Gibson ‘Feelie’ Object Files
I’ve created a OBJ and STL repository of printable / renderable versions of Gibson’s ‘feelies’ haptic dataset.
Go have a look and let me know if you download and/or use them.
Luxo Bookmark
Here’s something that might interest the Pixar fans/geeks out there-
I was cleaning up my lab today, and I found a copy of the RenderMan Companion (hey- I animated that cover image!) with a bookmark in it — more than a bookmark, it was a strip of Kodak 5247 with a sound track too!
Lo and behold- it is a piece of a print of Luxo Jr.
We used extra film / prints for all sorts of stuff. I remember going to a lunch at Colossal Pictures where they used scrap for cinematic silverware holders. Turns out, scrap prints were used all the time, especially when doing sound editing (so you could put place-holder stock where there was no sound, to get the timing right). Apparently, a lot of the old scrap was from pornos, so it turns out you had to be a little careful what you used for table place-settings.
Raspberry Pi Zero Clock Drift
So I built a device that uses the Raspberry Pi + a Hardware RTC.
I thought I’d look at the drift as reported by `hwclock -c` over the course of a day. So, here you go-
Colorful Visualization of the day
An in-development look at some age judgement data. Not sure what it is telling me yet, but I can’t let that stop me from looking at it?
Specularity and Shape from Line Drawings
Indications of shape via specularity in line drawings.
PDF here: Mazzarella2014
Volumes
Doing a little work on a fluid-flow project with some folks here- Wrote a volume renderer for it, basically re-visiting work I did in 1986 at The OSU CGRG, which eventually was applied to stuff we did with the volume renderer at Pixar.
Big difference here- this is about 1 page of code. The OSU and Pixar volume renderer was significantly more nasty.
On Labradoodles and Fried Chicken
(21 March – Updated with more images…) I recently saw a post on twitter from @teenybiscuit, subsequently picked up by NPR here . You’ve probably seen it, dogs, muffins and a whole bunch of other things that look similar. I decided to check out how well Wolfram’s (nee Mathematica) ImageIdentify[] function would do with it. Super quickly, here it is. Basically a 50% or so hit rate (for loose definition of ‘hit’) and not a terrible false-alarm rate (10-15%). It’s really good at guac and parrots. And it sometimes thinks Donald Trump is a virus.
I’ve put an image of the Mathematica output below. There’s a link to the WolframCloud™® website with the actual notebook below the image.
Click on the link to take yourself to the notebook’s web page and see the whole thing if you’d like. The cloud servers are being particularly slow/annoying, so you may have to wait a bit for it to load.
Source: whatarethey.nb – Wolfram Mathematica Online
My eyes!
OK, just sort of messing around while making some stimuli for this experiment. I’m as creeped out as you are. I promise.