The original aim of this project was to develop and demonstrate a QuickTime VR object movie that permitted the user to manipulate and view the specimen from any angle. However, as can be seen from the example, this is not the case; the specimen can only be viewed from one perspective. Over a period of a day and half, a total of 181 images were taken of Elphidium crispum, most of which were not used in the final product. As this was an experiment, much time and effort was expended to determine the proper procedure. Although we include a description of all our mishaps below, this is not meant to dissuade future creators of QuickTime VR movies for the purpose of micropaleontological illustration. Following the correct procedure, an adequate QuickTime VR movie could be assembled from as few as 20 images.
For the initial run it was decided images were to be captured every 10° in the X-Y plane for a total of 36 images and coverage in the Y-Z plane would be every 22.5°. The first series of images had to be aborted midway through when it was discovered that the increments used to turn the stage were too small. We determined that the control used to rotate the stage did not, as initially thought, measure degrees but instead used a relative measure in the form of a counter that was numbered from 1-100. This problem made it very difficult to adjust the counter for 10° increments of rotation. It was decided that the increments of rotation would be increased to a number that was easily divisible into 100. Images would be captured at intervals of 18°, thus requiring a total of 20 images to encircle the sample. Several series of images were captured at 22.5° increments above and below the equator (chosen to be the intersection of the X-Z plane with that of the two other planes) of the specimen. We also decided to incorporate a second QuickTime movie illustrating the aperture of the specimen through increasing stages of magnification (the hotspot).
After all the images were captured, they were burned onto CD-ROM for transport and further work. The images were imported into Adobe Photoshop 4.0 to be touched up and corrected. The backgrounds were removed and replaced with uniform black and using Apple’s QuickTime VR Authoring Studio software package a rough QuickTime VR object movie was assembled. At this point we determined that there were significant problems with the individual series of images obtained.
Each sequence on its own could be assembled into an object movie, but they could not be incorporated into a single movie. This was primarily due to the change in rotational increments from 10° to 18° (in X-Y plane). The most time consuming problem that had to be overcome was the wobble generated by the off-center mounting of the sample. This difficulty was eventually overcome by determining the point of rotation of each frame and then manually aligning each image as a layer in Photoshop. A total of 50 hours was spent correcting the images, with the majority spent centering the images. This time would have been unnecessary if the specimen had initially been mounted exactly in the center of the SEM stub.
Once the images were corrected, they were imported into Apple’s QuickTime VR Authoring Studio software package. This application allows for very quick and easy creation of both QuickTime panoramas and object movies. The basic QuickTime VR object movie was completed and a hotspot was created to link the second linear object movie (the zoomed aperture movie). Both the hotspot and the linear aperture object movie were very easy to create. The time devoted to assembling the final movie was about one hour.
The majority of the problems encountered during the experiment would have been avoided if we had taken more care in centering the specimen on the SEM stub. This would have reduced the time spent correcting the various images in Photoshop from 50 hours to less than 3 hours. From start to finish a complete QuickTime VR object movie could be completed in about 6 hours (including time spent capturing the images on the SEM).