Hello,

I’m developing a small module that reads text through Google-text-to-speech (see http://translate.google.com/translate_tts?tl=en&client=tw-mp3&q=What else ).
The url returns a video that plays an audio file. When I test it locally on my Mac, the autoplay works without problems.
But when I transfer all this to the Raspberry Pi, the video is added to the DOM but does not launch.

Has anyone ever encountered this problem?

There is the module MMM-TTS but unfortunately the librairy Festival has no French at its disposal. That’s why I’m testing other solutions.

Sincerely,

@Mykle1 Ah! I found what was bothering me. The image of the moon does not change … Indeed, the image seems more decorative than informative. As a result, this does not show the current moon phase at all.
It was rather in this sense that I had to understand my previous comment.
The moon-phases module did its job very well. Until the cache problem appears.

Hi,

IMO, the TV monitor is more expensive than a computer monitor.

For your project, you need to find your screen before you decide on the dimensions. And for the width, do not forget the width of your wood. If the mirror have to be 50cm width, then you have to choose a screen from 39" to 42" max. The panel (your nude screen) will be less than 50cm with one of those dimensions.

I don’t know the price in your country for the one-way mirror, but it’s expensive too… In France, my choice is too use an acrylic glass and then put a one-way film on it.

Your project looks more or less like mine:
https://translate.google.fr/translate?sl=fr&tl=en&js=y&prev=_t&hl=fr&ie=UTF-8&u=https%3A%2F%2Fwww.teddypayet.com%2FMagicMirror%C2%B2-un-socle-qui-lui-va-comme-un-gant&edit-text=

@Mykle1 : Yeah, I saw that (in another topic too). However, from what I see in your code, there is no identification of the current location. Which would allow to have information maybe more precise.
In addition, the various texts displayed do not interest me on my installation. And there is no possibility to disable line by line (CSS or javascript) these texts.

Hello,
Has anyone managed to make the Google Cloud Text-to-speech API work?

There is indeed a Text-To-Speech module that works: MMM-TTS.

Unfortunately, the libraries used do not take into account the French language. The alternative solution is the use of google-tts.
I tried to create my own module to integrate the module google-tts-api
Not being used to developing under Nodejs, I have trouble making it work as I would like.

Any suggestion is welcome.

Thanks.

Same here as well.

I found it! Perhaps, the answer will help someone.
MMM-TTS need a string as “payload” not an array. So, here is my solution :

    {
        module: "MMM-ModuleScheduler",
        config: {
            notification_schedule: [
                {notification: "MMM-TTS", schedule: "45 6 * * *", payload: "Have a nice day. Courage."},
                {notification: "MMM-TTS", schedule: "0 8 * * *", payload: "Do a barrel roll!"}
            ]
        }
    },

Easy-peasy!

Hi!

I understand this issue. I use a gitlab private repository. I clone it in “magicmirror_config” in /home/pi/
And then I created a symlink im /home/pi/MagicMirror/config/ for my versioned config.js

And all my personalized files are in symlink where I need it.

That’s really easy and like that I can test on local device (my Mac) before put it on the raspberry Pi.

Hello,
I am new to this forum. I have successfully installed my MagicMirror recently. I installed the following modules:

• MMM-AlarmClock
• MMM-FreeBox-Monitor
• MMM-page-indicator
• MMM-pages
• mmm-moon-phases
• worldclock

Everything is working. Now, I would like to test the plugins MMM-ModuleScheduler and MMM-TTS. My idea is to program the reading of texts at specific times. (Wish a good day to my wife when she awakes, for example.)

What should I put in my config.js file for it to work correctly?

Thanks for your help.