Musical Instrument Form Finding in Grasshopper?

Replies to This Discussion

Permalink Reply by Hannes Löschke on January 14, 2014 at 2:02am

Do you mean the whole instrument or just the bars?

Permalink Reply by Mark Loomis on January 14, 2014 at 9:41am

My intent is to use Grasshopper for the form finding of the bars. The instrument form would follow the bars geometry. I'm guessing it would probably require a live physics engine (Kangaroo) and possibly other GH plug-ins to pull it off. Thanks.

Permalink Reply by Mateusz Zwierzycki on January 14, 2014 at 9:54am

That's a good question. I'm not very musical, but maybe (I really don't know is it a proper way to approach this) we can simulate the material vibrations with something like this ? (with a greater wavelength ?)

Correct me if I'm wrong - if I punch a xylophone bar, does it emit one sound frequency or 3 (WxHxD) or more ?

Permalink Reply by Hannes Löschke on January 14, 2014 at 11:37am

In fact, most Xylophone bars are way more complex than just (WxHxD) boxes. They have an arc shaped undercut and some have n additional tubular resonator. Both features as well as the positon of the fixing holes determine frequency and spectrum (timbre) of the bar.

I once implemented a pretty straight forward algorithm for tuning brass instruments. So maybe there are some general formulas to calculate the form of xylopone bars. Other than that, you'd need to employ FEM to find the eigenfrequencies of the bars depending on the shape.

Permalink Reply by Mark Loomis on January 14, 2014 at 12:23pm

If FEM is required you'd have to also use Weaverbird to convert the GH definition into a mesh that the FEM could use.

I noticed that Mateusz's example (awesome work!) used both Kangaroo and Weaverbird.

Here's an analog description of building a xylophone that might prove helpful in creating a parametric equivalent: http://www.tidewater.net/~xylojim/xylocons.html

As Hannes points out, the shape of the bar can be hollowed out but it sounds like it might not be necessary to achieve acceptable results.

Thank you both for your feedback.