04. More than just Digital Signage

June 12, 2012

Another report, "Interactive Multimedia Display; More Than Just Digital Signage" was prepared for the Office of the Dean of Engineering Carleton University to update to elaborate on updated material and our plan from now on.


The term digital signage is what is generally used when dealing with this kind of project. A sign is something we look at and it displays the information that it wants you to see. A sign can point you towards somewhere you don’t necessarily want to go, rendering it useless to you. We don’t want our displays to be useless to anyone and in order to do this a certain degree of interactivity needs to be involved. The user of the board should be able to make the board useful and informative at any time with ease. In calling this project digital signage, it is automatically limited in what it is meant for. 

A more robust term that expresses what this project is capable of is Interactive Multimedia Display. Interactive as the user can decide what they want to see on the display. Multimedia because it isn’t going to just be showing some pictures, it can show videos of projects in action, websites, with the potential of even more.  

Software

• Both SpinteX and Axis do not allow for interactivity as far as I can tell.
• Powerpoint feels too simple for our requirements. Sure it can get the job done, but I doubt it’ll be the best looking solution. Might be able to do some kind of interactivity.
• Currently looking into Vodigi

1. It’s open source
2. Supports Microsoft Kinect
3. Mouse interactions
4. No keyboard (as far as I can tell)
5. Needs a web server

SpinteX/Axis

During the early stages I looked into SpinteX’s solution for digital signage as that is what the Kaliash Mital Theatre uses for their board. It did not feel very intuitive in my limited time with it. On top of that as far as I can tell it does not allow for interactivity, which makes it much less appealing then other software solutions. The advantage is a group on campus is using it and if problems arise they may have answers.
Axis’ solution is used on most of the boards on campus that are run by the campus card office. There was no demo available to use so I have no hands on experience with how easy or hard it is to develop content. It does have a lot of features such as using RSS feeds and other’s listed on their site. Since it is the software used on campus it would have a lot of support, even as much as hosting our content on their server. The advantage to this is less hardware for the Dean’s Office to manage, and our content will be displayed on boards around campus. However, other groups content will be displayed on our boards, which is not exactly desired.
There are two other disadvantages to these programs. One is that it costs money, and it is not directly disclosed how much. The other is both are closed source so no outside development can be done on it.
 
Vodigi

With Vodigi being an open source piece of software it allows for a lot of future development from other groups, such as Carleton Engineers, or other developers. I have already figured out how to change the colours of the buttons to make it more suited for Carleton and my knowledge of C# is limited at best. I am currently looking into implementing OC Transpo GPS data, but will not spend too much time on it. People who are well versed in the language can do much more than I, opening up endless possibilities to what the software can become.
Server
Vodigi is split up into three main components: Server, Player, and Uploader. The server is what hosts all the files as well as the administrative interface. The server allows for scheduling (Sunday to Saturday), without allowing to schedule end dates as far as I can tell. There is a setting called player groups that allows different groups to display different information, so it is possible for a board in Minto to have different information then a board in Mackenzie depending on requirements. The server allows for 4 file types to be uploaded so everything needs to be either: jpeg, png, wmv, or mp4. Thankfully those are extremely common and if someone submits data not in one of those formats it is likely a converter can do it problem free.
Figure 2: Vodigi Web Administrator View
 
There are a number of options in how the server can be implemented.
 
Option 1: Free hosting from Vodigi

Vodigi offers a free server from them, but it is limited to only one user. Uploading content will also be limited by the internet and there server speeds. It is however extremely easy to setup compared to the other options and it is free.


Pros:
· Easy Setup
· Free

Cons:
· Limited by the internet and their servers
· If there server crashes it is out of our control
· Our content is held by someone else
· Single user account
 
Option 2: Have our own server

We also have the option to run the software on our own server. This will give us full control over it; however it comes at additional costs. Setup is also fairly easy, but requires a fair amount of time.
 
Pros:
· We control the server
· Slightly greater controls (able to have multiple users)
· Increased speed (our internal network)
· If there is an issue we can address it ourselves as opposed to waiting for someone else to do it
· Increased options for in house development should that road be travelled


Cons:
· Additional costs
o Computer
o Software (Windows Server 2008)
· Slightly more difficult installation process

 
Option 3: Virtual machine on one PC

We are also able to install windows server on a virtual machine and run the server from there. It takes minimal extra time from option 2 to set up and eliminates the need for an extra computer. However there will be a performance hit on the PC running the server depending on the amount of RAM installed. The server will also be limited in read and write speeds as everything needs to go through the virtual machine (may be able to reduce this with the way the virtual machine is set up). If the player needs to be reset, the server will also be reset.
 
Pros:
· We control the server
· Slightly greater controls (able to have multiple users)
· If there is an issue we can address it ourselves as opposed to waiting for someone else to do it
· Increased options for in house development should that road be travelled
· Potential reduction in costs compared to option 2.
 
Cons:
· Decreased performance from running in virtual machine
· If something happens to that one PC (in open high traffic area), both screens will go down
· Still need to purchase Windows Server

 
Option 4: Install the server directly on PC

It is also possible to install the server software directly on Windows 7 and have the server running side by side with the player. This will eliminate the overhead caused by a virtual machine. However some of the issues associated with option 3 still exist.


Pros:
· No need for extra hardware
· No overhead from running server in virtual machine
· Further cost reduction from option 3 (no need to purchase Windows Server)
· All pros of option 3


Cons:
· If computer goes down, both screens will go down
· Potential decrease in performance because they are running on the same computer
o Not likely as it is not an extensive server/application
· Might not be the best practice to run server on the same PC as client

There are other options such as renting a server elsewhere, but I don’t believe that necessary as Vodigi offers a free one. In my opinion the best option overall is having our own server in a central location, that way it is in a safe environment and less likely to experience problems then in a high traffic location. However this is also the costliest option, which may leave it slightly undesirable. The next best option in my opinion is then installing the server software directly onto one of the machines. While it does pose a greater risk of needing maintenance being in a high traffic area, as well as performing two tasks at once (Client and Server), it likely will not be a huge concern. If problems do occur another solution can easily be adapted.
 

Player


The player is the software that will be on the computers directly mounted to the screens. The minimum OS requirement for the player is Windows 7, so it needs to be up to date, as well as needing a fair amount of Software Developers Kits to actually run however they are all free. The player itself can be configured slightly through a configuration file, but due to it being open source, so much more configurations can be done. It is in here I changed the colour scheme from blue to red, and am looking into adding in bus data.
Uploader
The Uploader is a Windows based application that you can sign into given a username and password, and upload data to the server. This data will need to be set as a video or image from the administrative view and then later in a slideshow or playlist. The main advantage is that there is no file size limit and it is possible to upload multiple files at the same time. From the administrative web view, there is a file size limit of 100MB and you can only upload 1 file at a time.
An additional use for this is to give someone you don’t always want in the administrative view, but still want to give the ability to upload content, and then the Media Uploader works well as unless they know where to look for the servers IP it’s not obvious.
Figure 2: Simple interface of the Media Uploader.
 

Layout

The way the software is set up there are two main kinds of content you can display, videos and images as well as being able to open websites, webcam, and the weather. The default screen has to be a video or a slideshow/image, but with the interactivity it is possible to select a menu option that brings up more options across the screen. This allows the user cycle through the options and makes a selection. These selections can open up a webpage, webcam, another slideshow or video, or the weather. This is where I think Carleton students can improve in terms of the way the information is displayed and especially interacted with through the Kinect.
Figure 3: Clicking the button in the bottom right corner brings up this screen
The way I currently envision the content being laid is simple yet can be extremely informative. The main page slideshow should have the directory as that is one of the main purposes of the board, or at the very least a notice informing students/staff/visitors how to access the directory. The main slideshow should also include other important information such as upcoming events, new professors, awards from professors, and whatever else is deemed important. Every couple of slides or even incorporated with the button should tell users that it is able to be controlled by hands (and potentially voice).

After the menu button is selected, a line of different pictures (with labels) will be available for the user; the first one will be the directory and “clicking” that button will open up the slideshow that displays the directory. In this directory I believe not only professors should be in it but also staff members and student groups, this gives users a centralized location to find data and makes it much easier to change if an office is moved. In addition to just a name, office, and extension appearing on screen, I also think department and position are important. If a student is looking for an academic advisor, but they don’t know the name, they are still able to find them using our directory. The look of the directory slideshow is unfortunately something above my expertises.

The directory can be further enhanced with increased development. For example, it could be possible to browse individual professors through the directory page and see what they are working on. This of course will require more updating on the administrative end, an alternative is that Vodigi can open up a website and for example clicking a professors picture can bring up there page on the Carleton site.
 
Content

o Directory
§ Professors Offices/Extensions
§ Staff Members (Technical support)
§ Other locations

o News from the faculty
§ Award winners
§ Upcoming events
§ New labs
§ Closed labs

o News from Society’s/other groups
§ Upcoming events
§ Services

o Information about a professor (In the future be incorporated in the directory with Kinect)
§ New professor every so often
§ Research being done

o Information about a project
§ Detailed enough so people can learn something, but simple enough that anyone can understand what they read

o Bus Schedule
§ Useful especially in Minto during the winter when people stand there to wait for buses
 
The above are what I think as a student is useful/interesting information, especially with regards to engineering. Some pages will have more information than others, but that is fine as often I find myself reading about the famous inventors/engineers around Minto, so new information would be nice to have. Even if only brief information is desired for the main slideshow, with the extra menu more detailed information is easily accessible.
I am currently in the process of compiling a list of all the professors, staff, and other groups in the faculty as well as extensions and room numbers in order to create the directory pages. The exact layout of these pages has not been decided, but the requirements of being easily editable and useable are priority.
 
 In terms of the other content, I was thinking having a template for the different departments and groups so it is more distinguishable when looking at quickly. Perhaps just a border with white in the middle and text in that space with pictures to allow for easy reading or something more exciting with more pictures and contrasting fonts to make it legible.
In terms of submitting content, I think it should be the responsibility of the group submitting it to have it formatted correctly. In respect to the actual submission of content there are two easy options.
 
Option 1: Email the Content

Then the group submits the content to whoever needs to approve it (the Dean), and once approved it needs to be added to the server. The file needs to be uploaded, file type selected, and then added to the slideshows (takes a few minutes to do). This can be taught to the Dean’s staff.
 
Option 2: Allow outside groups to upload

This way it is possible to give outside groups an account and allow them to upload the content via the Uploader. This will not put it directly into the slideshows and still needs to be added. If the content is deemed inappropriate it can be deleted, if it’s accepted it can simply be added to the slideshow or any section desired.
I am also working on a way to not have a username and password need to be entered on a modified version of the Uploader (still needs a user created in the server). The disadvantage to this is that if someone distributes the EXE then others can add content, though it could never be seen as it still needs to be approved. The advantage being that the person using it doesn’t know the user name and password and cannot enter the administrative view.
It may be possible to only allow “administrators” to sign into the web view, eliminating both problems, but by default it is not able to do this.
 


Hardware

In terms of being the very best boards, hardware plays a little role compared to content. This is with the exception of hardware that improves interactivity. In terms of interactivity, a button was the first idea. Press the button and it will bring you to the directory, very basic. Then the idea of multiple buttons came up, such as in a keypad. Allow the users limited interaction with the device, but not very exciting. From the keypad came the idea of a touch screen which can be exciting, but can be costly and it has been done before.
Something that hasn’t been done, at least not widespread is the use of motion tracking hardware, specifically the Microsoft Kinect. As mentioned above, it allows the user to interact directly with the screen at a distance and can even use voice commands. The amount of development that can be done using this device is endless. In my currently limited experience with the device I have gained a far greater understanding on how it works, and developed ideas on how it can be used to make an Interactive Multimedia Display an even more useful device.

Kinect

Please see my report on the Kinect for my preliminary research on Kinect. This section will outline what I have learnt hands on.
Through my initial testing using the provided tool kit, I learnt about the basics that the Kinect can do. It can change its physical angle from -27° to +27°, allowing for nearly unlimited mounting positions and could allow for other applications such as tracking.
It is aware of direction of sound as well as volume. More testing needs to be determine if it will pick up someone right in front of it even if people are being loud off to the side, though documentation says it can.
It can detect up to 6 people in front of it, and track the 20 point skeleton of 2 of those people. The 20 points allow for many options in terms of different gestures and movements to interact with the system. In our case it allows for the Kinect to act as a mouse cursor and when the user holds there hand over a button for 3 seconds it will open the content.
The Kinect has 3 different video outputs. It can display regular images, but it also can output depth information and 3D images.
In terms of how Vodigi is current set up a user will walk up to the screen and raise their hand, the cursor will inevitably move and the user can direct it to the bottom right hand corner and hold it there for 3 seconds. When this occurs the menu overlay will come up and this will allow them to either use the voice commands “Next, Back, and Close”, or continue to use their hands.
With a little bit of time, I could likely add a voice command to bring up the menu so that the whole system can be controlled through the use of voice or movement.

Screen

As mentioned the choice of screen does not have a dramatic effect on the overall quality of the display. In terms of longevity of it, I can recommend to get a well established brand name display such as Sony, Sharp, or Samsung. They are very well established in the TV market and have a history of producing quality units. There are some requirements that I would like to see in the screen and they are:
· Minimum 1 HDMI port
· 1080p resolution (allows for maximum quality content, and won’t limit anyone)
· VESA mountable (should be standard on all displays)
Everything else such as actual size of the display will depend on the overall budget for the project. From asking other groups the range of screens used is between 47” up to 60”. I believe something around 55” would be a good size and relatively well priced.
 

Mount

In terms of mounting it to the wall there are 2 main options: mounted to the wall, with or without enclosure. In the proposed location it is not feasible to hang from the ceiling like other groups have. The Minto Screen will be best served by no enclosure as the wall is curved and would not be pleasant to look at. The doors are far enough away and there is little potential for stuff to fall on it.
In the Mackenzie location, the screen will be right next to doors as well as the ceiling has been known to leak. In this situation I believe an enclosure will protect against the environment and also provide addition security for the lower mounted display.
 

Security

For the Minto screen, the mount should have options to lock it, plus there are cameras in the hallways. Security through the Carleton network can be attached to the wall so if the screen is removed and alarm will go off. In terms of securing the Kinect it is possible to remove the base and drill a hole in it to mount it. I would also recommend attaching a clip with a security screw to the cable and attaching it to the mount or the wall. It would also be easy to secure a security tab to the back of the device. The security should be subtle as many outside people will see this.
If an enclosure is used for the Mackenzie screen the security is obviously there. A camera is also in proximity to the display and security through the Carleton network can easily be attached. In terms of the Kinect, the enclosure would need to be taller than screen, which is less aesthetically pleasing, to fit the Kinect on the inside. The other option is to mount it on top of the enclosure by drilling a hole in the both the enclosure and the base of Kinect. As well attach a security tab and security screw mentioned above. This will leave the Kinect exposed to a dripping ceiling, and be in close proximity to the outdoors. This in my opinion is not a real issue as while the Kinect is not a military grade piece of equipment it has no open electronics. The one concern I would have is the piece connecting the body to the front plate has a little gap, but a clear piece of tape can cover this and prevent most, if any, water from getting in. It will also be angled down and any drops of water would drop off the front instead of resting on it.


Power and Network

In both cases I think getting Physical Plant to run a conduit with the network and electrical wires will provide the cleanest looking solution. Without it wires would need to be hung over the door way in Mackenzie, and down the wall in Minto. Each location would need 3 ports for power, and 1 network cable.


Computer

The requirements to run Vodigi Player are not very high; in fact they do not list any requirements besides the need for Windows 7. The server will need more RAM to accommodate the servers and Kinect in order to run smoothly and potentially a larger hard drive to hold the content. There are 3 options in my mind that are feasible if 2 machines are decided on, more if a dedicated server is used.
 
Option 1: ASUS Eee PC

Specifically the ASUS Eee Box EB1021-B059E, this PC is a small form and can fit directly behind the screen. I have installed similar models for the Systems Department. The hardware inside is not top of the line, but it should be good enough to run Vodigi Player with the Kinect. The one running the server will need more RAM as the 2GB included is a pushing it a little close for all that it needs to be running. A 320GB hard drive is included with, which should be large enough to store all the content.

Cost Approximations:

380$ Per Device x2 = 760$ +RAM Upgrade (4GB) 30$ +25$ Shipping +Tax (13%) = 920.95
Keep in mind these are not exact numbers and are subject to change
 

Option 2: Custom Built

This option is pricier but will not be limited in terms of processing power as there is not much room to upgrade in the ASUS Eee Box. It is possible to put together a PC that will have greater upgradability and initial power for approximately 485$
An example of such a PC is below. The ASUS Eee Box does also include a VESA mount, wireless keyboard and mouse, and has wireless built in. However with the exception of the VESA mount those are not necessary. The price can be reduced to below the cost of the Eee Box if a copy of Windows 7 does not need to be purchased.

Cost Approximations:

With Windows 7
$485 per Device x2 = $970 +45$ Shipping +Tax = $1146.95

Without Windows 7
$365 per Device x 2 = $730 +45$ Shipping +Tax =$875.75

Figure 4: Possible build for Custom PC

It is my recommendation if possible to go with the custom built PCs as they are fully customizable, and allow for greater performance. It also removed the concern if the PC is good enough to run a server and if the software is developed upon the PC will still be able to run it.


Conclusions

In conclusion, I recommend that Vodigi is used with a dedicated server is budget allows, otherwise the server installed directly to the player PC. The custom built PC is also my recommendation as it allows for more flexibility and power. The Minto Lobby display should be wall mounted with locks and security tabs. The Mackenzie display I believe is better suited for a full enclosure due to environmental concerns and being mounted lower. Power and Ethernet will need to be run by Physical Plant in order to ensure it looks as clean as possible. In regards to the screen itself it a rough idea on budget is needed as there are many variables that come into play, however Samsung is my recommendation as they make good looking displays and are consistently among the top rated.
 

What’s Next

In the coming weeks, more work on the Kinect and how to better its integration with Vodigi will be done. In its state now it is very usable, but I see it has the potential for much more. Content will continue to be developed based on the use of Vodigi.
A more in depth look into the hardware will also happen, as most of the time has been dedicated to the software and interaction components of the project.
As more details about budget constraints are known, an ordering plan will be drafted.