Moog, Buchla, ARP, Korg, Oberheim and Sequential Circuits, to name a handful, all made foundational and deeply influential synths. But you could argue that no company's gear steered the course of dance music quite like Roland's, particularly a trio of machines produced for a brief moment in the '80s: the TB-303 bassline synthesizer and two drum machines, the TR-808 and TR-909. (It's worth noting that the company also made a number of famous synths—the Jupiters, the Junos, the System-100, the SH-101—and effects, like the ubiquitous Boss stomp boxes and the RE-201 tape delay unit. The TR-606, TR-707 and TR-727 drum machines, from the same era as the more famous ones, garnered their own followings as well.)
The company has its roots in the organ trade, and early, non-programmable drum machines like those in the Rhythm Ace line, made by Roland's predecessor company Ace Tone, weren't seen as instruments in their own right so much as accessories to be used in the absence of actual players. The programmable 808 and 303, released in 1980 and 1982, were cut from the same cloth—Roland marketed them to guitarists and pianists as tools for practicing or writing parts. Early promotional posters showed Jimmy Page riffing along with an 808 placed on a stand next to a mixing desk, and Oscar Peterson playing an electric piano that has a 303 and 606 perched just above the keys. Neither sounded much like the instruments they were standing in for, so the company didn't imagine they'd have much of a use outside of the studio.
That they didn't sound real ended up making them the most important and enduring instruments of the era. No one had heard a kick drum boom quite like the one on the 808, and no cymbal had ever sliced through mixes quite like the ride on the hybrid-digital/analog 909, which followed in 1984. The 303 was its own animal entirely, capable of churning out endless, looping basslines that could morph into wild, squelchy leads. There were better instruments out there, objectively speaking—the LinnDrum, released in 1982, was a bit more convincing and flexible—but Roland's were the most distinctive, and their reasonably low price point (before they took off on the used market, at least) made them attractive for music-makers keen on experimentation. These units seemed to have come from another planet, bringing their own musical tradition with them. In 1982, before Jesse Saunders' "On And On" and Phuture's "Acid Tracks" came tearing out of Chicago, a Bollywood musician named Charanjit Singh used a 303, an 808 and a Jupiter 8 to make an album called Synthesizing: Ten Ragas To A Disco Beat; though the music isn't house, the sound the gear lends it is unmistakably acid.
In 1984, Roland stopped producing the 303 and 808, and by '85 they'd pulled the 909. The products, particularly the 808, were reasonably popular, but it was only after they'd left the market that they became smashes. Yet that was it. No matter how many rappers name-checked the 808, or how far 303-led tracks climbed the UK singles charts, or how hard techno producers like Jeff Mills pushed the 909, Roland didn't bring them back. Starting in the mid-'90s, the company rolled out a number of "grooveboxes" like the MC-303, which borrowed the sound and bits of the workflow from the classics but weren't really pitched at pros. Only last year, with the launch of the AIRA line of dance-music-centric studio gear, did Roland come close to reissuing their famous bassline synth and drum machines, and thus wholeheartedly embrace one of their most important legacies. To followers of electronic music, or even to fans of popular music in general, their logic is inconceivable at first brush, and that was a big part of why I was so excited about Roland's invitation—I wanted to find out what took them so long.
"Rather than moving forward with techno, we were progressing generally in music production," said Atsushi Hoshiai, a developer who's been with Roland for 33 years. His first project with the company was the 909, and when we met in a sunlit meeting room at the R&D center, he was carrying a Paiste ride cymbal on a stand. It was the one he'd brought in from home to sample when the synthesized versions they tried weren't cutting it. Hoshiai still hasn't wrapped his head around the impact Roland's machines have had, though he's philosophical about and flattered by their immense popularity. "The fact that people are using the sample of the sound of me drumming a long time ago—it means that we are playing together. So I feel amazing about being able to play with all sorts of musicians from all around the world. But at the same time, if you ask me if that means I listen to the music using those sounds, well"—he lets out a laugh—"actually, I haven't really."
I'd hoped to ask Hoshiai for a list of his favorite Chicago house records, but I quickly realized I'd need to scrap that line of questioning. What really made a product like the 808 important for Hoshiai, and for Roland as a whole, wasn't that it was used to push music forward; rather, that the product itself did something new. "A rhythm machine that could be programmed didn't exist at the time," he said, "so it was extremely innovative." I sensed he was less in awe of his own 909, which had hit shops that were suddenly more crowded with drum machines capable of sequencing "the nntsk-nntsk beat," as he put it. "We started to see all sorts of different maker brands, and then PCM drum machines were released, so it started to become a competition, meaning that the 909 didn't have the same impact as when the 808 was released." Though the 909's cymbal samples became some of its most important sounds, they were the result of time, budget and aesthetic constraints, not the kind of technological forward motion that makes a product a true success for Roland.
From the meeting room, we went into a studio control room to meet the team behind AIRA: Kazuhiro Kubo, who manages the line; Daichi Tawara, who works on effects programming; and Kenichiro Nishi, who wears a number of hats, like drafting first designs, designing the user interface and working on the sounds. Of the three, Nishi, a youthful 20-year veteran of the company, is the most invested in dance music outside of his day job. "I initially joined Roland because I liked the TB-303," he told me with a laugh, and he later showed us one he'd had signed by '90s German acid obsessives Hardfloor, one of his favorite groups.
In the control room, the whole AIRA line had been assembled on a table in the middle of the room, and Nishi started jamming out a lush but incisive techno beat, pushed to an earsplitting volume by a pair of Funktion-One monitors. The black-and-neon-green boxes he was fiddling with didn't look much like Roland's classics—only the TR-8, the line's 808- and 909-emulating drum machine, in its red, orange, yellow and white step keys, overtly references the 808's look—but they sounded like cleaned-up versions of them. I'd watched plenty of sound-by-sound video comparisons emphasizing the finer points of their disparate tones and timbres, but in the midst of a loud live PA, I'd be hard-pressed to differentiate. They sounded fat and convincingly analog.
Only they aren't. "We don't work on the premise of rereleasing an old machine," emphasized Shinsuke Takami, a member of the marketing department who was our de facto host for the afternoon. "The most important point was thinking that there must be a reason why TR and TB sounds have been used in dance music for over 30 years. We didn't know the reason why, so then we thought to disassemble the machines completely and take out the inside information. And that was how what we call our ACB technology came about, where we model each analog part, bit by bit, and look at how the TR-808 actually worked, and how the TB worked."
"ACB" stands for Analog Circuit Behavior, Roland's proprietary modeling technology. Put simply, this type of modeling tries to recreate an analog signal path in the digital realm by representing each step via complex digital signal processing. The whole AIRA hardware line—the TR-8 drum machine, the TB-3 bassline synth, the SYSTEM-1 synthesizer, the VT-3 voice effects box and the MX-1, a mixer with sequencing, syncing and effects capabilities—has a digital backbone, but the 303, 808 and 909 sounds the units produce aren't samples; they're a deeply rooted imitation of how the originals worked. Hoshiai characterized it this way: "The difference between PCM and modeling is that change you get in the sound when you move the knobs this way and that."
Computing power only recently got to the point where Roland could pull off this kind of DSP at a price point they could work with. That confluence of technology and affordability is how Roland felt confident dipping their toes back into dance music gear without merely doing a remake, something that countless other devices, from German manufacturer JoMox's XBASE 999 to the myriad x0xb0x 303 clones, had already achieved, and that would have gone against Roland's design philosophy. "I have a strong desire to tweak that sound and present it to the next generation," Hoshiai said. "If you make an analog circuit—if you don't remake that circuit—the sound won't change. But if it's digital, just by changing it in the program, you can make all kinds of progress."
Going digital had a number of other advantages—they could be endlessly tweaked and updated on the manufacturer-end after the release, and a single piece of hardware could theoretically become any number of synths or drum machines. For users, ACB-based sounds would be more flexible than samples, and the machines would probably sound more like the classics than fully analog recreations could, too. Because analog circuits age, there's a good chance that new circuits copied from old ones might not sound quite right anymore. And there were quirks to these machines resulting from them being built the way they were, when they were. "The product itself is different, so it sounds different," Hoshiai said of the AIRAs. "It's not going to be the same, and the original analogs have a different sound in each and every one of them."
Takami explained that AIRA was both a multidisciplinary project and a model of efficiency. "There was a project growth strategy, where you start something new in the company, and that was held for exactly three months, from September 2012. Teams were assembled, and all sorts of ideas started coming out from there. Until that point, if you were on drum development, you'd make drums; if you were on piano, you'd make pianos, etc. And in that mix, the idea was to make an instrument for dance music that was different from Groovebox." The team beefed up on the contemporary market—"hardware similar to that of AIRA," Kubo said, "but also controllers that were related to software, all sorts of pieces like that"—and set to work on their own entry. That January, the project got the green light, and 12 months later, the first photos of the gear began filtering out to the press.
In the year since, AIRA has grown both digitally and physically. The SYSTEM-1 had a brand-new synth engine right out of the box, but you can now load in an ACB version of the SH-101, the SH-2 and the PROMARS. To the TR-8, they've added the 707 and 727. (The sound sources themselves were always PCM, so the expansions feature the original waveforms, not ACB models; the analog decays, built in to compensate for low bitrates in the devices' DA converters, and clocking deviations, however, were modeled.) After getting lots of requests for mixers from artists who came in for demos, Roland began developing the MX-1 specifically for AIRA, which would take audio and sync information from each piece by way of USB to cut down on cable clutter. (They've dubbed this AIRA LINK.) I asked if Roland had any further plans to expand AIRA, but I hit the engineer-interview wall. Takami said it was likely but politely declined to give any specifics.
The answer to the question that brought me to Hamamatsu is that Roland weren't dragging their feet—they just weren't interested in getting involved until they had something new to bring to an old idea. If you've spent decades listening to house and techno, you have a kind of implicit understanding of what makes these machines important. But if you're a company trying to move forward with each new release, it might take pulling them apart and analyzing their intricacies to really see what they are. The engineers all said this to me in one way or another, but it took a walk through Roland's private collection of gear for their words to really sink in. Nearly everything they've ever made, from Ace Tone organs and obscure digital recording equipment, to each and every Juno and Jupiter, is laid out in clusters in a vast, open-plan room. It took some time for me to find the drum machines and the 303, but finally I arrived at their shelf in the back corner, the AIRAs facing them like attentive pupils just across the aisle. For us, these boxes achieved a peculiar kind of perfection, the source of critical sounds, rhythms, textures and whole ways of making music. For Roland, though, they're part of a progression that started before them, continued after them and won't be allowed to stop. I couldn't help but notice that the gallery was only half full.