Sharing ProPresenter lyrics to multiple clients through the web browser in real time using Node.js, socket.io, and Amazon EC2

Every year, my church has a “night of worship”, a worship service in the heart of the city at an outdoor stage, where we sing songs for a couple of hours. Because it doesn’t get dark enough to use projectors for lyrics until the service is almost over, in the past we have relied on using small flat screen TVs to try to show some words for people to follow along. Big white letters on a black background, nothing fancy. Of course, it’d be great if we could just rent an LED video wall, but the cost to do that has been too expensive for us to do in the past.

Photo Oct 19, 3 34 24 PM
You can see the screens we rented here. Pretty small (60″) for such a large crowd.

So, I had an idea: What if we could somehow send the lyrics out of ProPresenter to everyone’s phones, in real time, and let them use their own screens to follow along?

I gave myself a couple of limitations:

  • It needed to work in the standard phone browser so there was no barrier of installing a particular app
  • It needed to be real time or as close to it as possible

Awhile back, I started tinkering around with the undocumented ProPresenter API. I say undocumented because it is not officially offered as a way to access ProPresenter data and control it. Some people have done a great job at figuring out how ProPresenter sends data over the network between their apps which allows us to extend the software to meet unique needs. Basically, by using websockets, we can interact with ProPresenter which will return JSON-formatted data reflecting information about songs in the library, playlist, the index of the current song, the current slide and next slide information, etc.

I created a local Node.js project and in just a few hours, I had something ready to alpha test! My approach was to have a web browser open on the local network that could poll and listen to changes from ProPresenter, and then relay that new data to a web server. That web server would then relay those changes to all of the connected clients, much like a chat server would send a message to everyone.

I showed it to my team, but the idea was tabled for awhile because we thought it through and didn’t want people buried in their phones while singing. However, as we got closer to the event, we realized that the two TV screens we rented might not be sufficient, and I was asked to work on this again.

As I was preparing this for production, I discussed briefly with our IT team about setting up an internal server running Node.js that could be accessed on port 80 (the default HTTP port) outside the firewall, but bandwidth, security and performance for hundreds of clients connecting through the Internet all at the same time was a concern. With that in mind, I turned to Amazon EC2.

If you haven’t heard of it, Amazon Elastic Cloud Computing (EC2) is basically virtual servers “in the cloud” (i.e. remotely and accessible from the internet). It’s not too hard to set up, and they even have a free tier available for 12 months, so you can try it out for free! I had never used it before this project, so I actually followed a tutorial to help me get it going.

Screen Shot 2018-10-21 at 6.21.09 AM
This is my EC2 instance. It’s running Ubuntu linux.

Once I had my Linux server set up on Amazon EC2, I assigned an “Elastic IP” (Amazon’s term for a type of static IP), and then I bought a domain name, fglyrics.com, for $12 and tied it to that IP. It was up and online in minutes. I installed Node.js on my new server, copied over my code, and started it running.

About the software:

I call the software Presentation Bridge, because it acts as a “bridge” or connector between the presentation software and all of the clients.

Screen Shot 2018-10-21 at 5.19.59 AM
The initial Bridge page.

When you first load the Bridge page, you have two options: Configuring your ProPresenter connection, and Connecting to a Bridge. In order to connect to ProPresenter, you have to enable the network settings. It relies on both the “remote” and “stage display” controls to get all of the data needed.

ProPresenter 6 Network Settings
You need to enable the network, ProPresenter Remote, and Stage Display App. Be sure to assign a control password. The Network Port is the port we will use in Presentation Bridge. The Stage Display App port is not needed.

When connecting to ProPresenter in Presentation Bridge, you’ll need to supply the local ProPresenter IP address and port, as well as the control password and local library path. This is what allows the Bridge to pull all of the slide images and other information. This is standardized across all of our ProPresenter installs at our church, so it’s always the same path for us. It should be the full, absolute path from the root of your drive.

Screen Shot 2018-10-21 at 6.15.51 AM

When you’ve successfully connected, the gray dot at the top of the ProPresenter config box will turn green. If there’s an error, it will turn red. Status information is displayed in a log area further down the screen.

To connect to a bridge, choose one from the dropdown list. If it is configured to have a control password, you’ll have to enter that in order to connect. Adding bridges and making changes to existing bridges can be managed by clicking the settings wheel. It supports multiple bridges, which I added as a feature since we have multiple auditoriums and may want to use more than one simultaneously.

Screen Shot 2018-10-21 at 6.02.00 AM
This is what the page can look like when connected to both ProPresenter and to a Bridge room.

As the operator runs ProPresenter, the slides will be displayed on the Bridge screen with the currently selected slide showing a blinking red border, so it is clear which slide is currently being displayed. You can browse the playlists and items in the playlist. You have the option to send the current data from ProPresenter to the server (or turn it off), turn on a logo (configured in Bridge settings, useful if you’re currently not connected to ProPresenter, etc.). I also implemented the NoSleep.js library which will attempt to keep any connected mobile devices awake.

On the viewer/client side, I implemented three types of “listeners”:

  • Text Listener – just gets text data and displays it as big as possible on the screen
  • Image Listener – displays the actual slide image by using a base64 encoding of the slide
  • Stage Display – recreates the current slide/next slide layout
Screen Shot 2018-10-23 at 3.54.23 PM
This is the default “text listener” option, and what we used for our night of worship.
Screen Shot 2018-10-23 at 3.54.01 PM
This is the image listener. It uses a base64 encoded image pulled from ProPresenter at the time it connects. The quality is based on the slider value set in the config options for the ProPresenter connection. It does not include any background or other layers.
Screen Shot 2018-10-23 at 3.54.14 PM
This is the stage display, with the current slide and the next slide, and any notes attached.

All three listeners can be accessed through the browser. The data is relayed from the server using the socket.io library. I tested it on my iOS devices, Android devices, and even my Amazon Fire TV stick on multiple browsers and they all work really well. Across an internet connection, the moment a slide is clicked in ProPresenter, that slide is visible on the listener devices.

We used it during our Night of Worship this year and it worked great! I used a hotspot for the Bridge connection and then everyone connected to the Text Listener using the internet connection on their own phones. It uses very little data since it is just a text stream, which is really nice!

Photo Oct 19, 3 28 26 PM

Photo Oct 19, 7 17 43 PM

Overall, I enjoyed creating this software for our unique need. I plan to extend the functionality down the road as I have time, including attaching “triggers” to specific slides as they are activated, to send RossTalk messages, fire HTTP cues, etc. on the local production network.

If you’d like to try Presentation Bridge out for yourself, the code is freely available on my Github repository. You can also request to demo it using my live site running on Amazon EC2. I wrote the software to support multiple bridges at a time, in case you have multiple meeting spaces or venues that need to run simultaneously. When more than one Bridge is enabled and running, any users that connect are presented with a drop-down list and can select the Bridge they want to join.

Network control of window shades using relays, a Raspberry Pi Zero, and Node JS

At my church, we have a smaller auditorium that has a giant window to let in natural light. We don’t use it much during services because the light can wash out our projection screens, but it’s wonderful to have during the week to let in light. We also like to have it open before services begin and close it right as we get started. It helps to create a nice environment.

Photo Oct 03, 5 08 02 PM.jpg
These are the two big windows. Each one has a motorized shade attached.

These shades came with a very easy to use RF remote. Actually, it came with two remotes. Since we never need or use two, I thought, wouldn’t it be cool to hack one of them and turn it into a network-controlled remote? This would allow us to control the shades from anywhere, not just within the RF range! Plus, I like automation and making things more accessible and efficient and I wanted to try to create something inexpensive rather than just purchasing something already made.

somfy remote
This is the remote we have. We have two shades that we usually open simultaneously, so we have two remotes but don’t need both.

I purchased this USB relay off Amazon:

usbrelay
These relays are pretty cheap and available on a lot of sites. I paid about $12 for one off Amazon. They can work with both low and high voltages.

I figured I could use the relays to electrically “click” the buttons on the remote. So I took the remote apart to see how it looked on the inside. Here is a picture of the remote outside of its shell.

Photo Sep 20, 11 35 45 AM
The first and third buttons are the “up” and “down” controls. The middle button is the “preset level” where the shade can open or close to a preset level, but we don’t use this feature. The very bottom button determines which shade(s) to control.

I knew there was probably a way to use the GPIO pins of a Raspberry Pi and transistors to trigger the button presses, but that’s not my forte. I don’t know a lot about circuity and electronics at that level, but I know how to solder (well, sort of. Don’t look too closely at my work.) So I soldered some wires onto the remote buttons.

Photo Sep 24, 3 52 18 PM

When I touched the wires together, the remote activated! The concept worked, so I hooked it up to the relay to begin the development and testing.

Photo Oct 02, 1 00 27 PM.jpg
Here is the remote, connected to the relay via the NO (normally open) pins.  I bought a small electronics project box to fit all the parts in.

I knew I wanted to control the remote using Node JS. It’s a great lightweight platform that runs on multiple operating systems, including Raspberry Pi’s with Linux. I found control software written in C, C++, Python, shell scripts, etc. for the USB Relay that I had bought, but nothing in Node JS. At least, nothing native without using DLLs and that sort of thing.

It’s been a long time since I wrote in C++ and I’ve never used Python, but I can read both languages enough to learn about the control protocol. I wanted to keep mine running with the least amount of dependencies so it could work on multiple operating systems (not just on the Raspberry Pi, for example).

In about an hour, using the Node-HID library, I had a working script in Node JS to turn my 2-relay device on and off. I decided to make the control code for the relay into a separate Node JS module so that it can be used in other projects. If you want to use it, I’ve made it available on Github as well as published it in the NPM registry.

Once I had the control of the relay that I needed, I whipped up a quick Node JS project running express to host a simple web server that would accept a “shades up” and a “shades down” HTTP request and then trigger the appropriate relay.

const USBRelay = require("./USBRelay.js");
const relay = new USBRelay();

//express API variables
const express = require('express');
const app = express();

//http server variables
const http = require('http').Server(app);
const listenPort = 4400;

//routes

app.get("/shades_up", function (req, res) {
   ShadesUp();
   res.send({returnStatus: "shades-up"});
});

app.get("/shades_down", function (req, res) {
   ShadesDown();
   res.send({returnStatus: "shades-down"});
});

http.listen(listenPort, function () {
    console.log("listening on *:" + listenPort);
});

function ShadesUp()
{
    relay.setState(1, true);

    setTimeout(function () {
        relay.setState(1, false);
    }, 1000);
}

function ShadesDown()
{
    relay.setState(2, true);

    setTimeout(function () {
        relay.setState(2, false);
    }, 1000);
}

And by making an HTTP request like:

http://ipaddress:4400/shades_up

 

I can trigger Relay 1 on my device to turn on for one second, and then turn off, simulating the button press.

Because we use Ross Dashboard as our go-to production interface, I added Up/Down buttons to our FOH custom panels to send those commands to the relay server.

Screen Shot 2018-10-03 at 6.48.16 PM
Even though the remote is easy to use, this works from anywhere on our network!

I then deployed the ShadeController Node JS project to a Raspberry Pi Zero, and put all the parts into my project box. I had our IT department make a DHCP reservation for the Pi so that it would always have the same IP address, making it easier to program around.

Photo Oct 02, 3 24 42 PM
Here is the remote, the USB relay (hot glued to the bottom of the box), and the Pi Zero all in the project box.

So, what does this really gain us, if the remote was already pretty easy to use?

  1. We are no longer limited to RF range of the remote to control the shades. We can control it from anywhere on the network, including our control rooms which are located far away from the receiving end of the RF controller for the shades.
  2. We can use the URLs in the ShadeController program as hooks in other projects to control the shades through automation. Time for the service to start? Just run the appropriate light cue on the light board and it fires a network midi command to a server that then relays the HTTP request needed.

Overall, this was a fairly inexpensive project.

I ended up spending a little bit more because I bought a Raspberry Pi Zero W kit with a case and some other adapters that I didn’t already have, like a mini HDMI to HDMI connector, but if I end up buying another Zero, I’ll take the Pi case out of this project box and use it. I also bought a pack of project boxes for about $8 that were way too small and didn’t fit my parts, but I plan to use them for some other projects down the road.

I hope this gives you some ideas on ways you could automate or more efficiently control your non-networked equipment through the use of relays! I certainly learned a lot during this process and am looking forward to implementing it in other areas.