How We Made Thayer360

Credits

Thayer360 grew out of strong interdepartmental collaboration among Thayer School's Communications, Computing Services, Development, and Career Services offices. The team members are (in alphabetical order):

The Idea

Dean Joseph Helble provided the vision: a way for people to "visit" virtually to better experience the magic of Thayer School than is possible from our normal web site. This vision came out of discussions Dean Helble had with Thayer School alumni. We assembled the team and brainstormed how to best provide the desired experience for our virtual visitors. We considered a number of candidate technologies, but settled on panoramic photography with rich multimedia content relevant to the panoramas. We realized that today's typical broadband connection could sustain the traffic required for full-screen, beautiful, high resolution, 360-degree panoramas which provide a very different experience that the typical "real estate ad" panorama people are used to seeing.

Panoramic Photography

We used a DSLR with a wide angle lens. That was the most expensive part (we used nice Nikon gear), but one should be able to use just about any digicam with a wide angle lens. We used a Manfrotto indexing tripod to facilitate pivoting around the right point and quickly getting the pictures from multiple angles. Here are the details of the equipment we used:

Taking the many photos for each panorama (we used 13) just right takes some careful setup, practice, and skill. But one can get started with reasonable results without too much effort. There are excellent technical notes about panoramic photography on the Virtual Reality Photography web site. We highly recommend you read the articles there.

Here's a video showing how we shot the 360-degree panoramas:

After taking the 13 photos, we used PTGui to stitch them into the panorama. Again, this is relatively easy to do, but to get it just right requires practice. Getting rid of the tripod in the panorama's down view is the hardest part, but you can simply leave the tripod in the panorama. Depending on the room, this took between 30 minutes and two hours per panorama. The tricky part is getting the stitching control points right. PTGui does a good job inserting most of them automatically, but has trouble with certain situations like repeating patterns (e.g., the floor in the Great Hall), great curving lines (the tabletops in Spanos Auditorium).

Putting the Panoramas on the Web

We display the panorama and add content using krpano. This involved some flash and HTML5 programming, and no small amount of skill too. We designed and programmed: the navigation (including transitions between rooms), menus and icons for interacting with the tour, the map, initial orientation when entering a room, how rooms connect, displaying slides and videos within a room, etc. We developed the user interface incrementally and had people outside the team test and evaluate it periodically (even recording video of some of those sessions so we could analyze the users experiences and reactions and make improvements based on our observations and analysis).

We used our existing web server to deliver Thayer360. Each panorama is around 6 megabytes.

Content

Finding, generating, and adapting all the content is the most time consuming part of Thayer360. Students contribute some of the content, especially videos. Staff does the rest because logistics and timing make it impractical to have students work on the other pieces. We use Apple iMovie and Final Cut Pro to edit videos.

Much of our content is about events that happen at particular times. We need to plan ahead to make sure we are ready to capture appropriate video, CAD materials, interviews, etc., at the right times. We waited for sunny days and blooming flowers to take the outdoor panoramas and sought bright days for panoramas of rooms with outside windows.

We put a lot of thought into what spaces and content to include in Thayer360 and into what content to place in which spaces.

Costs

By far the most expensive part of Thayer360 was the labor we put into it. We estimate about 12 person months total went into this project. A good portion of that labor was spent brainstorming, evaluating alternative technologies, selecting tools, and setting up the framework for Thayer360. Making another similar tour would go more quickly because we wouldn't need to repeat that work. But most of our time went into refining the presentation and generating content.

The camera and lens and tripod were the next biggest cost, with the DSLR being the biggest ticket item. We chose moderately expensive equipment, but one could use significantly less expensive gear without undue loss of quality. The PTGui and krpano software were relatively inexpensive.

Future Plans

We plan to update and add content regularly as events happen and to convey different messages to our visitors. We also plan to take new outdoors panoramas to show the winter season.

We have been monitoring Thayer360 usage. We can monitor the landing page and the tour page using Google Analytics, but we can't monitor the flash-invoked elements that way, so we wrote a simple bash script to analyze loads of panoramas, videos, slideshows, info windows, the map, etc. We plan to fine-tune the user interface based on these measurements to try to ensure that people see the most important information (we noticed that early users were much more likely to view outdoors videos than the rest).