Archimedes Silverball & Silverball 2

[Rocketeer]

Silverball, a full-featured pinball simulator for the Archimedes series of computers, along with the follow-up Silverball 2 that featured a built-in table editor. These have been available on a few public domain discs and internet sites tucked away out of sight for a few years now. Thought I'd bring them to your attention as part of a number of RiscOS applications and games I wrote in the late '90s. Silverball was originally published by The Fourth Dimension at the end of the 90's as part of their VFM range (value for money). Silverball 2 was released as freeware.

Instructions attached and also within the applications. I've got these running on RPCemu 3.71 and are provided here as regular 7-zip files, just unpack the contents into the hostfs folder on the emulator then use as normal RiscOS applications from within there.

Sample screenshots from Silverball 2 and the editor where custom tables can be created. Silverball is very similar with 10 pre-defined tables to play but no editor.

screenshot 05.png (27.85 KiB) Viewed 2087 times

screenshot 04.png (36.49 KiB) Viewed 2087 times

screenshot 03.png (31.62 KiB) Viewed 2087 times

screenshot 02.png (65.73 KiB) Viewed 2087 times

screenshot 01.png (60.51 KiB) Viewed 2087 times

screenshot 00.png (43.32 KiB) Viewed 2087 times

[tricky]

Looks grea.
I have an early days pinball for the beeb, or it will be once I have converted the C fixed point maths version on Windows to 6502.
If love to know hope you choose to do your physics and what limitations were built into the emulator.
Do you know of any good videos of either?

[Rocketeer]

Looks grea.
I have an early days pinball for the beeb, or it will be once I have converted the C fixed point maths version on Windows to 6502.
If love to know hope you choose to do your physics and what limitations were built into the emulator.
Do you know of any good videos of either?

Physics wise I chose to have all movements/forces as summed vectors, for instance object collision was at pixel level and each object would have it's own routine to determine what vector should be added onto the ball's. A triangular side piece would calculate simple reflection on the two straight sides while the powered side would take a vector from the center of the object. This would be scaled according to where on the powered side it hit - the strongest being at the center. The same principle for the physical environment properties such as gravity, viscosity etc.

The limitations of the emulator is that I didn't strive for a physically accurate model, just one that behaved realistically to a game player while altering the variables. There were probably other properties that I just didn't cater for such as ball spin.

Don't know of any videos to help out because at the time they simply weren't available.

[tricky]

Thanks, that's roughly where I am except I went for 10.6 fixed point pixel position with 4 (or maybe 5) updates per 1/50 of a second.
I found for small posts that I had to interpolate the ball position back to where it crossed the edge and bounce from their otherwise a glancing blow could accelerate the ball.
But otherwise, like you have customer code for the various constructs and tuned things to feel right.
The flippers have been a pain as have a few other bits where the ball sticks to edges or keeps rolling up slight hills!

[Rocketeer]

Thanks, that's roughly where I am except I went for 10.6 fixed point pixel position with 4 (or maybe 5) updates per 1/50 of a second.
I found for small posts that I had to interpolate the ball position back to where it crossed the edge and bounce from their otherwise a glancing blow could accelerate the ball.
But otherwise, like you have customer code for the various constructs and tuned things to feel right.
The flippers have been a pain as have a few other bits where the ball sticks to edges or keeps rolling up slight hills!

Went for 24.8 fixed point position (too easy with 32bits to play with) for all maths, don't recall the frame update rate but there were many iterations of the physics math between frames to keep things precise with regard to collisions etc. Sticky balls were definitely a problem to begin with!

[tricky]

Yes, running on an 8bit CPU means lots of compromising and in some places changing the scale of the world.