Balkentrol:

Lords of Acid

Balkentrol is a simple 3d racing game implemented entirely in hardware (no processor or computer of any sort).

It consists of a ball bouncing along a race track. The track consists of levels at two heights. The lower level is texture mapped. The second level is transparent. In addition, the game has two animated backgrounds, and portions of the lower level move. The player can accelerate in the four directions, and jump. The player can "boost" before a jump to jump higher. In addition, I implement a very simple lighting model, where the track darkens towards the back, and the ball has a shadow. For added headache value, I also support LCD shutter glasses for a more 3d effect. All of the above are implemented in the simplest/crudest way I was aware of. The ball cannot turn on the race track (primarily due to number of gate limitations); just accelerate in the four directions and jump. The textures have no interpolation. The systems has no antialiasing. The system runs at VGA resolution (640x480x24bppx60fps)

The game is hippie/acid-themed, although (to the best of my knowledge) no one involved in it has, at any point, used illegal drugs. This is primarily because colors generated through bit manipulations tend to look very fluorescent, so it was an easy default theme. I wanted a consistent theme, so I ran with it.

Hardware used:

Thanks to a generous anonymous donor for donating a UP1 10k20 board, which I used to develop the initial version, before I was loaned a larger 10k70 board (which I had to return; I am investigating possibilities for acquiring one permanently, but pending that, only the smaller 10k20 version of the game is functional). Also, thanks to Intersil for donating the high-speed DAC chips, and assorted MIT supply rooms for donating the remaining parts.

The game was a group project, although the portions of the project were entirely modular, and as such, we met maybe three times for about an hour each before final integration (at which point, my code and circuitry were fully finished, and required no changes). I was responsible for the 3d render engine and game play (including everything on this web page, except for portions of the lab report). My partner, Kristin Carr, developed a sound system (where the music changed along with the game play in fairly sophisticated ways, with multiple tracks of sound, filters, etc.). The two interfaced through a one-bit connection following the standards of a parallel interface. In addition, there was an original sound score composed by Daniel Roy. Only the portions of the project I developed are documented on this web page (outside of this paragraph).

Although I began developing this project independently of coursework (primarily to learn more about digital design), I started working on it the semester I was taking MIT's digital systems lab, and around the time of the final project (shortly after I had finished the last laboratory). As a result, I decided to use it for the final project as well. It was nominated for top laboratory project at MIT, although, as of this writing, it remains to be seen whether it will win.

Resources:

If you are trying to reimplement this, interfacing to the 10k70 FPGA is easy; the hardware around it must supply 8 input pins; 4 from arrow controls, 4 for buttons. One is active-low, the other active-high. The output has 2 pins for the 3d glasses (optional; the game has non-glass mode as well), which switch every other frame. It has 24 pins, in sets of 8 each, corresponding to the color output. It has two others, corresponding to hsync and vsync (I believe the outputs are inverted). There was a DAC clock pin in the .acf file, but it is not used. For details, see the laboratory writeup above.

The 10k20 runs on the UP1 board. It has the same inputs as the 10k70, the same video DAC output, but lacks the stereoscopic glass output. It has significantly inferior graphics, and much less functionality (no boost, no transparency, no second level, no shadow, no stereoscopic glases, no screenshot mode, no communications port to broadcast game status to other boards, and so on). I have several screenshots of the 10k20 version:

Bear in mind that the 10k20 version does not have a stop-frame/screenshot mode, and as a result, there is a significant amount of blur from the background rotating. The camera shutter time was set to as short as possible to minimize the effects, but this made the images very dim, and they had to be brightened in xv to reflext the actual game. As a result, they look fairly blurry, and the colors are a bit off.

Since this page is getting some traffic, I figured I'd mention I'm looking for a job. Although this was a primarily digital project, I'm actually a good deal better at analog than digital.

Copyright (c) 2003. Piotr F. Mitros.