Category Archives: vintage

Synderella Stories – Yamaha DX7

I hesitated before classifying the famous (infamous?) Yamaha DX7 as a “Synderella” synth, that is to say, an underdog. In fact, its sound dominated the 1980s, and it held on to the title of best-selling synthesizer ever until that was eclipsed by the Korg M1 workstation in 1988.

Here’s why I decided in favor. The DX7 is often maligned for sounding harsh and strident, not to mention being difficult to program. This was due to a number of factors: PCM sample-based synthesizers (such as that M1) becoming more affordable beginning in the late ’80s, the virtual analog and then the real analog renaissance, and the return of knob-per-function interfaces on keyboards. And perhaps the difficult-to-program criticism was not entirely unfair.

Yamaha DX7 synthesizer. Image via Wikimedia Commons.

Yamaha DX7 synthesizer. Image via Wikimedia Commons.

The thing is, I still have mine. Not a DX7-II, but the original that didn’t even have stereo outputs. It was the second real synth I ever bought, one summer after I acquired my beloved Korg Poly-800. As with the Korg, I learned everything I could about programming the DX7 to get the maximum sonic flexibility out of it. Among other things, I got some very serviceable B3-style organs (if paired with an external Leslie simulator), and some pads, leads, and strings that sounded surprisingly analog. Of course, there were the usual things it was good at due to FM synthesis being able to produce non-contiguous harmonic spectra in a way subtractive synthesis couldn’t: electric pianos, that infamous marimba, a shockingly good harmonica, and more. Suffice to say, if you had a DX7 plus a more traditional analog synth, you had the bases covered for gigging throughout most of the ’80s. 

Okay, so it did have some quirks. The output had a hiss that was audibly louder than my supposedly cheapo Poly-800. The keyboard, while still one of my favorite semi-weighted synth actions in terms of speed and feel, was on the noisy side, physically speaking. It was also scaled to sense MIDI velocity from 0 to 100, as compared with the standard maximum of 127. This meant if you drove a different MIDI keyboard with it, you wouldn’t get maximum velocity or timbre out of its notes. If you drove it with another MIDI keyboard, and played hard, the sound engine would sense those higher velocities and create some surprising and clangorous tones. 

On a current trip to visit family in Vermont, where my DX7 still sits on a dresser in my old high-school bedroom, I powered it up to find that many of the sounds I’d programmed were intact, either in internal memory or on a RAM cartridge. (Oddly, the internal battery still reads a nominal 2.8 volts, as though the thing has been in some kind of suspended animation.) So I figured, what the heck, post some audio examples here! These are all ones I’d programmed or heavily edited myself, and I tried to pick ones that are unlike the stereotypical DX7 sounds. 

ANABRSST is halfway between a string and brass sound, and I was going for as analog-sounding as possible. I used this for comping on Kool and the Gang’s “Get Down On It.”

ANALOGSTR was meant to be an all-purpose ballad string patch.

 With CHIKLEAD, I was going for that wormy lead sound Chick Corea is known for. The song that inspired this was “Flight of the Newborn” off Return To Forever’s album No Mystery.

I couldn’t afford a Fairlight or even an Emulator (who could?), but I really wanted that marcato strings sound reminiscent of “Cloudbusting” or “Under Ice” by Kate Bush. 

Okay, FULLDRWBR sounds much better if you add some Leslie simulation or even just a chorus pedal, but the harmonic balance is close to a full-registration B3 patch. This used algorithm 32, which just placed all six operators side-by-side in additive fashion, like drawbars.

FUNKEDOUT was a cool comping patch I found worked well for the synth intro to “New World Man” by Rush. Sadly, the part where I had velocity mapped to FM (to produce brighter harmonics as I played harder), seems to have been lost in this example.

Trying to cop that “Whiter Shade” ballad organ, and it doesn’t sound too bad. ‘Nuff said.

Factory sounds in the DX7 had plenty of solo saxes and trumpets, which could get very expressive if you used the BC-1 breath controller (I never did), but I wanted a big, funk brass stack for T.O.P. and EW.F. covers. TWRBRASS was the result. 

Sort of French horn-ish, sort of string-ish, and dark to offset the DX7’s rep for being too bright and bell-like. That’s WARMPAD. 

The FM synthesis pioneered by Dr. John Chowning of Stanford and first widely commercially applied in the DX synthesizers is still going strong today, and much more sophisticated and flexible, in instruments like Native Instruments FM8. The Yamaha Montage synthesizer also has a full, eight-operator FM engine that lives next door to its sample-based engine. 

Stay tuned for more Synderella stories! 

Synderella Stories – Kurzweil VA-1

Synderella Stories is our column about underdog synths that were under-appreciated for their time and/or are not now remembered with the fondness they might really deserve. In this installment, that’s the case because the Kurzweil VA-1 virtual analog synth, first demo’ed publicly in 2004, never made it to market. I know because I was writing the instruction manual at the time the project was shelved — in spite of the instrument attracting huge buzz at trade shows such as NAMM and Musikmesse, and generally being one of the most anticipated new synths of the early to mid-2000s.

A search of forum posts and vintage synth sites shows little information but more than a little misinformation. As someone who was there and working closely with Kurzweil’s stateside engineers (and the owner of one of a handful of prototypes left in existence) I’d like to take a little time to talk about the synth itself, its incredible sound and power, and the real story of why it never saw light outside of a convention center floor.

In the late ’90s and early 2000s, knobs were finally making a comeback after the digital user interface desert of the previous decade. The real-analog renaissance — marked by the introduction of Dave Smith’s Evolver and Bob Moog’s Voyager — hadn’t begun in full yet, though. If you didn’t own vintage synths, and wanted polyphony and lots of real-time controls at anything like an affordable price. virtual analog was the name of the game.

The first Nord Lead (1994) is widely recognized as having kicked off this category, and other notable (and still desirable) entrants included the Waldorf Q, Korg Z1 (which also did a fair amount of cool physical modeling stuff), and Yamaha AN-1X. In the early 2000s, Kurzweil — whose K2600 already enjoyed the reputation of Rolls-Royce of keyboard workstations — decided to up the ante and embark on an argument-ending ultimate virtual analog synth based on their then-new proprietary DSP chip, nicknamed CLARA. 

I should mention that the VA-1 was entirely the brainchild of Kurzweil’s stateside think tank, the Young Chang Research and Development Institute (YCRDI). Located just outside of Boston in a region often called the Silicon Valley of the East Coast, they continue to work on cool stuff I can’t tell you about yet. But that’s another article.

The VA-1 was to be a 16-voice polyphonic synth with three oscillators per voice and four-part multi-timbral capability. Its oscillators were modeled and in their final form were supposed to be able to morph from one waveform to another. In addition to the oscillators, the mix in a sound program could include two DSP processors (these provided such functions as sample-and-hold and ring modulation), noise, and external audio — for a total of seven sound sources. FM and hard sync were also supported.

Dual filters could be arranged in series or parallel, and covered all the expected types as well as both 12dB and 24dB-per-octave slopes. (Two-pole and four-pole by another name.)

The modulation matrix was really something special, with up to six sources (three of them assignable) per destination. You selected a destination, usually right on the nearly one-knob-per-function panel, then chose the sources you wanted to modulate it. The process was very quick and intuitive. In addition to the wheels and joystick, sources could include the three LFOs, two ADSR envelopes, and two ASR envelopes.

Effects were based on the high-end KSP-8 rack mount unit, which at the time was earning a place next to the likes of Eventide and Lexicon for its sound quality. On top of this, there was a 48-band vocoder, not one but two XLR combo jacks for audio input, 24-bit optical digital out, and “roll bars” on the back to protect plugged-in cables in the event the synth was tilted rearwards. YCRDI was really going for a fully pro, no-holds-barred, synth nerd’s dream at a price point between two and three grand.

You can surf the web and get more specs still, but that doesn’t capture what captivated me and just about anyone who got to hear one in person: The analog authenticity and sound quality were off the charts, and would be so even by today’s higher standards — in my opinion even standing up next to the excellent real analog polysynths available now. The VA-1 was particularly excellent at Oberheim-like sounds, but its palette was very broad. The only thing I recall us not talking about at the time was wavetable synthesis, but the DSP power to do so was there and that could well have been part of a firmware update.

Even with about half the functions working, which was the case with my prototype as the firmware was a work in progress, the thing sounded completely killer and still does.

So what happened? As I was about halfway into writing chapter 4 of the manual and pleased with the positive feedback I was getting from the mothership, I got a panicked call from the head of the engineering team. “Stephen, we love what you’re doing,” he said. “But stop all work on the manual. We’ll make sure you get paid for what you’ve already done.” (Props: They did.)

I’ve read online speculation that perhaps Kurzweil saw the writing on the wall with real analog coming back (given the success of the Evolver and Voyager) and decided not to do a dedicated VA but rather incorporate the technology into their next series of workstations. The latter did eventually happen (keep reading) but not for that reason.

South Korean piano maker Samick, known for inexpensive acoustic and digital pianos, bought Kurzweil’s parent company Young Chang in 2004, or rather they tried to. Samick was interested only in exploiting CLARA and other tech developed at YCRDI for the home digital piano market, and would have killed a lot more than the VA-1 had the buyout proceeded. During the worst of it, engineers came to work to find offices locked, files removed, and computers gutted and thrown on the floor in a pile. Of course there was a “red wedding” of layoffs. Some had the foresight to safeguard their most important work offsite — work that eventually directly figured in to current Kurzweil pro keyboards.

In September 2004, the South Korean government disallowed the buyout, citing anti-monopoly statutes. This left Young Chang with the hope of finding less rapacious overlords, but also bankrupt. Miraculously, YCRDI kept the lights on and a skeleton crew of engineers kept their gigs and forged ahead. But just about every pro keyboardist I knew was freaking out that Kurzweil would soon be no more.

Fortunately, this was not to happen, as in spring 2006 Hyundai Development Company acquired Young Chang. (Not the car company per se; HDC is the ultimate parent of the car company but also of a bunch of other things in heavy industry and shipping.) While their largesse hardly compared to Prince Charming’s, they did prove committed to pro keyboard development, and the brain trust at YCRDI began to rebuild. By 2007, Kurzweil was showing their first new workstation in years: the PC3.

By then, market conditions had shifted and it was indeed too late for the VA-1 as a stand-alone instrument. But its story is truly the Cinderella story of the company, which went from glory to destitution under a “wicked stepmother,” then rose again. The VA-1 lives on as the KVA analog modeling mode in the PC3, PC3K, and Forte instruments.

My VA-1 prototype? I still have it, though it’s currently more pumpkin than coach. This is because of the incomplete firmware, not any defect. MIDI isn’t working, some of the knobs adjust their parameters very slowly, and so forth. But the core sound that turned heads is all there, and I plan to get it running as smoothly as I can for you and post what will surely be the first video demo in over a decade. For now, check out one of the only videos on the searchable web, of sound design wizard Dave Weiser demonstrating one at Musikmesse 2004, courtesy of synth blogger Matrixsynth. Yes, just like its subject, it’s gone sideways.

 

 

Synderella Stories – Korg Poly-800

Welcome to Synderella Stories, a new blog about synthesizers that were (A) not very popular when they were new, and/or (B) not especially in demand on the used market today. However, (C) they were and are much more powerful, useful, or interesting than they got credit for.

First up is the first actual synth I ever owned, the Korg Poly-800 I purchased during the summer of 1984 (note: photo above depicts a different unit). Money from a lot of teenage odd jobs, combined with a “matching grant” from my grandparents to reward my diligence in saving, allowed me to buy the first programmable polyphonic synth priced under $1000 ($795 list). As I recall I was able to get it for around $700 out the door.

Critics blast the Poly-800 for having cheap plastic construction (true) a basic, limited signal path (also true), and sparse “data entry” driven user interface (yup). However, every synth player remembers that when an instrument is your first and only and you poured what then seemed like all the money in the world into it, you programmed the heck out of it to squeeze out every last drop of tonal flexibility (I did).

On the Poly-800, you got eight notes of polyphony with a single oscillator per voice, or four notes with two oscillators per voice. There were 64 memory locations. The oscillators were … weird. You had a choice of sawtooth or square waves, but the machine pretty much built the sawtooth by stacking up squares. Then, the DCOs had four square-wave harmonics labeled like organ footages: 16, 8, 4, and 2. You could turn these on and off separately for each DCO, and in combination with DCO 2’s interval parameter, thereby create interesting stacks of tonics with, say, thirds or fifths.

1983 Keyboard magazine ad for Korg Poly-800

1983 Keyboard magazine ad for Korg Poly-800

While this may have been a design workaround to make the 800 sound like more of a synth than it really was, it worked, and I was able to use it to great effect. It was great for nailing Prince-like pads and comping sounds, and turned my bandmates’ heads when I played it on covers of “1999” and “Let’s Go Crazy.” I have to say as well: the resonant VCF sounded great.

The ultra-simple step sequencer could hold one pattern at a time, and achieved a shockingly realistic signature arpeggio for The Who’s “Baba O’Reilly.” In fact, when the Korg Radias virtual analog synth came out in 2006, it featured a custom patch for this called “Hooz Next,” and I marveled at how close my old Poly-800 got.

Last but not least, it had MIDI, stereo outputs, and strap pins so you could sling it like a keytar. It fell to such a low price in the 1990s ($200 or less) that it became popular with modders and circuit-benders. If you see a unit with non-original knobs above the joystick, like this …

… it’s been hacked to provide real-time control over the filter cutoff and resonance. Later, the EX-800 keyboardless module would come out, then the Poly-800 Mk. II., which added effects. But for me, there’s never been anything quite as wonderfully odd and surprisingly useful as the original.

Synderella needs her glass slipper of course, and today that comes in the form of rising prices on the used market. US $349 seems about the median, but I’ve seen the limited-edition model with reverse-color keys go for over $600.