Roland Boutique is a new series of instruments that faithfully reproduce the sound and appearance of the historic JUPITER-8, JUNO-106, and JX-3P synthesizers in a compact form. In this interview with the lead engineers involved with creating the original instruments, Kazuhisa Takahashi, Hideki Izuchi, and Akira Matsui, the three legends share some untold stories about the original product development, and also offer their impressions on the Roland Boutique recreations. Mr. Takahashi and Mr. Izuchi are still active engineers at Roland, while Mr. Matsui has since retired. In the second half of the interview, Takeshi Tojo, Masato Ohnishi, and Hirotake Tohyama—developers of the Roland Boutique series—joined in the conversation.
What are your honest opinions regarding the Roland Boutique series? As developers of the original products, what were your impressions when you heard about the modern recreations?
Takahashi:I'd heard about this product series rather early on thanks to my position at the company, and my honest feeling about it was that I was excited. I was truly excited at the prospect of a simulation being released of a product that I was involved with decades ago. Synthesizers from those days have a unique feel that you can only get on analog, so I was interested in seeing how much of that could be reproduced in the digital realm. I knew that the Roland Boutique series development team was working on this very seriously. More than being intrigued, I was truly very excited.
Matsui:I'm no longer with the company, so I hadn't heard about this until very recently when someone mentioned, "We're working on this project relating to the JX-3P." Honestly, I was very happy to hear that. I was very happy to hear that Roland will be releasing a product that I had a part in developing in the past in a new form. That's because the JX-3P was a very special product for me.
Do you have any interesting memories relating to the JX-3P?
Matsui:That synth was developed at Roland's Matsumoto factory. I was working at Matsumoto at the time and was involved in guitar synth development. Guitar synths up to the GR-300 and GR-100 were based on converting pitch into a CV (Control Voltage) signal, which would then trigger an analog synth. But for the next generation of guitar synths, we decided to use a design where pitch would be converted to MIDI, which would then control the sound engine digitally (though the sound engine itself would still be analog). We studied how guitarists were using the GR-300 and GR-100 and realized that they didn't do much sound creation on their own. (Laughs.) So if that was the case, we decided to go with an easy-to-use synth with presets. And this was the start of the GR-700 project.
The [GR-700] would feature a built-in pitch-to-MIDI converter and a sound engine with presets. But even if this was to be a preset synth, it would still be too difficult to create sounds without any knob controls. So we made a programmer—solely for development purposes—with which we could adjust the tones. Sometime later in the development process, one of our superiors remarked, "You can attach a keyboard to that and make a polyphonic synth out of it, can't you?" [Laughs.] So that's how the JX-3P came about. The reason it's a six-voice polyphonic synth is because it started out as a sound engine for guitar synths, guitar being a six-string instrument. (Laughs.)
The GR-700 was released in 1984, and the JX-3P in 1983. So are you saying that development for the GR-700 had been started earlier, and the JX-3P was derived from that?
Matsui:Yes, that's right. The release of the GR-700 was delayed somewhat because the controller took a bit more time to develop. We decided to release the JX-3P, and we also decided to release the programmer that we used for development and named it the PG-200. Just to let you know, the "3P" in JX-3P comes from "Programmable Preset Polyphonic."
Can you share any memories of the JUNO-106?
Izuchi:We used a one-oscillator design for the JUNO series to reduce its price, but its sound naturally ended up being thinner than say the JUPITER-8 or JX-3P, which used two oscillators. So our main priority in developing the JUNO series was to produce thick and dense sounds with just one oscillator. To this end, we employed a variety of strategies, such as adding a chorus function and boosting the lows when the high-pass filter was not being applied. So the flat setting is actually just one increment up from zero on the high-pass filter. (Laughs.)
The JUNO-106 continues to command unshakeable popularity. The JUNO-60 is also highly regarded and sells for very high prices on the pre-owned market. Are the circuitries in the JUNO-106 and JUNO-60 quite different from each other?
Izuchi:A lot of money went into the JUNO-60. [Laughs.] We used a lot of premium parts. If memory serves, the JUNO-60 was priced at around 240K yen, whereas we slashed the price of the JUNO-106 by more than 100K to about 130K yen. With the JUNO-106, we just wanted to lower its price, whatever it took.
What about the JUNO-60 and JUNO-6?
Izuchi:Everything that related to sound in those two models was exactly the same. So the sounds that you got should've been the same. The JUNO-60 is essentially a JUNO-6 with an added patch memory feature.
Do you have any memories relating to the JUPITER-8?
Takahashi:The sheer scale of that synth made everything quite challenging. The circuit boards inside the unit were arranged in two layers, and there were lots of boards, although I don't remember how many in all. The number of parts that went into it was also quite large, and they all had to work perfectly. Since it was an analog unit, there were a lot of things that needed adjustments. You would adjust one part just right and that might put another part out of whack, so it was a lot of work. With that much packed into it, it was a heavy synth and it took two people to carry.
The JUPITER-8 cost 980K yen in those days. Did you have a target price when you were developing it?
Takahashi:The concept for that product was to make the best synth there was, so we didn't pay any attention to cost.
Matsui:Mr. Takahashi just touched on analog circuit adjustments, and that reminds me of when I had just joined Roland and was assigned to a section that was responsible for assembling and inspecting SYSTEM-700 synths. There were more than 100 semi-fixed resistors in the SYSTEM-700, and I remember working from morning to late in the evening making adjustments to them. (Laughs.) It was from this experience that I decided that, if I were to develop a synth, I would design one that didn't require all this onerous work. So with the JX-3P, we employed a digital design and minimized the number of parts that needed adjustments.
Izuchi:The JUNO series was the first to use digitally controlled oscillators, and the core part of that setup was the NEC μPD8253 chip. Actually, the JUNO wasn't the first product to use this chip. To be precise, the first time we used this chip was on an earlier product called the EP-09 (an electric piano). It worked pretty well for that product and so we decided to use it in synths as well.
Matsui:DCOs would provide stable tuning, but they were difficult to fine tune. DCOs in those days were 16-bit, and it was difficult to adjust subtle deviations in oscillating frequency. So it was a challenge to get fine tunings that sounded good. So after some brainstorming, we designed an analog-controllable master oscillator for the DCO on the JX-3P, and made fine tunings by adjusting this oscillator. So you could say that this was an analog/digital hybrid circuit. The oscillator is digitally controlled, but it retains some of that analog character and gives stable tuning.
The JX-3P was the first synth that supported the MIDI standard. Do you have any memories about that?
Matsui:When our JX-3P prototype was nearly finished, we conducted a verification experiment for MIDI with the people from Sequential Circuits in the U.S. I couldn't be there in the U.S., so other staff members went instead. I think it was the year before the 1983 Winter NAMM Show that the MIDI standard was announced. We connected two synths, our JX-3P prototype and a Sequential Circuits synth—I think it was the Prophet-600—using MIDI, and showed how you could play the keyboard on one synth to produce sound on the other. The audience burst into applause when they saw this. (Laughs.)
That said, although the note numbers were perfectly matched, the pitch bend didn't work right, and I remember receiving a phone call from a flustered staff member saying, "The pitch bend is going the opposite way!" [Laughs.] To that I remember saying something like, "We haven't settled on a protocol yet, so it doesn't matter whichever way. You decide." This became the template for the MIDI standard. Today, we can communicate easily by email, but in those days, we only had the telephone. On top of that, when you come down to details like pitch bending, you have to deal with expectations and preferences on both sides, so coordinating all of that was quite a chore. (Laughs.)
At Roland, are there times when veteran engineers teach know-how and techniques to younger engineers? Something like a transmission of skills?
Matsui:I can't speak for young people today, but I don't think we ever taught anything to our juniors in those days.
Izuchi:Yes, yes. We just made things that we wanted to make.
Takahashi:In those days, if I was involved in a project as a member of a team, I would just appropriate a lot of know-how from watching my seniors. I don't think that's changed much with young people today, so I think it's more like appropriating whatever know-how you want without actually being taught by your seniors.
These days, because of these kinds of products (simulations of classic products), I get asked a lot more questions from younger engineers. I don't see myself as being some kind of legend or anything, but I think the younger people refer to me as "Takahashi-san who made analog synths." [Laughs.] So I do give advice when I get asked questions in this area. I've also seen something similar with the AIRA TR-8, where the engineer who was involved in developing the original TR-808 would write a program based on the circuit diagram and give advice on how they could reproduce the original. If it weren't for that product, I don't think we would have had the opportunity to pass down the design of the TR-808 or TR-909 to younger engineers.
What about the new Roland Boutique series?
Tojo:This series is based on ACB (Analog Circuit Behavior) technology, so we modeled the circuitry of the originals, and took it as far as we could by ourselves first. But even then, we would come up against things that we couldn't figure out, so we would go speak with our seniors.
Izuchi:My advice was, "Make sure to get 'just this' right."
What did you mean by "just this"?
Izuchi:That would be [the JUNO-106's] chorus. I stressed, "Just make sure you get the chorus right." (Laughs.)
What are your impressions of the finished Roland Boutique series?
Izuchi:When I first saw [the JU-06], I thought that it was a neat gadget. But when I played its sound, I realized it was no gadget. It was like playing a real JUNO-106, and I was impressed with the degree of simulation. Although I don't know if it was by my insistence, the chorus was also reproduced exceptionally well. A few weeks ago, I decided that I wanted to hear its sounds more seriously, so I played it in a soundproof room and it truly blew me away. I was impressed with the ACB technology in that it could realize this degree of simulation just from the circuit drawing and an original unit.
Tohyama:It's been many years since the JUNO-106 was first released, so when we went through original [units], we would come up against individual differences [due to age]. It was difficult to decide on which of those to simulate.
Tojo:A BBD (a component for delaying the audio signal) can get noisier as years go by. The first original unit that we checked had a lot of noise, but another one that we tested later was better maintained and made less noise. So we adjusted the amount of noise to where it sounded good.
Matsui:I've only been trying out the JX-03 for a short while, but my impression is that it is well made. I will be taking it through its paces in the next little while.
To the developers of the original synths, do you have any questions or comments for the developers of the Roland Boutique series?
Matsui:We old-timers were just thrilled when we were making these products. [Laughs.] Excitement was the most important thing for us. Were you all excited when you were developing these [new] products?
Tojo:Yes, we were. I joined Roland because I loved Roland synthesizers and had a burning desire to make these kinds of products myself. So it was exciting to resurrect synths that I admired back then with my own hands. I had a lot of fun.
Ohnishi:Same with me. I had been involved in the development of only non-keyboard instruments up to this point since I joined Roland, and this project was truly a lot of fun as it was my first opportunity to be involved in the development of a keyboard instrument— one that was based on the JUNO-106, a synth that I used to own.
Tohyama:I get the feeling that, with these kinds of products, the more features you pack into them, the more they become like tools and less like musical instruments. All three of the synths that we based our development on have such powerful and unique characters, so all throughout the process, I had a greater feel than ever that I was making musical instruments.
Matsui:It goes without saying that you can't make a good musical instrument unless the person who's making it is excited about it.
Takahashi:Just like Mr. Izuchi mentioned earlier, in our days, we made products that we really wanted for ourselves. Even if we got chided by our superiors, we would secretly make what we wanted. [Laughs.] As a result, we were able to come up with products such as these that have garnered much acclaim. Of course, there were [also] those that ended up being commercial flops. And I think that's how it should be.
Matsui:At any rate, ACB is an amazing technology. When we were developing the JX-3P, we placed a great emphasis on the sound, so we spent a lot of time on tuning the hardware and software. We worked very hard to make the sounds in that synth, and now they can be easily reproduced in the digital realm. In a sense, I feel envious of the engineers today—or should I say, I wish I were still a middle manager at Roland today. (Laughs.)
Tojo:You say so, but there would be no JX-03 if you hadn't developed the JX-3P when you did.
Matsui:That's true. (Laughs.)