plug line-out in to Sonos line-in
Sorry for bad explanation i m not very good in English writing.
I will not go in this direction but note that Google translation v2 is not so expensive if you don’t use a lot of text. It depends on the number of characters: $20 for a million per month. It could result to less than $1 per month for our usage. But no free quota.
By the way, the Google translate API v2 seems to provide translate feature but no TTS feature !
Another one to throw into the hat is:
They are the voice behind Amazon Echo, I just signed up and you get 50’00 units per month for free but I dont have a clue what a unit is lol!
Though here is that we could eventually have Echo as a full time listener and a unified voice for the Vera created alerts…
Just a thought!
Thanks for the attention.
Johnny
[quote=“watek2k, post:29, topic:188206”]Another one to throw into the hat is:
They are the voice behind Amazon Echo, I just signed up and you get 50’00 units per month for free but I dont have a clue what a unit is lol!
Though here is that we could eventually have Echo as a full time listener and a unified voice for the Vera created alerts…
Just a thought!
Thanks for the attention.
Johnny[/quote]
Are you sure it is free or that you have a free quota ?
Maybe it is free because at a beta stage ? What exactly afyer the beta stage ?
I found the explanation about how is calculated a unit:
Number of units is calculated by dividing number of characters in the input text by 200 and rounding up the result. SSML tags are removed from the calculation. There is no pre-paid model, so no upfront investment is needed to use the service.http://developer.ivona.com/en/speechcloud/introduction.html As I understand, each TTS request will cost 1 unit if your text is less than 200 characters.
Where did you find that there is a free quora ?
API looks a little more complex. But that is probably doable.
Ivona is using chunked data to provide the audio file.It is apparently not compatible with wget that we use until now to get the audio file. So implementing the use of this new engine would probably be more difficult and require more lua code.
[quote=“kazenami, post:22, topic:188206”]Thanks to this, I’ve installed MaryTTS (http://mary.dfki.de/) on my Raspberry PI (that was only running the Hue Server for the Amazon ECHO Workaround). This is for those that don’t have a 24/7 Mac server running. Hopefully it helps
Here’s what I did (Full Steps).
- SSH into Raspberry PI
$ cd Downloads
$ wget https://github.com/marytts/marytts/releases/download/v5.1/marytts-5.1.zip
Unzip the application to the /usr/bin directory
$ sudo unzip marytts-5.1.zip -d /opt
At this point it will not run unless the you have already installed Java 1.7 you can determine the current version of Java by executing:
$ java -version
Install the required version of Java (also add openjdk-7-jdk if you intend to do any java development):
$ sudo apt-get install openjdk-7-jre
After installing the new java runtime (jre) it will still not be the default. To set the new jre to your default use:
$ sudo update-alternatives --config java
Selection Path Priority Status
- 0 /usr/lib/jvm/java-6-openjdk-amd64/jre/bin/java 1061 auto mode
1 /usr/lib/jvm/java-6-openjdk-amd64/jre/bin/java 1061 manual mode
2 /usr/lib/jvm/java-7-openjdk-amd64/jre/bin/java 1051 manual mode
Press enter to keep the current choice[*], or type selection number: 2
Having selected option 2 the java version should return something similar to:
$ java -version
java version “1.7.0_65”
OpenJDK Runtime Environment (IcedTea 2.5.1) (7u65-2.5.1-5~deb7u1)
OpenJDK 64-Bit Server VM (build 24.65-b04, mixed mode)
The runtime package delivers the scripts necessary to run the MaryTTS Server, which can be used via a browser of the client to synthesize speech. The server can be launched with:
$ /opt/marytts-5.1/bin/marytts-server.sh
This can then be used either through a browser or via the MaryTTS Client. The browser address will be:
http://(RaspPI_IP):59125
The MaryTTS Client, which is a Java GUI can be launched with:
$ /opt/marytts-5.1/bin/marytts-client.server
Launch the Server
- Edit L_SonosTTS.LUA (Download Apps → Develop Apps → Luup files)
- edit the SAY_EXECUTE variable, replace the 5th line with the Google URL to:
http://(RaspPI_IP):59125/process?INPUT_TYPE=TEXT&OUTPUT_TYPE=AUDIO&FAKE=%s&INPUT_TEXT=%s&LOCALE=en_US&AUDIO=WAVE_FILE
- Please replace the (RaspPI_IP) with your device IP and make sure to put them inside the quotes.
- Upload, replace, restart LUUP.
- Test and enjoy! Congrats, you’re done. No more relying on other TTS servers.[/quote]
Please Can you Post L_SonosTTS.LUA corrected put the correction in capital letters
Thanks
MaryTTS and Microsoft Translator should be available very soon, I am working on it since yesterday.
Maybe today or tomorrow…
Thank you! I actually miss Vera talking to me. #firstworldproblems
hmmm, sounds like they did this to block Hola traffic and inadvertently blocked this site too
glad this thread was here, save me from researching and thinking new vera release broke my system
who knows google may fix on their end too, as sure many apps blind people use are also effected
not sure if this helps anyone with a fix
but someone posted this online
got the link from this article
https://mojolingo.com/blog/2015/importance-of-vendor-relationships-in-apis/
+1
Please Can you Post L_SonosTTS.LUA corrected put the correction in capital letters Thanks
I just committed changes. You can now use 2 new engines. Please switch to the 2 new topics.
[quote=“JS007, post:36, topic:188206”]not sure if this helps anyone with a fix
but someone posted this online
got the link from this article
https://mojolingo.com/blog/2015/importance-of-vendor-relationships-in-apis/[/quote]
I will take a look. Maybe just changing the user agent in the request header could help…
The most important is that a good alternative (Microsoft) is now vailable. 8) So solving Google problem is less urgent.
The result of the wget command run by the plugin is:
wget: server returned error: HTTP/1.1 503 Service Unavailable