Archive for June, 2011


Making it Move pt. 2: Banana with Detail

Yesterday I posted about my early animations on the Atari 8-bit computers using MovieMaker, and linked to video of the same, but after watching them I thought I’d post notes about the different segments.


Part of a longer animation which featured the Bloom County “Banana Jr.” computer walking in and displaying a countdown. The walk was clumsy, but I loved the 3.5″ floppy shooting in, and the little “hop” the computer makes on contact.

The disk flies in perspective

Click to start animation here (opens new window)

The “zoom” in on the Banana Jr. at the start of the countdown was a feature built into MovieMaker and let you zoom in on the center part of the animation by taking that ¼ of the screen and doubling the size of its pixels, reducing the apparent resolution from 160×96 to 80×48!

ZOOM made the image bigger but at the cost of lowering the resolution


Artech was what I was going to call my imagined animation & graphics biz. I did two different animations of the logo I’d created. The first—appearing at the end of the video—was simply sliding around of the pieces. MovieMaker allowed you to animate up to six items (“Actors”) on a given frame, and since the logo had six letters, I opted to animate four sets of letters. Had I been able to animate seven objects I’d otherwise have animated all the letters in separately.

First version of the logo featuring 4 overlapping elements (Actors)

Click to start animation here (opens new window)

Sometime later (God I wish files included time and date stamps in those days!) I took another crack at the logo, this time spinning the two halves in, so I created 16 frames of each half of the word, rotating (30 degrees per increment), so you’d see them head on and as they turned you’d see the edges.  This was partially effective, but it didn’t have any dimension, so to create the illusion that there was some kind of light source, I changed the palette with each frame, so that the faces and edges of the letters would brighten and darken as they spun, which is a neat trick for making it look like you have more than the three colors (the fourth being the background) per frame actually possible!

Second version of the logo with multiple animation drawings and color cycling

Click to start the animation here (opens new window)


Making It Move

Some months ago I was posting about my earliest computer graphics work, starting on the Apple ][+ and moving onto the Atari 8-bit systems.

Memory is probably failing me on this account, but what I recall is that I had developed a growing interest in animation in the years leading up to and through 1984 and one of the reasons I bought my first computer was in order to do animation in a way that wasn’t going to be as monstrously expensive as traditional film methods. I bought an Atari 800XL and immediately set about getting tools. I don’t recall the order in which I purchased them, but amongst my earliest purchases were an Atari Touch Tablet, and Atari Light Pen, and, most importantly, a program called MovieMaker (no relation the Windows program of the same name).

As I’ve discussed in recent posts, the Atari computer’s graphics were fairly limited (albeit in some ways superior to other low-end computers of the era). Most paint software worked in a mode called 7-Plus (or Graphics 7½), which displayed 160×192 pixels in four colors chosen from a palette of 128, which was arguably the best multicolor graphics mode the machine could do. However, MovieMaker utilized Mode 7 graphics, which took half as much screen memory, but at a cost of half the vertical resolution, so animations and the artwork used to build them were at a resolution of 160×96 pixels.

There were other limitations.

  1. One screen for the background
  2. One screen that had to contain all the graphic elements you would animate, so you could only animate as much stuff as could be crammed onto one 160×96 screen
  3. Four colors at once, but you could change the palette frame by frame if you so desired.
  4. A maximum of 300 frames per animation file
  5. Three different sets of nine sounds each, but you could only use one sound set in a given animation

Here’s a set of animations created with these tools, most by me and two by Jennifer Voigt (the island dancers and Easter Eggs). (The native computer-generated sound were retained when these animations were assembled to video, but in a few clips recorded music and voices were substituted, as you’ll hear.)

A lot of these animations were created after I purchased the more powerful 16-bit Atari ST computer, but no real animation software was available to me for that system for the first 15 months I had it, hence I was forced to stick with MovieMaker and its limitations for quite a while.

A number of MovieMaker animations (and others) were assembled on video as a demonstration for potential business uses, but by the time I finished putting that video together I knew it was already obsolete. Still, working within MovieMaker’s rather stringent restrictions again taught me how to do a lot with very very little, something that continues to serve me well to this day.


Rigel Seven Up

My first computer was purchased in December 1984. It was an Atari 800XL with 64K of RAM.

My second computer was bought some months later (maybe May 1985). It was an Atari 130XE, which was basically the same machine as the 800XL but with better video output and 128K of RAM. The extra RAM was useless for programs, but could be configured as a “RAM Disk” which made some tasks easier.

Some the of the earliest software I purchased for these computers were art and animation tools. These were easy to get because by that time the Atari 8-bit computer line was going on five years old.

The third computer was purchased in August 1985. It was an Atari 520ST, a low-cost 16-bit marvel with a staggering 512K of RAM built-in. I was an early adopter. Hell, I got one the first week they were available! I recall it cost me $937.50 for the computer, one single sided 3.5″ floppy drive, and and SC1224 color monitor.

The trouble with buying a brand new computer model, especially one with a new operating system, is that upon release there’s typically scant software to accompany it. Such was the case with my brand new ST. I had a computer many times more powerful than my other ones, but I had so little software that it was barely useful for anything much for the first few months I owned it.

Fortunately, one of the few programs that shipped with the ST was an early version of a paint program called NeoChrome, so at least I could work on improving my computer graphics skills as I waited for something approaching an animation tool to arrive.

I posted one of my earliest ST graphics yesterday and the day before, but the most elaborate image I drew in those early days of the ST was the following recreation of a scene from the first Star Trek pilot: an abandoned fortress on the planet Rigel VII.

What circle drawing tool?

What I remember about drawing this image was how hard it was to do.

  1. It was an attempt to recreate digitally a very colorful matte painting using only 16 colors.
  2. My source was a tiny screenshot on the back of a book, less than 2″ wide.
  3. The NeoChrome program was v.0.5, which means not even a full-featured paint program, and amongst those missing features were any tools for drawing arcs, ellipses or circles!

So, how to draw a giant moon looming on the horizon sans a circle drawing tool? I knew eyeballing it would be an exercise in futility, so what I did was use a protractor to draw a circle of the right size on paper. I then cut out the circle so I had a stencil, stuck it onto the monitor, and—one pixel at a time—drew in one-quarter of the outline of the circle. I then used the cut and paste tool to duplicate and flip this to complete the circle. Insane, but it worked!

Below is a screen-grab of the original matte shot. The book I was working from had the colors way oversaturated, so I got that wrong. Still I think I didn’t do badly for 16 colors out of a palette of 512 possible. Heck, the image was good enough that when I used it to illustrate an article on graphics I wrote for ANALOG Computing Magazine they used it on the cover of their ST-Log Magazine insert for issue number—you guessed it—seven.

The original matte shot I was aping.


A Pixel Is Not A Little Square

(With apologies to Dr. Alvy Ray Smith and his technical memo of the same name.)

Yesterday I posted about some of my early graphics work and included a couple of images illustrating how differently graphics would be drawn with relatively modest improvements of computer display capabilities. I noted that the images displayed weren’t necessarily true representations of how the original looked because of changes in display technology. While that’s true, it’s an incomplete explanation, as not only does the display type affect the appearance of the image, but there’s the issue of aspect ratio of the pixels that make up the image.

Because of how CRT TVs and monitors like them worked, pixels were never truly square. On the Atari 8-bit computers I was first working on each pixel of the display in the mode used (160×192) was roughly 1.8×1, and later on the 16-bit Atari ST (low-resolution 320×200 pixels) each pixel was roughly .9×1 As such, any graphic drawn to look correct on such a display ends up looking slightly distorted on modern system with truly square pixels. To better illustrate what these graphics looked like back in their day I’ve rescaled them to correct for the aspect ratio and blurred them a bit to give a better sense of how they displayed.


Atari ST image corrected for aspect ratio


Twice something is four times something

Some months back I posted about doing some of my earliest computer graphics work.

My first computers were Atari 8-bit machines, and the most widely supported graphics mode for “art” purposes was had a resolution of 160×192 pixels, and in that mode I could have precisely FOUR colors on screen picked from a total palette of 128 colors. Below is a very early image I drew in that mode, probably created in the first few months of 1985.

Early pixels. 160x192 in 4 colors.

Now, the things about computer tech is that it’s always changing, and by the time I got my first Atari its days were numbered. Eight months later Atari released a 16-bit computer with much better graphics. Instead of 160×192 pixels with 4 colors from a potential palette of 128, this new machine allowed for 320×200 pixels with 16 colors from a potential palette of 512 (eight levels each of red, green, blue). The older computer used two bits per pixel, meaning there were four possible combinations of zeroes and ones (00 01 10 11), thus any pixel could be one of four colors. The newer computer used 4 bit per pixel, meaning there were 16 possible combinations of zeroes and ones (0000 0001 0010 0011 0100 0101 0110 0111 1000 1001 1010 1011 1100 1101 1110 1111) ergo sixteen possible values per pixel. In the case of binary, doubling the number of bits squares the possible combinations. So, 2 bits = 4, but 4 bits =16. Twice something is four times something.

Anyway, one of the first things I did when I got a simple paint program for it was to see how much of an improvement the resolution and more colors on screen would permit. I did this by recreating an early image from the older Atari on the new one.

What I could do with twice as many pixels and four times as many colors.

By modern standards it doesn’t look like much. But still, just having four times as many colors not only allowed me to add a background, but also to have enough colors to make the energy bolt animated via color cycling. The animated GIF here doesn’t cycle these colors anywhere near as fast as the Atari would, so the strobing effect the original had is somewhat lost, but you get the idea. Yes, it looks gaudy as Hell, but then these modern flatpanel monitors don’t have the same softening effect the older CRT monitors did, so every pixel stands out more sharply than it did back in the day.

I’m still amazed that these two images were created within nine months of each other.

But I still had a lot to learn.


Polaris Process

I’ve been spending most of my working hours in recent weeks plugging away at editing together the rough assembly of STARSHIP POLARIS and it’s already been an interesting learning process. As of Friday the 24th I’ve actually completed a slightly polished rough assembly (how’s that for a contradictory term?) of the first 16 scenes of Polaris comprising roughly the first eight minutes of the show.

The process of getting to this point has been an education.

Oh, sure, I’ve been doing video editing on my own projects for a few years now, and also on the final section of another film that was basically abandoned by its producer, but none of those projects individually has been very big. In the case of my own work it’s because they’ve all been less than 10 minutes long, and in the other case because it’s only the last 20% of a larger show.  In the case of POLARIS. I essentially got the raw footage (already transcoded) and the separate raw audio and had to build the entire project pretty much from scratch. This has meant doing tons of file renaming to make sure all the footage and audio files maintain a consistent naming convention and  then running the video and audio through a program (DualEyes) that figures out how the second sound lines up with the sound recorded by the camera and replaces the latter with the former, and even how to fix audio files with a bad channel and manually syncing the sound to picture on a few shots where no automated solution worked.

Dr. Valerie Young reacts with bemusement to Captain Paul Fredericks's skepticism.

Even after all that there’s still the process of importing all the footage into Final Cut Pro and logging it. This means going one file at a time and tagging it for the Scene number (say 72A-A as seen on the slate), the take number, who’s in it, what the angle is, etc. This for for literally hundreds of files. To complicate matters, in some cases we did rolling takes where a file slated as take 1 might actual include five runthroughs of the same bit of action or dialog.

As part of this process I had to work out a system that would let me find things easily when editing. By getting all the scene numbers logged I can sort clips in the Browser by the Scene column to see all the clips in more or less the order they happen in the story. But sometimes I’m looking for shots of a given character from a different scene (say I want to steal a reaction from a shot not part of the current scene, or we covered a bunch of different scenes in one setup), so I sort by the Description column, which now contains the name or initials of all the characters in a given scene, so all the files containing that character get listed together. Additional comments columns are used to note which takes are shot MOS, have only camera sound, or contain defects or interesting accidents. Sorting by these columns further makes it possible to find what I’m after quickly.

It’s a virtual certainly that there are even better ways to do all of this, and it’s equally certain I’ll slap myself in the forehead once I realize my “easy” way may be a “hard” way.

But that’s the fun of learning, ain’t it?

And, no, I don’t mean the self abuse.


The Monty Lisa

Some more of my old computer artwork. In this instance, I had this scan of the Mona Lisa which I reduced to 14 colors out of the 16 total the Atari ST could normally display in this graphics mode, leaving two colors with which I could work to make a sort of Terry Gilliam take on the material.

Scanned image of the Mona Lisa

Monty Lisa, by Maurice Molyneaux, 1988

As with the Bugatti image yesterday, this was one of several images I created for Epyx Inc. as box-art for the Art & Film Director product. The scanned image was altered using Art Director on an Atari ST computer at 320×200 resolution in 16 colors out of a palette of 512 possible.