Detangling the Complexity of Waves with Acoustic Voxels
Anne Nasato posted on August 17, 2016 |
By optimizing the structure of acoustic voxels, Columbia Engineering researchers can control the acoustic response of an object when it is tapped and thereby tag the object acoustically. Given three objects with identical shapes, they can use a smartphone to read the acoustic tags in realtime, by recording and analyzing the tapping sound, and thereby identify each object. Credit: Changxi Zheng/Columbia Engineering.
Columbia Engineering researchers were able to control the acoustic response of an object when it is tapped and thereby tag the object acoustically. Given three objects with identical shapes, a smartphone can read the acoustic tags in real time by recording and analyzing the tapping sound and thereby identify each object. (Image courtesy of Changxi Zheng/Columbia Engineering.)
A novel way to simplify the design of acoustic filters is being developed in a collaborative effort among engineering researchers via simulation methods.

The engineering research team behind this development decided to study a rather simple shape (a hollow cube with holes on some of its six faces) in order to enable 3D printing it as their base module. This new technique is capable of determining optimum filter designs, which then enables the selective reduction of sounds at specific frequencies.

This approach has been named “Acoustic Voxels” by its creators. Acoustic Voxels aids in veering away from using trial-and-error iterations in the design of acoustic filters. Instead, this program precomputes the acoustic properties of an item. It also enables the user to simulate the filter with varying properties. 

Additionally, the engineering research team behind Acoustic Voxels created a technique for computationally optimizing attachments between filters in order to achieve a desired effect. Acoustic Voxels operates 70,000 times faster than current algorithms used to predict acoustic qualities.

An interesting outcome of Acoustic Voxels was that the team could design acoustic tags into objects that seemed to be the exact same as each other. However, when tapped, each object would give a distinctive sound. Although the frequencies affected often demonstrate significant dependence on the shape of the cavity, the exact influence of the shape is complex and difficult to understand. 

Acoustic Voxels not only sped up and computationally optimized the design process, it also enabled the design of more advanced geometries. Current computational tools are limited to more simplistic shapes.

When waves are transmitted through a cavity, some of them are reflected back and forth. These reflected waves either result in a constructive superposition, which amplifies the sound, or destructive superposition, which muffles the sound. This is how acoustic filters operate.

Wojciech Matusik, associate professor of electrical engineering and computer science at the MIT Computer Science and Artificial Intelligence Laboratory explained the current state of this study: Thus far, the method is mostly suitable for controlling impedance and transmission loss at discrete frequencies, such as in traditional muffler design. 

However, the scope of this study only covered one shape of a single material. “Extending our method to additional shapes and materials could offer a larger palette for better acoustic filtering control,” said Matusik.

The engineering research team behind Acoustic Voxels was a mixed, collaborative group. It was composed of members from Disney Research, the Massachusetts Institute of Technology and Columbia University. This development was supported by the National Science Foundation. 

For more information on Acoustic Voxels, visit the project website.

Recommended For You