Stereoviews were the first commercially mass-produced media designed to be accessed with a special piece of equipment—namely, a stereoscope or stereoviewer. In that important sense, they anticipated the later media of cinema (motion pictures) and phonography (sound recordings).
But stereoviews are at a significant disadvantage when it comes to twenty-first-century accessibility. Today it’s fairly easy to present historical examples of cinema and phonography on the Internet. People can access early sound recordings and films using the same familiar applications they use to watch other videos and listen to other audio, even if translating the content into modern formats (say, from wax cylinder to mp3) can be technically daunting.
Stereoviews are another matter. There’s no commonly available means of displaying and viewing images in 3D through the web, if by “commonly available” I mean that I can reasonably assume you have ready access to it if you’re reading this blog post. Therefore, anyone wanting to mount historical stereoviews on the web for general public viewing has been forced to choose among several imperfect solutions, as follows:
1. Display a scan of the pair of images side by side, as they appear on the original card, and invite people either to look at them through a stereoscopic viewer (such as the “Pixie”) or to learn freeviewing, a technique for viewing the image pairs in stereo without using special equipment, as described here: “Focus your eyes to infinity. Hold the card [or the computer screen] 12-18 in. away and raise it into the center of your field of vision maintaining infinity focus with your eyes. You may see 4 images at first or only 2. You may need to focus the image by moving the card closer or further away. You want to see 3 images with the center image appearing 3 dimensional.” This is the simplest way to present historical stereoviews online, and the most consistent with the original philosophy of stereoscopy, but it also demands a lot from people who want to access the images in 3D. Not only is freeviewing a skill that takes practice and effort to master, but it has other drawbacks besides: Wikipedia says that “the physiological depth cues resulting from the unnatural combination of eye convergence and focus required will be unlike those experienced when actually viewing the scene in reality, making an accurate simulation of the natural viewing experience impossible and tending to cause eye strain and fatigue.” Casual visitors to a website probably won’t take the trouble to learn freeviewing, and they certainly won’t order a stereoscopic viewer and wait around for it to arrive.
2. Another alternative is to convert the stereoview into an anaglyph—a single composite image filtered for viewing through 3D glasses with (typically) red and cyan lenses. This approach still requires a special piece of equipment, and while it’s easier to obtain 3D glasses than a stereoviewer, casual site visitors aren’t likely to have a pair readily on hand. If you don’t have 3D glasses, you’re out of luck: an anaglyph doesn’t permit freeviewing, and instead of two individually acceptable 2D images, you’re faced with a single, unpleasantly blurred image.
3. A third option is the wiggle GIF, an animated GIF that alternates rapidly between two frames corresponding to the two images that comprise the stereoview. This approach is popular that the New York Public Library has created a web-based software tool, the Stereogranimator, that lets visitors create their own wiggle GIFs dynamically from stereoviews digitized from the library’s collections. The great advantage of this approach is that it lets viewers experience a kind of 3D effect without using special equipment or acquiring special skills. The drawback is that it’s sensorily jarring. The comments section below an article about “3D ‘Motion Pictures’ From the Civil War” —meaning wiggle GIFs made from stereoviews—contains various remarks about “the annoying ‘animated’ viewing method” with its “dizzying effect,” said to produce “a kind of odd visual dis-equilibrium with all the flashing and wigglin’ going on”; “my eyes are physically freaking out from the strain,” writes one visitor.
Each of the three approaches described above has advantages but also shortcomings. Bearing that in mind, let’s consider a fourth alternative, the tweened animated GIF:
Instead of alternating rapidly here between the two frames, I’ve tweened them in Photoshop—that is, I’ve treated the original pair of images as key frames and generated sequences of intermediary frames between them so that the transition is gradual rather than abrupt. In the timeline below, frames 1 and 13 correspond to one of the original images, and frame 7 corresponds to the other:
This allows the “wiggle speed” to be reduced to a rate that’s less sensorily jarring than in the case of an ordinary wiggle GIF; a complete cycle in this case takes 1.2 seconds instead of 0.2 seconds. It also produces an illusion of continuous motion so that objects really appear to be shifting slowly in three-dimensional space: in the above example, watch the girl, card, and desk. Objects that lie further forward or back may seem to jump a bit, as with the chairs and paintings in the above example, but even this effect resembles the familiar transitions between viewpoints in lenticular prints, and so doesn’t strike me as a fatal flaw. The first stereoview I converted into a tweened animated GIF—an alligator photographed by Engle and Furlong of Fernandina, Florida, circa 1870—does a nice job of illustrating how tweening in Photoshop affects different parts of objects depending on their distances from the reference point we use to align the two images (in this case, the alligator’s hind leg):
Tweening hasn’t been applied at all extensively to historical stereoviews, as far as I can tell. In fact, the only prior instance I’ve turned up online is a YouTube video of Japanese stereoviews created by Colin Davidson of Start3D, a startup that invited people to upload 3D image pairs to a website where patented software would convert them, for a fee, into tweened animations. Davidson’s 3D renderings in the video are considerably smoother than the ones I can create myself using the tween function in Photoshop, mainly because they interpolate intermediate positions for far-shifting background or foreground objects (with occasional glitches, most notably with the spokes of the left wheel on the cart at 1:26). I’d like to have seen what Start3D could do with the head and tail of my alligator! But after a burst of publicity in 2010, the operation now seems to be dormant—I can’t get the website to load. The Start3D software would clearly be superior to Photoshop for this kind of work, judging from Davidson’s video, but it seems to be beyond reach at the moment.
And I haven’t been able to find any previous example anywhere on the web of an actual tweened animated GIF created from a historical stereoview. As noted earlier, Davidson presents the results of his experiment via YouTube as a video, perhaps because he also plays around with focus—a cool effect in its own right. But the tweened animated GIF has substantial advantages over the more common alternatives described above, and it’s only slightly more time-consuming to create one than it is to create an ordinary wiggle GIF. This is one technique I’ll plan on using when I have stereoviews I want to display and blog about here.
Postscript: Since writing this blog post, I’ve hit upon yet another technique for animating stereoviews using image-morphing software, which is a lot more time-consuming but yields better results: see here.
Postscript #2 (October 8, 2015): I finally bought some red-cyan glasses and found to my chagrin that the sample anaglyph I’d posted here had the colors assigned to the wrong “halves” of the image, garbling the visual results. I’ve substituted a corrected anaglyph above, but I’ve also reposted the original, incorrect version below for comparison. This blog post had already received 493 hits by the time I discovered my mistake, and nobody had ever pointed it out to me. Are my readers just really polite, or are there that few people out there who actually use red-cyan glasses for 3D viewing online?