Real-time animation and music using a 1985 home computer.
By Free Tutorials on Oct 10, 2008 in Windows Animation Tutorials
The Amiga home computer shined brightly through the dark ages of computing, before the words “Windows 95″, “3D Acceleration” and “Multimedia” ever existed.
God I fucking loved my Amiga!
Jodezza | Oct 10, 2008 | Reply
Thanks for this video! I think I saw this too on my friends Amiga. And if I remember (20 y ago) it was stored on one floppy. But I may be wrong. Anyway. Amiga 500 was great machine. The time when “computers” were made, before the gray uniform machines.
pavels68 | Oct 10, 2008 | Reply
Had one, loved it, the A500
Trew it away after 20 years, when it broke, with a lot of pain in the heart
But until the end of days I used this demo to prove the gratsness of this machine
Bring back memories
OtB1977 | Oct 10, 2008 | Reply
Comparing the 1985 computer with Flash Animation (the latter not as smooth, maybe ‘cos of the OS) I can safely say that we had a very good machine in those days… When I first saw it on my Amiga 500 with 512Mb RAM I was impressed… we had never seen such an impressive demo.
kevincarbonaro | Oct 10, 2008 | Reply
Ooh, they are smarter than I gave them credit for.
The “trail” effect is, in fact, a by-product of the rendering method, which they purposefully exploited for a creative effect.
Instead of blanking it out to the background colour, they blanked it out to all different colours to leave a “trail”.
Oh, that’s quite clever. Taking something you’d normally hide, and actually making it visible with psychedelic colours to create a cool effect.
Very creative thinking there.
KlaxonCow | Oct 10, 2008 | Reply
p.s. Actually, just thinking some more, another way to speed it up and not have to draw the entire shape every frame, is just to store the differences between frames.
Just draw the bits that have changed since the last frame.
And what those “changed bits” would look like is, lo and behold, what we see in the “trail” effect.
KlaxonCow | Oct 10, 2008 | Reply
Indeed, half the battle with programming is working out the best method of doing something (especially demo coding, where you’re constantly trying to outdo what’s already been done before).
The coding is actually the relatively easy part, in comparison to that. ;D
KlaxonCow | Oct 10, 2008 | Reply
Yeah, if you were doing things in 3D, then you would need a “spot suit” (and at least two cameras to film it from different angles) because you’ve got the extra dimension of “depth” to worry about.
So, generally, you’re not wrong about those “spot suits”. That is exactly what Hollywood uses to do “motion capture” these days.
But it’s just that, in this particular case, it’s a simpler problem (you only need the flat 2D outlines here), so you can use a simpler solution.
KlaxonCow | Oct 10, 2008 | Reply
“Spot-on” I would say
Your theory seems very likely…anyway, I have never dealt with these concept in practise, so I guess I tend to think too complicated…I guess the “spot-suits” are more for 3D objects (in movies), not just moving polygons
kolabeist | Oct 10, 2008 | Reply
And you thought Apple was inovative.
PS Gotta admire the flow. If they didn’t first tape people, then really… big up.
someman7 | Oct 10, 2008 | Reply
Todas las pendejadas que Quieran, solo piensen que esta animacion en tiempo Real fu realizada en el AÑO 1985, entonces no existia la tecnologia con la que ahora contamos. No flash no Nvidia, no suits 3d. etc. etc. COMMODORE 64 & Amiga was . JUST PURE FUCKING GOOD PROGRAMERS AND REAL HACKERS…
nuezoo | Oct 10, 2008 | Reply
Still, it’s an excellent vector graphics engine for the time.
It’s smooth as silk, and adds “trail” effects and other stuff too. That’s some very impressive programming.
So, no, knowing how they did it doesn’t, for me, detract from the fact that it’s still exceptionally bloody good.
KlaxonCow | Oct 10, 2008 | Reply
Ever watched one of those Flash-based cartoons on the web?
Well, Flash uses vector graphics in the same way. The Flash files contain vector shapes - sets of coordinates to define shapes that take very few bytes - and then the plug-in draws those shapes in real-time.
It can compress simple shapes amazingly well.
Because you’re not storing all the pixels of the shape, you’re just storing a “description” of the shape which tells the rendering engine how to draw it.
KlaxonCow | Oct 10, 2008 | Reply
p.p.s. It’s somewhere between precalculated and real-time.
The vector shapes are precalculated. But they are drawn real-time.
Vector graphics are highly compact.
For a triangle, you’d just store three sets of coordinates - very few bytes needed - and then the rendering engine will draw a filled shape from that.
The drawing requires some real-time calculation. But not a great deal.
It’s real-time rendering of precalculated vector shapes. Thus, somewhere in the middle.
KlaxonCow | Oct 10, 2008 | Reply
p.s. That’s funny, I had to change the word “dot” to “spot” in the above, or YouTube’s anti-spam filter rejected it. Because it thinks that the word “dot” appearing too many times in a comment must be spelling out a web address in words, or something.
KlaxonCow | Oct 10, 2008 | Reply
Just film it against a background that makes it trivial to differentiate between what is and isn’t part of the shape. That defines the outline. Turn it into vectors.
If they did use some kind of “dot suit” then they did more than they needed to.
KlaxonCow | Oct 10, 2008 | Reply
Making such a suit, detecting the spots, dealing with the problem of obscured spots, recognising that a particular spot in one frame is the same spot in the next, …
In practice, this would actually prove to be more complicated.
You only need the outline. There’s no need to start trying to do complicated “motion capture” to get that.
KlaxonCow | Oct 10, 2008 | Reply
(There’s a trade-off there between memory used and quality - experiment to find a happy medium.)
Store all the co-ordinates into a file.
Then just write a simple vector graphics “engine” that reads back those co-ordinates and plots the vector graphics.
Would be fairly simple to automate the whole process.
A “vector dot suit” would actually be more complicated.
KlaxonCow | Oct 10, 2008 | Reply
I think it was probably simpler than that.
Film someone in front of a green screen (well, the colour doesn’t matter - so long as it’s a “flat” single colour).
Whatever’s “not green” defines the outline of the shape. Once you’ve got the shape, then convert into vectors.
To do that, create points tracing around that outline that defines a complex polygon.
Optimise it by reducing the number of points to a minimum - simpler polygons - that still retains the overall shape.
KlaxonCow | Oct 10, 2008 | Reply
I guess this is vector-gfx? Filming a person wearing a “vector-dot-suit”. So, the only thing needed is to feed these captured dot positions to a 3D-screen-figure…dont know, but my theory..allthough I was better in ripping than inventing
colabeist | Oct 10, 2008 | Reply
I enjoy reading you insights on various CPU’s…sheesh the last time I tried to code assembler was for the 68000 on the NEOGEO gaming system..
MoneyMatthew | Oct 10, 2008 | Reply
Thank you for remmebering me for that spot, I have enjoy wathing that video on Amiga !!!
idovecer | Oct 10, 2008 | Reply
Great fun to watch! TY for posting it (at ANY quality!)
billmills | Oct 10, 2008 | Reply
hahalol
thizzboy99 | Oct 10, 2008 | Reply
and whole this thing came on one floppy disc 880kb…that’s power
NiNDz4aaa | Oct 10, 2008 | Reply