Month: July 2019

Joe Britt, system developer / computer engineer

In 1999, he bought his first machine, sight-unseen. He had just learned the difference between the Series I, II, IIx and III and he had no idea what he was up against. He took on the challenge and dived right into the seemly endless depths of the system. Along the way, he gained enough knowledge to patch up some wonky CMI’s, and to come up with a few inventions to extend the lifespan and usefulness of these beautiful machines.

“I bought my first CMI, a Series III, sight-unseen from a guy in Sweden. Then I was down the rabbit hole.”

When he was in high school, Joe Britt already knew his way around computers; he built his own Macintosh out of junk parts, he loved all things electronic, and his favorite book was The Soul of a New Machine. And, he was curious about this mysterious instrument called Fairlight. Joe: “I first learned of the CMI in the mid-eighties. As a long time fan of artists like Tangerine Dream, Jean- Michel Jarre, and Isao Tomita, I was blown away when introduced to the music of Art of Noise and Kate Bush. I was in high school then, and it was not easy to learn much about this “Fairlight C.M.I.” that was listed in the album notes I pored over. And what kind of instrument was this, that it needed a programmer? There was clearly something very electronic and very magical going on.”

Exploring

“I was introduced to Kate Bush’s Hounds of Love through a classmate’s modern dance performance to “Jig of Life.” That led me to explore the rest of the album, which just pulled me in like no other record had. I found it very deep, and I wanted to understand how she was creating those sounds. Playing things backwards could be done by tape of course, but doing that digitally, and being able to play it on the keyboard; that was really something. It’s hard to overstate the impact that music had on me. 

Electronic music of all kinds provided the soundtrack for my young life and later work. Time passed, I went to university, and I ended up in California working on consumer electronic products at a few different companies. I remember very clearly one day in 1999, working in the lab and listening to some music. The Web had become quite popular by that time, and I remember wondering what I could find out about the Fairlight now on the internet. I found the (famous!) Greg Holmes Page and was really excited! I voraciously consumed everything that was on that website, spending so much time scrutinizing the photos of the machines, trying to understand what was inside. There was also a classified ad section where people could buy and sell Fairlight equipment. (We did have eBay then, but the CMI was too exotic — there wasn’t anything listed.) And that’s where I found my first Fairlight. I didn’t really know anything about it. I had learned the difference between the I, II, Iix and III, and really wanted a IIx, but there on the Holmes Page classifieds was a Series III listed for sale. The machine was located in Sweden, and it wound up costing me US$7000. Its original owner was Puk Recording Studios, Denmark. According to the fellow I bought it from, the second owner was one of the guys from Ace of Base. I bought it sight- unseen.”

Troubleshooting

The day his Series III finally arrived from Sweden was a memorable one. He unwrapped the whole package and he set it up. Everything seemed in order, except for one thing: there was no sound. “I remember pretty clearly tracing out the signals from the music keyboard into the Fairlight and checking all the connections inside and outside. I didn’t even know the Series III keyboard was MIDI, because the connector was not a DIN5. I also tried driving the CMI from an external MIDI keyboard connected to one of the CMI MIDI inputs. That didn’t work either! Finally, I decided to just dig in and start debugging the hardware. If you’ve never taken apart a CMI, especially a Series III machine: they are beasts. First I took off the top of the keyboard, did a visual inspection and got a rough idea for how it worked. I went through the usual first checks: power supplies looked good, clock to the processor looked good. I recognized the a serial port chip, and could see data coming out of it when I pressed keys on the keyboard. So, the keyboard was alive; it was sending data. Looking at the timing of the data with an oscilloscope, I saw that it was at 31,250 bits per second — for anyone who has worked on MIDI hardware, that’s a dead giveaway! So I realized that it was just sending MIDI over a different kind of connector, and it looked like the keyboard part was working. “

“Next I had to follow the signals into the CMI and to verify that they were reaching the circuitry inside the mainframe. This was made more exciting by my lack of any technical documentation! All I could do was trace the wiring. In the Series III, the signals from the keyboard come into a connector on the back of the mainframe. They then transition down to another board that actually is part of the power supply, it runs across that PCB, to another connector, through an internal wiring harness, and up to a PCB that the rear mainframe boards plug into. A board with the CMI MIDI connectors plugs into that board. The CMI has multiple MIDI inputs, and it turns out that the music keyboard is just another one of them.”

“MIDI inputs use a component called an “optoisolator” to electrically isolate the connected MIDI devices. Once I found the optoisolator used to interface the CMI music keyboard, I replaced it, hit a key, and literally jumped when the CMI played a really loud brass sound! So, clearly, there was my problem.
The interesting part was this: the same type of optoisolator circuit was used for the external MIDI inputs. All of those optoisolators were blown out, too! That is really unusual. I never found out how that had happened. Anyway, I replaced the parts and got the machine basically working again. But, it turned out, that wasn’t all. There seemed to be a problem with one of the voice cards. Those things are super complicated, and I really needed a service manual or at least a schematic to figure out what was wrong. So, the next day I called Fairlight in Los Angeles and just asked, like: ‘Hey, I’ve got this machine, I’m looking for a service manual, could you help me?’ They transferred me to, as it turned out, one of the famous guys who had worked on the development of the Series III in Australia, Andrew Brent. Of course, at the time, I didn’t know who he was! But he was extremely nice. I think once I described what I had gone through to debug the keyboard problem, he probably took pity on me. He was incredibly kind, and offered to mail me a copy of the Series III service manual. Wow! I was ecstatic when I got it: man, I just couldn’t believe it! It was like a thick phone book, and It contained all the schematics and the entire theory of operations of the machine. This was also a time before document scanners were common. At work, we had a scanner, so I turned it into a PDF, and I think that PDF is still the one floating around on the net. For me, this experience was also emblematic of the whole Fairlight community: brilliant, friendly, and empathetic.”

The Fairlight-collector

“So, I had a lot of fun working on that machine, and I still have it. It’s a Rev 6 machine, one of the earlier ones. Around the year 2000, Peter Wielk came out with a memory extension board for the Series III. I bought two of them, and that machine has one of them in it now.”

But it didn’t stop there. Joe: “By that time, I was much more aware of the differences between the machines and had more clearly defined interests. I got another Series III, also a Rev 6 machine. It was kind of a basket case, which I originally bought a parts machine. I’m restoring it now, but need to find some more time! Then, I found a guy in LA with a Rev 9 machine, and I bought it for US$2000. Compared to the US$7000 I paid for the Rev 6, that felt like (and was!) an absolute bargain. I put the other Peter Wielk memory card in it and now have 2 great Series III machines, one rev 6 and one rev 9.”

“Back in the early 2000s I continued to visit the Greg Holmes page. And then one day there it was: a Series IIx for sale! And it was an interesting one. The guy I bought it from worked for Sony Music, New York, and he had bought it from Bill Laswell. Bill is an extremely famous bass guitarist and producer, and he worked quite a bit with Herbie Hancock. As It turned out, Herbie Hancock was the owner before Bill. The flight cases — still in my garage — have ‘Rockit Band Herbie Hancock’ stenciled on the sides. It’s also obvious that they toured with this machine. It’s got really cool stickers on it from all over the world: stickers for shows like Japan’s “Live Under the Sky ’87” and “World Destruction”.

When I got it, it wasn’t like a “normal” IIx. It was clearly originally a Series I that had been upgraded over time. This would be consistent with an early owner like Herbie, who would have likely had access to upgrades before they had been fully productized. The MIDI interface on this machine is unlike any I had never seen before: instead of a 68000 CPU, Herbie’s uses a 6809 on a board that was very clearly hand soldered.
A typical IIx also has a big box on the back which provides the MIDI in and out connectors. Mine only has three DIN5 jacks, wedged into a hole in the bottom, very hand-done. But it works great!”
This machine also has a very specific cosmetic detail: there’s a cigarette burn on the low F key. I love to imagine someone playing it, perhaps at “Live Under the Sky” in 1987, pausing to park a cigarette on this UD$25000+ CMI… It’s like a cool tattoo!”

“I love developing products, and love taking apart other products to see how they were made. I love thinking about the people who built them, and all the decisions they had to make. With an incredibly complex product like the CMI, built at a time when such an amalgamation of technology really was alchemy, looking behind as many of the curtains as possible becomes a kind of addiction. So I kept looking for more. The internet did not disappoint.
A couple of years later, probably around 2009, I ran across a listing for a Series II on eBay. Actually, the guy thought it was a IIx, but it was a II. He was an accomplished studio musician and teacher, with a career that included work with bands like Bananarama and studios like Paisley Park. Incredibly, this machine had been sitting in a garage for the past 20 years in a house just a few miles away from mine! Time had taken its toll on the machine, and it was not operable. There was also a bit of corrosion throughout from the humidity. It would be a fair amount of work to restore, but it would also be fun. I bought the machine and spent the next couple of weeks fixing it up and getting it back in good shape. So, that’s how I got my Series II, IIx and the III’s. Now all I need is a Series I!”

Nifty little boxes

“As a member of the Fairlight community I saw over and over again reports of trouble with the light-pen on the Series I/II/IIx machines. They break and are difficult to repair, and sometimes they are just lost. For certain parts of the UI, the light-pen is critical, and getting a replacement is practically impossible.”

“This inspired me to build an interface which lets you connect a USB mouse instead of or in addition to the light-pen. The response was great! I built and sold about 20 of them back in 2010, and I guess I should have built a few more — I still get emails from folks who want to buy one. It’s a neat little device, just a simple box that plugs in between the monitor and the CMI. If you move the mouse you’ll see the cursor moving across the screen, and you can click the mouse button to have the same effect as tapping the tip of the light-pen.”

 Any famous users? Joe: “I know one of my interfaces made its way to Jean-Michel Jarre! Jean-Bernard Emond {the famous Fairlight-fixer-upper from France} sent me a photo of JMJ’s studio and he explained he was using the mouse with my interface connected to his old series IIx. So, that’s super cool, you know… I saw JMJ play a couple of times, here in California, and It makes me very happy, knowing one of my childhood heroes is using something I’ve invented. The artist Benge also has one, and is using it on his project with Neil Arthur from Blancmange, FADER II. It’s really cool!”
”Recently, I started working on another interface that lets you plug a modern USB keyboard into a I, II, IIx or III. For the III, you can also use a mouse instead of the G-pad and stylus. Its also got a MIDI in port, which lets you drive any of those machines from MIDI — even Series I and II, which lack MIDI!”

In retrospect

“It’s such a deep instrument, especially considering the timeframe it came from. Take Page R, for instance. I’m sure there were similar concepts, like, the Roland digital sequencers, but they didn’t have a UI like that. They didn’t have a screen, so they couldn’t show a high fidelity representation of what the computer was going to play. Also, consider the display of the audio waveforms, in pseudo-3D on the screen. I remember thinking, “that is the sexiest thing I have ever seen.” It’s so cool. But not only that, it’s very useful as well. You can shift the data forward and backward and everything lines up just the right way when the loop point is right. You get a nice, topographic-like picture: it’s like the CMI gives you synesthesia, letting you see the sound. Genius!”

“I feel very fortunate, having been able to spend some time talking with Peter Vogel, learning about the history of the machine from him. As it turned out, we knew some people in common at Stanford University’s Center for Computer Research in Music Applications. It is well known for John Chowning’s research on FM synthesis (used to develop Yamaha’s DX-synths). Peter told me how he and Kim Ryrie were trying to create a digital synthesizer back in the ’70’s. They weren’t even thinking about sampling. What they really were aiming for was what we now call physical modeling synthesis. But back then, computers didn’t have enough horsepower to pull it off. Peter visited Stanford University in the 1970s to check out the digital synthesis work being done there, and that visit led to the conclusion that it probably just wasn’t practical given the limits of microprocessor systems at the time.”On the plane back to Australia, Peter realized he could kind of ‘cheat’ by capturing a sound as a starting point: the CMI could then take a bit of, say, a real piano sound and then modify it digitally to create a new but related sound. I think the big surprise came when everyone saw how mind- blowingly compelling it was to capture any sound and play it up and down a keyboard. The Mellotron existed of course, but the plasticity of digital took the concept to a whole new level.”

People in this picture are: Yoshi Yoshikawa, Jory Bell, Peter Vogel, and Joe Britt. Photo credit: Yoshi Yoshikawa
People in this picture are: Yoshi Yoshikawa, Jory Bell, Peter Vogel, and Joe Britt. (Photo credit: Yoshi Yoshikawa)

“There are these pivotal instruments which, from an engineering point of view, if you take them apart, contain surprising choices that are beautiful. Sometimes that beauty comes from brilliance, sometimes it comes from naiveté, and it’s really fun when you can’t tell which it is!”

“Like the Linn LM-1, the first digital drum machine. I saw the schematics and I thought: ‘Holy crap! That’s how it works!’ The timing of the playback, it’s magical in part because of its imperfections. The LM-1 uses a very famous early chip called the 555 for sample rate timing. This is an unorthodox choice. Normally one uses a crystal oscillator for a very precise sample rate. In contrast, the 555 is rather un-precise, and its frequency changes over temperature and voltage. Usually, that’s a bad thing. But, that is actually part of its sonic character of the LM-1! Its sounds are definitely digital samples, but the 555’s frequency variations due to its environment give them an unusual organic aspect.”

“The CMI also has some unorthodox (by modern standards) design choices that contribute to its sonic character. For example, most modern sample playback hardware runs at just one frequency, say 44.1kHz, 48kHz, or 96kHz. By comparison, the CMI varies the sample rate depending on the desired pitch! To understand this, imagine a buffer full of audio samples for some sound. To play the sound at a lower pitch, modern hardware would likely “play” some samples multiple times, so they consume more time. To play at a higher pitch, modern hardware would likely “skip over” some samples, effectively squeezing other samples closer together in time. The CMI, though, plays all samples every time — but changes how fast it plays them. That contributes to the character of the device. At the time, they chose the most straight forward way to accomplish that, and it was the least computational. The result was artifacts that were sometimes undesirable but seen as a reasonable tradeoff. In hindsight, they actually were at the core of the soul of these machines. And I think that’s a big difference with modern instruments. We have lots of memory and processing power, but we have lost some of the beautiful messiness. Recovering that messiness, now, that’s an interesting challenge.”

Final thoughts

“So, absolutely, this machine had a transformative effect on me. There’s a great, positive community of people who still love and use all things Fairlight. I feel very privileged, not only being part of that community, but also to be able to give back in a way, by creating new devices that can extend the life and usefulness of these incredible machines.”