The Uncanny and Impressionist in 123D Catch
I’ve written a few times now about the photo-modeling platform 123D Catch (formerly Photofly). While I haven’t posted anything lately about it, I have continued to expirment with the technology with some interesting outcomes.
The more I’ve used Catch, the more the technology has fallen into the uncanny category for me. It started with my attempts to construct a model of myself:
Based on the comments on YouTube from the second of these videos, I’m hardly alone in feeling like this is somehow repulsive – but there is an attraction here for me as well. Seeing myself so obviously mangled is strange – especially when I think back to the feeling of being photographed. Voluntarily trying to keep absolutely still – to become an object, fighting my human urge to settle or move – was a bit foreign, and yet it was my ultimate failure to remain motionless that caused such a foreign outcome.
Moving further (and beyond the realm of Catch, specifically) it was exhilarating to see that 3d model made physical when it came out of the 3d printer where I work:
Beyond the process of capturing myself, I’ve had other strangely satisfying uncanny outcomes – like what happened when I tried to model the Statue of Liberty:
This model was generated from still frames extracted from a Creative Commons-licensed video of the real Statue of Liberty – but when I fed the stills through Catch, they turned the statue inside out.
Yet another bizarre outcome came from my first attempt to model the west facade of Ohio Stadium:
The cause of this mishap is more easily understood; I inadvertently included some images from a second photo session. That second session was focused on a section of land on the bank of the Olentangly River – but portions of Ohio Stadium were visible in the background, obscured by trees. The analysis software attempted to reconcile the two image sets, but was stymied by the occluded view of the stadium through the trees – with somewhat amusing results.
To be fair, a second attempt at modeling Ohio Stadium met with better – but still odd – results:
A second conclusion that has become more apparent to me the more I use this tool is that Catch is incredibly Impressionist in its character. Not only are the best results generated when the model is stationary, even such subtleties as the quality of light and transparency enough to stymie the package. I have tried on a few occasions to go back and re-photograph some locations, and the outcome usually bears the scars of two objects having been grafted together.
In a way, I’m reminded of a passage from the Walter Benjamin story, “Funes the Memorious.” In it the character of Funes is described as irritated or bothered that the cat which he sees at one moment from the side is named the same as the cat he sees the next moment from behind. So fragile is his mental construct of the world that even slight variations are problematic – his ability to extrapolate from isolated impressions to a general comprehension of a subject is extremely minimal.
More to come
If this sounds like I’ve given up on the technology, allow me to clarify: I love this software. If anything, the fact that many models have these quirks or failures makes the process that much more interesting – and provides an opportunity to be intentionally disruptive.