My goodness. Never thought I'd see this again. We did that at the Dallas, Texas, acmecartoon company. It was a "sister" shop to Sundance Productions, owned and operated by Rush Beasley, who later formed Rushworks. The storyboard called for the video inset as you see it here, but there was no way to put a full CRT in the robot costume along with the actor. We had to figure out how to track the blank cardboard "screen" in his chest, and warp the flat video onto it with the proper perspective. A big problem was making the corners of the warped video smoothly track the actor's motion. These days, motion-capture and real-time perspective mapping with a GPU would make this pretty easy. Back in the early '80s, none of that existed. We had a PDP-11/60 computer and a Fortran compiler. I wrote the code to do the warping and smooth the animation. Our artists laboriously picked the corners of the screen, frame-by-frame, with a tablet (not the ones you know today, but a big flat touchable surface we called a "digitizer"). They had to paint out the warped video, again frame-by-frame, whenever anything passed between the screen and the camera. It worked pretty well, didn't it? Rush marketed the process for a while under the name "Motion Track-Matte." Yeah, those were the days...
First thing I thought when I started watching this clip was, hang on, they've little display is overlayed, can see it wobbling a little, of course they couldn't fit a CRT in there, how the F did they do that? you did it in Fortran on a PDP-11???? jesus
Taped this ad along with Superman: The Movie when I was a kid in the 80s in Australia. Watched it a billion times. Lol you never know what influence you're going to have.
@@danielsanichiban Yup. The 11/60 had 96k of ram in it. (Yes, I said "k," not "M," and no one knew what a gigasomething was in those days.) It couldn't hold an entire video frame, so we used an external frame buffer. The image mapping was done with a two-pass method published in 1980 by Catmull and Smith (alvyray.com/Papers/CG/2pass80.pdf). It was perfect for our limited machine, because it operated on the image one row (or column) at a time. I believe the authors really intended it for hardware implementation, but it was a good fit for our little PDP-11 and its 16-bit segmented address memory model.
wow...found a badge with prime on it in a box of old junk, googled it and this came up. Thanks for putting it up! sure clears up what the hell it was, poor mans c3po lol.
My goodness. Never thought I'd see this again. We did that at the Dallas, Texas, acmecartoon company. It was a "sister" shop to Sundance Productions, owned and operated by Rush Beasley, who later formed Rushworks. The storyboard called for the video inset as you see it here, but there was no way to put a full CRT in the robot costume along with the actor. We had to figure out how to track the blank cardboard "screen" in his chest, and warp the flat video onto it with the proper perspective. A big problem was making the corners of the warped video smoothly track the actor's motion. These days, motion-capture and real-time perspective mapping with a GPU would make this pretty easy. Back in the early '80s, none of that existed. We had a PDP-11/60 computer and a Fortran compiler. I wrote the code to do the warping and smooth the animation. Our artists laboriously picked the corners of the screen, frame-by-frame, with a tablet (not the ones you know today, but a big flat touchable surface we called a "digitizer"). They had to paint out the warped video, again frame-by-frame, whenever anything passed between the screen and the camera.
It worked pretty well, didn't it? Rush marketed the process for a while under the name "Motion Track-Matte." Yeah, those were the days...
First thing I thought when I started watching this clip was, hang on, they've little display is overlayed, can see it wobbling a little, of course they couldn't fit a CRT in there, how the F did they do that? you did it in Fortran on a PDP-11???? jesus
Taped this ad along with Superman: The Movie when I was a kid in the 80s in Australia. Watched it a billion times. Lol you never know what influence you're going to have.
@@danielsanichiban Yup. The 11/60 had 96k of ram in it. (Yes, I said "k," not "M," and no one knew what a gigasomething was in those days.) It couldn't hold an entire video frame, so we used an external frame buffer. The image mapping was done with a two-pass method published in 1980 by Catmull and Smith (alvyray.com/Papers/CG/2pass80.pdf). It was perfect for our limited machine, because it operated on the image one row (or column) at a time. I believe the authors really intended it for hardware implementation, but it was a good fit for our little PDP-11 and its 16-bit segmented address memory model.
@@froggyundul Cool! Looks like we have done some similar work since then. Nice to know it had some life beyond the commercial break.
wow...found a badge with prime on it in a box of old junk, googled it and this came up. Thanks for putting it up! sure clears up what the hell it was, poor mans c3po lol.