views:

125

answers:

1

In considering the design of marble-in-maze games where you tilt the table to get the ball to the end of the maze without going down one of the holes, I wonder whether anyone here has considered the modelling of the sound of the ball hitting the walls...

The ball doesn't always make the same sound.

This other question covers the rolling sound:

http://stackoverflow.com/questions/1325229/sound-of-a-rolling-ball

But I am more interested in the bouncing sound - I am often struck by how unrealistic it is in most people's version of the game.

What are the factors to consider to work out how to produce a realistic sound?

How must the sample or raw data then be processed or generated?

+2  A: 

There are some good links in the Sound Modeling section of this page from a course at Carnegie Mellon: http://www-2.cs.cmu.edu/~djames/pbmis/index.html. The instructor, Doug James, is now at Cornell as does similar research there (http://www.cs.cornell.edu/projects/Sound/).

I've never tried to implement any of these methods, but I suspect that they're overkill and/or too slow for a small game. However, you might be able to generate several samples offline and choose an appropriate one at runtime.

Hope that helps.

celion
Woah! I think I see why this is not often done. Serious maths there. Thanks for the answer. At least I now have an idea of where to start looking to understand this further.
martinr
Yeah, it's a lot of work to get physically accurate sound. However, the sampled-based methods that people mentioned in the other thread should be good enough to fake it convincingly - and that's what game development is all about! :)
celion