When people see my VR panoramas for the first time their comment is usually something like "Wow, incredible, awesome!" followed immediately by "How do you make them?".
Many people assume the panoramas are some sort of video, made by panning a video camera around in a circle. Actually it is much more sophisticated than that. What you are seeing is a single image, a cylinder or a cube, viewed from the inside and geometrically corrected in real time as you look around.
The panoramas are made by piecing together a number of photographs, taken rotating around in a circle. The resulting image can be a cylinder (its left and right edges match up) or a sphere (360° around by 180° up and down).
Flash software from Macromedia allows you to view the image from "inside", making it seem as if you were there. That's why we call it virtual reality - a VR panorama.
The first generation of VR panoramas were cylindrical - they had a limited view up and down. Now we make spherical panoramas - all the way around plus straight up and down. The older images on this site are cylindrical, but all the newer ones are spherical.
How To Make a Spherical VR Panorama
Step One: Take the Pictures
Using a full-frame fisheye lens, take a series of 6 pictures from a single point, rotating around in a circle, plus shots aimed straight up and down. There should be generous overlap between adjacent images.
For best results the pictures should be taken very precisely, so a tripod and special camera mount such as the Nodal Ninja are usually used.
Step Two: Stitch the Images Together
Special software such as PTGui or AutoPano is used to to combine the individual frames into a single image, procedures known as stitching and blending.
The program first warps each image to a sperical projection. This allows the geometry of overlapping images to match. Then it correlates features in the overlap area of each pair of images to get a good match, and blends the overlapping area to produce a smooth transition. It continues with the following frames to create a single stitched image.
The result is a seamless panoramic image, twice as wide as it is high, known as an equirectangular. It represents 360 degrees of view horisontally - the left and right ends match up. The top and bottom of the image are the views straight up and down, greatly distorted.
Step Three: Retouch the Image
The resulting image will pobably have a number of flaws such as stitching errors and lens flare. These are usually corrected using Adobe PhotoShop. This also afford an opportunity to tweak color and to lighten shadow areas.
The panoramic image is typically much larger than needed for interactive use, so it may also be reduced in size, or this can be done automatically in the next step.
Step Four: Create the Flash Movie
A program such as Pano2VR, KR Pano, or Flashificator is used to transform the equirectangular image into cube faces and package them up in a special VR movie format. In this step you can also set the beginning view's pan, tilt and zoom.
I routinely make two sizes, standard and fullscreen, and create a thumbnail based on the beginning view.
Step Five: View the Image
For the movie to play as a VR panorama it is necessary to use Flash from Macromedia, either a standalone player or a web browser plug-in. This warps the view in real time from cube faces to a natural looking perspective, and allows panning up, down, and all the way around, as well as zooming in and out.