Back in May, I blogged here about some experiments I’d carried out using animated GIFs to create illusions of perspective and motion from simultaneously-taken pairs of tintype portraits. Two of the pairs of “twin” tintypes I shared then had been taken from lenses mounted side by side, and these yielded convincing stereoscopic effects—they were basically accidental stereoviews. A third pair had been taken from lenses mounted one above the other, and those presented a different kind of perspectival illusion, giving an impression of the subjects rocking back and forth, towards and away from the viewer (similar to what I accomplished here with two versions of the Mona Lisa). As I wrote at the time, this third pair demonstrated “that animated GIFs aren’t merely a second-best substitute for ‘real’ stereoscopic viewing, and that in some cases they enable us to achieve an illusion of depth that stereoscopy proper wouldn’t permit. By the same token, there’s no reason we should be limited to juxtaposing pairs of images for presentation to corresponding pairs of eyes.” After all, tintype cameras were often designed to take many more than two pictures at once, and I illustrated the possibilities for animating larger numbers of images using a scan of an uncut set of four tintype portraits which I’d found posted online by the Minnesota Historical Society. The animation was jerky and hampered by the low resolution of the scan, but I wrote: “I’m eager to obtain some other sets of four or more ‘twin’ tintypes, whether cut apart or not, to see if I can improve on this result.” With that in mind, I was pleased to see the following turn up on eBay, advertised as “Photo Lot x 12 Tiny Miniature Gem Tintype Photos CA Late 1800’s Victorian Women”; and I was even more pleased when I was the only bidder and won the lot for $14.95 including shipping (which seems a fair market price):
In the top row, the tintypes at the left and the right—showing two women with their heads uncovered—appeared to be “twin” images, so I assumed they would lend themselves to the same kind of animation I’d demonstrated here back in May. The tintype in between is a singleton, although evidently taken during the same session as the other two, with the same two women now wearing hats. But what really excited me was the set of nine smaller tintypes underneath. The subject looked to be the same woman seen on the left in the tintypes in the top row, but clothed differently; and all nine images looked identical, suggesting that they had all been taken at the same time, but also from different lenses in slightly different places, since no two tintypes are ever exactly the same.
ANTICIPATION AND SPECULATION
As I write these words, the lot is still in the mail on its way to me, so I don’t yet know how my effort to animate the various images will pan out. But I can still record some of my thoughts and expectations. There are nine seemingly “identical” images, so they could be a complete set of portraits taken at one time using a nine-lens camera (something like this), or they might instead be the remnants of a set of images taken with an array of twelve or more lenses (with some missing), or perhaps even a mix of images taken at multiple moments with no immediately obvious change in pose.
Ideally, if the images turn out to comprise a three-by-three grid, with rows we can label ABC / DEF / GHI, then it should be possible to contrive a variety of virtual motions: a Z shape (ABCFEDGHI); an N shape (GDABEHIFC); continuous rotation around a square (ABCFIHGDABCF….); diagonals (AEI or CEG); and so on. I don’t know quite what I should expect the results to look like, but they might resemble animations made from images taken with the four-lens Nimslo camera, as seen here; or they might even hint at the “frozen time effect” pioneered by inventor Dayton Taylor and his company Digital Air.
ANALYSIS AND RESULTS
The images reached me safely, and even included an extra freebie—a photographic print on paper that seems to show the same woman at a later point in life, wearing a veil decorated rather sparsely with diamond shapes:
The pair of “identical” tintypes showing the two women with hats off produced a good stereoscopic effect, much like the examples I blogged about earlier:
On closer scrutiny, these gems turned out to belong to two different poses, one with the sitter’s head tilted up (A, B, G, H, I) and one with her head tilted down (C, D, E, F). It took me a while to figure that out, since the difference in the tilt of the head happens to look deceptively similar to a difference in camera perspective (except that the body doesn’t move with it, and there are other discrepancies besides). My technique was first to rotate the images into line with each other, choosing the sitter’s collar button and a point on her sleeve as static reference points, and then to flip back and forth between pairs of images to work out the perspectival relationships between them. The up-tilted set can be organized into a two-by-three grid, but the down-tilted set is more complicated—C, D, and F fall neatly into a two-by-two grid, but E is an anomaly: its angular perspective appears identical to F, but at a vantage point a little further away, such that a broader swath of the sitter’s hair is visible in back behind her neck. I’m not sure how to account for this phenomenon based on such knowledge as I have of multiple-lens tintype cameras, but there you have it.
Based on the analysis laid out above, it’s a pretty straightforward matter to create animated loops in which the viewer experiences an illusion of movement around a circle of perspectives. Here’s HAIBG (with a tweened frame inserted between H and A to compensate for the missing sixth image in the grid):
Both examples display a bobble-head-like effect which I imagine groups of images such as these will prove capable of yielding more generally. And that’s as far as we can go here while remaining true to original photographed reality. However, if we want to be a bit more creative, we can also merge the two different poses (FDC+HAIBG, sans tweening) so that the motion of the sitter’s head appears smoothly continuous between them, although her body doesn’t quite follow and so has an unfortunately jerky appearance:
Now, I’m no expert in historical fashions, but the clothing style here looks circa 1900 to me (would anyone care to venture a better-informed estimate?). That said, the multiple-lens tintype camera was already in widespread use during the 1860s, so there’s no reason why we shouldn’t be able to contrive similar animations of subjects from the era of the American Civil War.
In fact, I think I can already furnish one such example, thanks to a scan of an uncut set of four tintypes just posted online by the Indiana University Digital Library Program as part of the Wylie House Museum Image Collection. The subject of the portrait is Theophilus A. Wylie (1810-1895), an early professor at Indiana University who, among other things, is “rumored to have installed the first telephone in the state of Indiana” (this according to Bloomingpedia). Judging from the medium and from Wylie’s apparent age, I’d tentatively date the portrait to the 1860s. The top two images appear to have been taken separately from the bottom two images, with a slight shift in posture between. Either pair of images, top or bottom, would work as a stereoview. But when we combine all four images together into a single animation, the results are enough to make your head spin:
Source images courtesy Wylie House Museum, Indiana University, Bloomington, Indiana. The hunt for further promising candidates goes on.