Sampled acoustics percussion instruments |
Roberto Aimi | |
|
How can we make digital musical instruments as immediate and rich as acoustic ones? I am exploring methods for designing digital percussion instruments that aim to extend the capabilities of the player while maintaining much of the richness and understandability of acoustic percussion instruments. Unlike conventional percussion controllers that measure and convert the intensity of strikes into a trigger message, but ignore the timbre of the hit and fail to track more ambiguous input, this method processes the raw acoustic output of the physical controller object to add the resonance of a sampled instrument. This is achieved by employing realtime convolution algorithms which are well know in computer music, but have not typically been used to create realtime percussion instruments. In an example of the system, audio from a contact transducer mounted inside a highly damped controller object is continuously convolved with a sampled instrument impulse response. The result is as if the sound of the real object were played through the sampled instrument. One of the advantages of this approach is very light taps, scrapes, rubbing, or swishing with brushes all take on a hybrid timbre of the real and sampled sound that is surprisingly realistic and controllable (based on preliminary results). This gives a level of believability that the instrument is real, and the player's intuition about physical objects can be applied to controlling it. The ability to transform the apparent acoustics of objects also suggests interesting opportunities for applying similar systems to interactive installations. Design of the physical interface and the corresponding digital processing is different from typical percussion controller design because the acoustic properties of the controller are a critical component of the sound, and the usual division between controller and sound source is much less clear. Part of the challenge is to identify meaningful ways to control the convolution algorithm that are consistent with the physical design while still expanding on what real systems are capable of. |
||
|
[Home]
roberto@media.mit.edu |
||