In this TED Talk Ge Wang of Stanford geeks out about computer music (hooray!). He talks briefly about the hemispherical speakers used by the Stanford Laptop Orchestra. Their design (using an IKEA salad bowl) informed my hemi speaker experiments here.
I recently bought a 6-channel amplifier to upgrade the 3 2-channel amps I originally installed in the speaker. I’ll post about the new amp as soon as I drop it in.
One of my goals when I moved to Albany was to set up a good-sounding room for my studio. I had done enough recording in my previous spaces to realize that the biggest limiting factor for me to capture a good sound had become the acoustics of the room itself. Sure, I’d still love a boutique preamp and collection of the finest mics, but they can’t overcome the properties of physics at work in a poorly designed acoustical space.
One corner of the finished studio, showing broadband absorbers mounted floor to ceiling (black) and two wall-mounted absorbers (orange). There's a glimpse of a white ceiling panel visible in the upper right corner.
My original plan was to treat my new studio before unpacking from our move, but there were more pressing renovation projects around the house: roof, kitchen, two baths, and nearly every other room but my studio. I also had various projects that required immediate work recording and editing sounds, so I unpacked and set up what I hoped would be a temporary studio.
Two years later I was still working in an untreated room, so I finally made a plan for the acoustic treatment I needed. I used Mitch Gallagher’s book Acoustic Design for the Home Studio as a starting point for my plans. Soon after, I began building some broadband absorbers. I followed the well-known technique of using rigid fiberglass panels–gift wrapping each one in funky, bright IKEA fabric chosen by my wife. The panels then languished in my basement for almost another year before I was able to pack away everything in the studio and clear the room for painting and installation.
Bye-bye Pepto Bismol pink walls!
While the room was empty and untreated I took a series of audio measurements which confirmed the laws of physics–modes predicted by the dimensions of the room–and my own experience recording and mixing. I conducted a series of listening tests using familiar musical material. (More on these test results in Part 2.) Clapping in the empty room produced that characteristic boxy sound with fluttery echoes.
After installing only half of the wall panels I already noticed a significant difference in the sound of the room: now the echo I heard when clapping seemed to come from the hallway outside the room rather than the room itself. The broadband absorbers were clearly working.
Mounting the Panels
For wall-mounted panels I screwed eye hooks into drywall mounts which screwed easily by hand into the rigid fiberglass. With a little picture wire it was a simple matter to hang them from a nail in the wall. To increase bass absorption, I velcroed 2″ or 4″ spacers cut from scrap wood to the back of some of the panels.
For ceiling-mounted panels I built simple wooden frames that were hung from eye hooks mounted in the ceiling.
Back of the wall-mounted panels, showing the picture wire used for hanging.
Detail of the nylon threaded drywall anchors used to secure eye-hooks to the backs of the panels
I recently completed my second salad bowl hemi speaker. (See info on the first one here, including links to the Princeton and Stanford laptop orchestras which provided excellent guides to construction.) My second speaker followed the design and construction of the first very closely, with the substitution of Polk Audio DB401 speakers. The Polk speakers were significantly cheaper than the Infinity speakers. I haven’t directly compared the speakers, but I remember the Infinity speakers to be heavier and louder than the Polk speakers. The frequency response of both seems very similar. I do prefer the mounting tabs on the Infinity speakers to the broad flange on the Polk speakers. When mounting them on a hemispherical surface, the Polk speaker flanges don’t lie quite as flat (er…curved).
Here are some photos I took during construction.
This is the bottom of the speaker. I simply scribed the circumference of the bowl on a piece of 1/2″ plywood and cut out the circle using a jigsaw.
Salad bowl with speaker holes marked and taped
Here’s the bowl with the position of each speaker marked and taped. If you look closely you can see a small red mark at the center of each circle. I used a string attached to the center of the bowl to mark the center of all the equatorial speakers. The tape is simply to prevent the bottom of my jigsaw from scarring the surface of the bowl.
Salad bowl with holes drilled for jigsaw blade
Next I drilled holes in each speaker cutout large enough to fit the blade of my jigsaw.
Salad bowl with speaker holes
Here’s the bowl with all the speaker holes cut. It’s easy to crack the salad bowl after removing so much of the material, so take care with all subsequent drilling and cutting.
Detail of cut used to enlarge back of speaker opening
One result of working with the elliptical geometry of a hemisphere is that the back side of the holes we cut is slightly smaller than the front side. This might prevent your speakers from sitting flush against the surface. I only needed to trim a few places from the back of each circle to get the Infinity speakers to mount flush, but I had to cut the entire back edge of each opening in order to mount the Polk speakers.
Finished hemi showing knobs and connections
Here’s the finished speaker. (I know, I skipped plenty of intervening steps! I was having too much fun putting all the pieces together to stop and take pictures.) The volume knobs (one for each stereo amp) are on the left. In the middle is the power connector. On the right is a 6-conductor Neutrik connector for all the audio signals.
Finished hemi on top of subwoofer
A front view of the finished speaker. It’s sitting on top of a Sony subwoofer I picked up at a yard sale. The two together have a nice R2-unit look. I’ve set the crossover fairly high (around 300Hz). I expect I’ll back it down after some more listening tests.
I’ve already performed once with this hemi. So far, the biggest problem I’ve encountered is having the amps cut out on me when the input levels get too high. The Dayton amps seem to have a protection circuit that shuts them down when they’re driven too hard. It’s better than having the amps blow up, to be sure, but a bit of a drag having the audio suddenly drop out. Sending the low frequency signals to the sub seems to ease the load placed on the hemi amps. I’ve also been experimenting with limiters and high-ratio compressors, but I haven’t yet found the silver bullet. I’ve only scratched the surface of spatialization possibilities with this setup, and I’m looking forward to working with it even more.