Spangler Studios Blog

Some animation in the great outdoors.

One day last week, I spent more than a few hours trying to hook the little video feed camera up so I could see through the viewfinder of my Canon DSLR. This Unibrain Fire-i Board Camera is what I had been used to working with at Handcranked Productions, so I did some refresher tests at their studio to see which version would work best for me.

Unibrain offers 3 different Board Cameras (unlike the regular webcam they sell, these come with the camera board separated from the rest and the lenses and cover are optional). There is black and white, color and raw color. The problem with the color cam is that the images are not very crisp. This is because, from what I’ve managed to glean from the Unibrain informational downloads, the video signal can only provide one piece of information per pixel, which is perfect for monochrome cameras. But for a color image, three pieces of information are needed (YUV). A quote from Wikipedia: “Y stands for the luma component (the brightness) and U and V are the chrominance (color) components.” So groups of neighboring pixels carrying different pieces of color information must be merged to get the full color information, resulting in reduced information overall. Anyway, here are two comparison images. The monochrome was taken with an 8mm lens, the color with a 12mm lens (which gives a wider view, but cuts off the image a little at the edges!).

This obviously wasn’t the most scientific test. I should have used the same lens for both cameras to have a perfect comparison, but I think you can see how the color image is noisier than the monochrome.

So I ordered a monochrome camera, as well as an 8mm and 12mm lens. The 8mm lens has an IR filter, but the 12mm does not. I emailed a very helpful fellow at Unibrain support, and he informed me that “an IR filter allows only visible light and not close IR. With a lens without an IR filter you would want to avoid IR, Halogen and sunlight as these could damage the CCD chip.” So I had already purchased a small clip-on halogen light to try using on set. I wonder if even this little amount of halogen light would be damaging to the sensor?

The next challenge was connecting the little camera to the DSLR, a common dilemma I believe. I actually managed to get it in place with some bits of hardware laying around my studio (an amazing feat). At one point I became very frustrated and grumpily told my roommate that I felt like MacGuyver on a bad day, as I was attempting to shave down toothpicks to fit alongside the metal bar I had stuck into the camera’s flash hot-foot. I got the basic idea thanks to Nick Hilligoss on stopmotionanimation.com. Here’s a link to the discussion (with pictures!).

Here’s my version (until I get a little hotfoot adapter in the mail and attach the rig to that, which is probably safer for the camera):

I should note that the Board Camera I ordered came with an opaque gray case, with no attachment point on the back, as shown in these photos. This translucent case with attachment was given to me and was taken from a regular color Unibrain Webcam. Maybe Unibrain would be willing to substitute, or maybe I missed the option for this on the sales page. As you can see, this extra little piece of plastic was very helpful for this rigging situation, and even provides some rotational adjustment for the camera, which so far seems like a good thing, but since it can’t be tightened and secured, this could be bad if the camera were ever bumped during shooting.

Oh, and while I’m talking about things attached to the camera, I should point out that the lens on this camera is a manual 52mm Nikon lens, connected by a Novoflex Lens Mount Adapter, specifically for connecting Nikon lenses to a Canon EOS body. This is a method a friend of mine has been using successfully, and so far it has worked perfectly on this camera too.

This is a very rough animation test, done over a month ago, testing character interaction as well as a somewhat unsuccessful method of rigging that I like to call “the umbilical cord.” If you look closely, you can see that both characters are attached to a long wire that is covered with tissue paper and glue and painted dark colors to blend in with the background. This method worked alright for a quick test, but overall it was difficult to position the characters because of extreme “bounce-back” from the long wire, and the cord is quite obvious when it passes in front of a light color (or in front of the character itself!). Since this was a test, I didn’t make it so that the wire could be removed and repositioned when necessary. I think a better method would involve using straight metal rods with joints, maybe with bendable wire at the end closest to the puppet?

I should note that I’m attempting to rig the puppets externally, in a way that blends well enough with the look of the world that the rigs will not need to be removed digitally and will instead be accepted as part of the animation style. Even though this test proved a bit difficult, I enjoyed the freedom of not worrying about drilling holes to tie down the feet after each move, and also not thinking about rig removal.

Last week, the shrewmole film (tentatively titled “The Duvetyne Shrewmole”) was awarded a production grant through the LEF Moving Images Fund for New England artists. I’m very excited and looking forward to taking some time out specifically for work on this film. I will continue to post updates and share what I learn along the way.

Below is an image gallery of the test set I put together. Everything is made of pretty economical materials - aluminum foil, wire, cardboard, brown paper. The set elements are mostly on movable flat layers, except for the bigger flowers and the ground surface.
(click on the big images to advance)