Often at the climax of a hike in Hawaii, I've come across breathtaking vistas. The still photographic image pales in comparison to the realities of actually standing there, basking in the full panoramic magnitude of the scene. Quicktime VR panoramic movies recapture some of the original experiential nature of the original scene by empowering the viewer to tilt, pan and even zoom around an entire 360 degree rendition of the scene.
Quicktime VR movies are extensions of Apple's ubiquitous digital video format. Quicktime VR's can also be object-based although we'll discuss them in a later demystifier module. QTVR's use Quicktime's video compression engine to project a spliced up version of your entire scene onto an imaginary viewing sphere. QTVR movies dynamically jump to specific frames of the video based on the user's input. For example, if the user moves the mouse left or right within a QTVR window, the movie jumps to the appropriate adjacent frame in the movie to create the illusion of moving your field of vision. This illusion is further strengthened by angular distortions the QTVR applies to your view as you pan up or down. As an added bonus, you even have a built in pair of binoculars, courtesy of the option and control keys.
Although my examples do not feature this option, Quicktime VR panoramic nodes (individual scenes) can be linked together allowing for exploratory navigation.
Numerous subjective factors contribute to a compelling VR node. It is useful to develop an awareness of where your subject matter falls in relation to an imaginary horizon. Because the viewer pans up and down based on this line, interesting subject matter should be distributed throughout their navigation of the scene. Realism is further enhanced by wide variations in objects proximity to the viewing center. Many other factors such as the sun and your own shadow play an important role in selecting a good location for a successful panoramic VR node.
|Cameras, digital cameras or even video cameras can be used to capture the panorama. I prefer 35mm slide film for the maximum quality of the resulting VR. Lens selection will determine the number of overlapping images needed to capture a full rotation. I use a 15mm non-distorting rectilinear Nikon lens that requires 12 images (30 degree rotations) per node. In order to minimize efforts in photomontaging the individual shots back into a seamless single image you need your camera held in perfect horizontal and vertical alignment and rotated around the focal point of the lens. In addition, precise angular rotation all make a Kaidan Professional tripod head a serious consideration. You'll probably want to bracket your exposure to ensure a complete set of usable images.
Although there are much easier ways to combine images through auto stitching applications, they will require more software and usually result in slightly degraded images. The first task is to digitize with enough resolution and use PhotoShop to combine the individual images back into one seamless panoramic source file. Aim at about 800 to 1000 pixels high for your files. Use a combination of layers and layer masks in PhotoShop to selectively erase each adjacent image into the next. The offset filter can be useful in getting the final image to merge seamlessly with the first.
The second task is to prepare the file for splicing and compression with Apple's "QTVR Make Panorama 2" software. The source image needs to be flattened, rotated 90 degrees counter clockwise and have the height and width devisable by 96 and 4 respectively. You can downsample the source file in order to decrease the final size of the QTVR node. Save the document as a PICT file and import it into the "QTVR Make Panorama 2" software to add the final variables of viewing window size, initial zoom and pan location.
If you're hungry for a discussion on Quicktime 7's VR production, read Apple's QuickTime VR page.