Michael Flynn — scientist, designer, artist, and lecturer in the School of Art and Design — designed a bowl that could sing; now all he had to do was figure out exactly how to make it. Naturally, he came to the University of Michigan Library.

Specifically, he came to Spatial and Numeric Data Services (SAND) on the second floor of the Hatcher Graduate Library. He arrived without a bowl, but toting a computer full of data — quite a lot of it. “He opened his laptop,” says Jennifer Green, spatial and numeric data librarian, and manager of SAND. “There were many, many open programs and windows.”

Among them were Excel spreadsheets, .wav datasets, and other files, all containing various samples and renderings of the one-second piece of music — Flynn’s own voice, singing the line “Love is all you need” — that would be somehow recorded in the bowl.

But Flynn was so deep into his project, so full of questions about what he might or might not be able to do with all of his data, that it wasn’t immediately clear to Green what he was trying to accomplish. It wasn’t the first time that Green, who’s been working at SAND for five years, had to build a bridge between her own skills and expertise and what a researcher was trying accomplish. But she knew pretty quickly that Flynn’s would be one of the more unusual data problems she’d ever encountered.

Eventually, Flynn mentioned those plastic strips that “talk’ when you run your thumbnail over them. “Well, I’ve never actually seen one of those, but I got the idea,” she says. Those strips, it turns out, use the same technology as a phonograph; they generate sound by means of recorded vibrations on a physical object. Flynn’s bowl would use a rolling ball, as opposed to a thumbnail or a phonograph needle, to make those vibrations, but to make it work he had to translate the computer-based data he had onto the physical media (plywood, it turns out) that he planned to use.

So what Green and her colleagues had to work out was how to render multi-dimensional data — sound waves, represented as data points in an Excel spreadsheet — into a “flat” format that retained information about amplitude and frequency, and could be fed to and correctly interpreted by the fabrication equipment.

SAND was not Flynn’s first stop for technical assistance. By the time he met with Green, Flynn had already determined that the 3D printer at the Digital Media Commons (DMC) was not the best way to produce the arced pieces of contoured plywood that he needed. Next, he visited the FABLab at the Taubman College of Architecture and Urban Planning, and determined that the most likely candidate was its laser cutting equipment. Flynn envisaged something like a belt with bumps on it, resembling a bar code, which would wrap around the top of the bowl.

That’s when Flynn was directed to Jennifer Green and SAND, which is on the second floor of the Hatcher Graduate Library.

“People told me that she’s game for a challenge,” he says. “I rolled in there and described this problem, and she said, ‘Wow.’” Though the data problem was outside the realm of the typical research challenges that come to SAND (see sidebar), Green concluded that she probably had the tools and expertise to take it on. It was a stretch, she admitted, and “she was fired up about that,” Flynn says.

Green is an expert in government data and maps, applied geographic information system (GIS) tools. Using GIS software, she and her colleague Justin Joque rendered Flynn’s data into a graphic, essentially a .png file containing a series of gray-shaded boxes, after ascertaining that a .png file could support a sufficient number of shades of gray—1000-50,000, depending upon which sample set Flynn decided to use. This data would be fed to the FABLab’s laser cutter, which would produce four arcs with raised surfaces of varying heights corresponding to the shades of gray in the .png file. Flynn would put these arcs together to make the musical hoop—the bowl’s voice box.

But when Flynn returned to the FABLab with his newly-formatted data, he ran into one of the limitations of the laser equipment. (Photo left: Close-up of the laser cutting on the bowl.) It turned out that the FABLab’s laser cutter is very good at punching all the way through a substance, and is less good at controlling cutting depth precisely enough to produce the contoured surface that Flynn needed. Wes McGee, a lecturer in architecture and an expert in digital fabrication, advised him to flip it—to instead cut the sound waves out of the top edge of the plywood. This complicated the assembly a bit—it meant stacking four of the cutout arcs to create a track large enough for the ball to reliably play—but in the end it made the most sense to leverage the equipment’s strengths.

Though her data wasn’t used—the final input was a graph of the intensity of air pressure generated by Flynn’s voice — Green is happy to have been one of Flynn’s stops along the way. “It was interesting, and it was a challenge,” she says. “And I really enjoy working in a place that fosters this kind of creative collaboration.”

Flynn says that Green played an important role by filling in his initial plan, and making it possible to determine what the equipment could and couldn’t do for him. He also says “I was amazed at how willing people were to share their time and expertise.” In addition to Green and McGee, Flynn mentions a long list of others, among them Shawn O’Grady, a rapid prototyping expert who works in the Digital Media Commons; Stephen Rush, faculty member in the School of Music, Theater & Dance; Mojtaba Navvab, faculty member in the Taubman College of Architecture and Urban Planning; Michael Vitale, studio coordinator at the School of Art and Design; and Maciej Kaczynski, lecturer in architecture, all of whom gave willingly of their time, creativity, and enthusiasm in order to help him realize his vision.

This collaborative spirit seems fitting. The bowl, which has a circumference of approximately ten-and-a-half feet, sits on a pedestal, and making it sing requires the cooperation of at least two people, who must together rock the entire bowl, until the orbiting ball is high enough to reach the textured surface that produces the music.

The Singing Bowl cooperative phonograph is on display at the Grand Rapids Public Museum – Exhibition Center through Oct. 10; it is also entered in the ArtPrize competition.

What is SAND?

MLibrary’s Spatial and Numeric Data Services (SAND) offers assistance with spatial data, numeric data, and statistics for the University of Michigan community. SAND works closely with other library and campus units that provide related support, including the Map Library, the Government Documents Center, the Center for Statistical Consultation and Research (CSCAR), and the Environmental Spatial Analysis (ESA) lab in the School of Natural Resources and Environment.

When SAND staff isn’t working on projects like the Singing Bowl, which is to say most of the time, they’re helping researchers to:

• Locate spatial or numeric data; for example, projected global temperatures and emissions for the next 100 years.
• Locate demographic data, such as census information.
• Convert data into usable form, sometimes by writing small software programs to handle conversions.
• Map data; for example, putting data points on a map to identify least cost paths to hospitals.
• Generate contour lines (for architects making site models) using digital elevation data from the U.S. Geological Survey.
• Put data into statistical packages.
• Use GIS, including installing and using the various GIS software plug-ins, some of which are difficult to install and use.
• Data analysis.

SAND librarians usually meet with patrons in person due to the complexity of both their queries and the technology required to resolve them. The problem solving is collaborative, and often involves verbal and visual descriptions, as well as joint examination of data and the software used to manage it.