KVR Audio

Leslie Sanford · Post by **Leslie Sanford** » Sat Jun 20, 2009 4:15 am

I was wondering if anyone has had experience with this synth. It's a 4-op FM synth like the DX7's smaller brothers, but it wraps this architecture in a VA like interface.

I've been searching for info, and I found Sealed's page, which has a lot of useful information, but doesn't answer all of my questions.

Specifically, I'm interested in how it wrapped up the FM engine underneath to make it look like a VA. According to Sealed's page, it used algorithms 1 and 5 from the FB01. That's helpful, but how did it set the ratio between the operators? Was this hardwired, or could you change the ratio settings?

Any info will be appreciated. Thanks.

bstageboss · Post by **bstageboss** » Sat Jun 20, 2009 6:25 am

Hi Leslie,

at one point in time I had a Korg 707, which was the same synth engine minus the effect section. You're correct about the programming; they did put a sort of analog spin on the FM engine.
To the best of my recollection, the translation for the basic sound parameters was as follows:

OSC 1 Pitch = Operator 1 Ratio
OSC 1 Amp = Operator 1 Output Level
OSC 1 Timbre = Operator 2 Output Level
Waveform/Spectrum = Operator 2 Ratio

OSC 2 Pitch = Operator 3 Ratio
OSC 2 Amp = Operator 3 Output Level
OSC 2 Timbre = Operator 4 Output Level
Waveform/Spectrum = Operator 4 Ratio
Brightness = Feedback level (OP4) (just on/off or 0/7)
Xmod Off/on = Algorithm 1/Algorithm 5

There was a Detune parameter affecting OSC 2 (or operators 3 and 4 equally), as well as Amplitude and Timbre Envelopes for both OSC 1 and 2, and an LFO that had a Pitch and a Timbre/Amplitude level. Most of the parameter values were exactly the same as in the FB-01, such as the Timbre, which would go from 00 to 99. Classic 4-OP FM architecture, different name.

There's a decent shot of the DS-8 parameter list on the panel here:

http://www.flickr.com/photos/matrixsynth/420802757/

The thing did get some interesting sounds, since at that time there were a few programmers that had figured out FM's strengths. I once managed to get my 707 to sound like that DX7 bass patch on Kenny Loggins' "Danger Zone" (from Top Gun). It even managed to get close to some D-50 patches, such as "Soundtrack" and "Fantasia". (That was a big deal back then!) The factory patches were completely new, although I was able to duplicate many of them on a DX21 later.

One thing Korg did to up the ante over the other 4-OP Yamahas was to add a digital delay line that could also do modulation effects like chorus and flanger. The delay was technically in mono but split to left and right with one of the channels out of phase to give a pseudo stereo feel to the otherwise mono delay line. (Kind of a neat trick actually; I only found out when I panned both outputs to the center on a mixing console and the delay all but disappeared. My DW-8000 and my Poly 800 mkII also had the same type of delay circuitry.)

Hope this provides enough information for you,

STV

Leslie Sanford · Post by **Leslie Sanford** » Sat Jun 20, 2009 6:43 am

bstageboss wrote: To the best of my recollection, the translation for the basic sound parameters was as follows:

OSC 1 Pitch = Operator 1 Ratio
OSC 1 Amp = Operator 1 Output Level
OSC 1 Timbre = Operator 2 Output Level
Waveform/Spectrum = Operator 2 Ratio

OSC 2 Pitch = Operator 3 Ratio
OSC 2 Amp = Operator 3 Output Level
OSC 2 Timbre = Operator 4 Output Level
Waveform/Spectrum = Operator 4 Ratio
Brightness = Feedback level (OP4) (just on/off or 0/7)
Xmod Off/on = Algorithm 1/Algorithm 5

This is very interesting and helpful. I'm starting to get a good picture of what was going on under the hood of the DS-8/707.

I took at look at the instruction manual, and it showed the waveform selections, I think they were sawtooth and square variations. This makes sense given the limitations of the algorithms. My FM foo is rusty, but if I recall a 1:1 ration will give you a saw-like waveform, and a ratio of 1:2 gives you a square like waveform. The modulation levels determining the brightness of the waveforms.

Hope this provides enough information for you

Absolutely. I enjoyed reading your entire post. I learned a lot.

bstageboss · Post by **bstageboss** » Sat Jun 20, 2009 4:36 pm

Just took a look at the manual and found out I made an error about the feedback situation. Turns out that OSC 1 actually has 4 Waveform types: Saw, Square, Bright Saw and Bright Square.

Saw = Basic C/M ratio is 1/1
Square = Basic C/M ratio is 1/2
Bright Saw = Basic C/M ratio is 1/1 and modulator feedback is maxed (at 7, if you're converting from the Yamaha FB-01)
Bright Square = Basic C/M ratio is 1/2 and modulator feedback is maxed

What this essentially means is that Oscillator 1 actually corresponds to the operator pair 3 and 4 on the FB-01 and OSC 2 is actually ops 1 and 2.

The Ring Modulation parameter is technically another modulator ratio control (possibly the fine ratio; I'm a little fuzzy on that one), where as the waveform selection just controls a 1 or 2 multiplier for the modulator (so in essence using the saw wave and 1 increment in the spectrum parameter is the same as using the square wave). Note that you have to use one of the saw waves if you want your modulator to have an odd number for the ratio, say 1/7 for that classic DX7 marimba sound. With the square wave your C/M ratios will be 1/2, 1/4, 1/6 and so forth.

My guess is that Korg licensed the 4-op engine from Yamaha, but had to agree to limit some of the parameters (only 2 algorithms, Feedback either off or full on) so they wouldn't compete with the DX/TX lines.
Korg's idea was to make FM programming easier to grasp for people coming from a subtractive background, but it ended up making the simple FB-01 structure waay to complicated.

Cool sounding synth though. Now if only FM8 could get their LFO speed to match that of the 4-op Yamaha synths, all my FM needs would be met completely.

STV

By the way, I like your Cobalt synth; looks like I'm gonna have to pull $25 bucks together from somewhere...(yes, I miss my DW-8000)javascript:emoticon('

')

foosnark · Post by **foosnark** » Sat Jun 20, 2009 6:45 pm

Beat me to the answer there. I have a DS-8 sitting around under my bed that I don't play anymore, but I loved that thing in its day.

Leslie Sanford · Post by **Leslie Sanford** » Sat Jun 20, 2009 6:49 pm

bstageboss wrote:Just took a look at the manual and found out I made an error about the feedback situation. Turns out that OSC 1 actually has 4 Waveform types: Saw, Square, Bright Saw and Bright Square.

Saw = Basic C/M ratio is 1/1
Square = Basic C/M ratio is 1/2
Bright Saw = Basic C/M ratio is 1/1 and modulator feedback is maxed (at 7, if you're converting from the Yamaha FB-01)
Bright Square = Basic C/M ratio is 1/2 and modulator feedback is maxed

What this essentially means is that Oscillator 1 actually corresponds to the operator pair 3 and 4 on the FB-01 and OSC 2 is actually ops 1 and 2.

The manual says that in XMODE you don't hear the output of OSC1. So that makes sense.

The Ring Modulation parameter is technically another modulator ratio control (possibly the fine ratio; I'm a little fuzzy on that one), where as the waveform selection just controls a 1 or 2 multiplier for the modulator (so in essence using the saw wave and 1 increment in the spectrum parameter is the same as using the square wave). Note that you have to use one of the saw waves if you want your modulator to have an odd number for the ratio, say 1/7 for that classic DX7 marimba sound. With the square wave your C/M ratios will be 1/2, 1/4, 1/6 and so forth.

This further clears things up. I was wondering if the DS-8 allowed for ratios greater tha 1/2 because in the demos I've listened to, you definitely hear sounds (particularly metallic ones) where the harmonics are spread out.

My guess is that Korg licensed the 4-op engine from Yamaha, but had to agree to limit some of the parameters (only 2 algorithms, Feedback either off or full on) so they wouldn't compete with the DX/TX lines.
Korg's idea was to make FM programming easier to grasp for people coming from a subtractive background, but it ended up making the simple FB-01 structure waay to complicated.

Maybe if I want to investigate this further, I'd be better off grabbing a decent FB-01 and using an editor to play with it? Basically, limiting myself to a subset of its parameters to simulate a DS-8.

I wouldn't mind getting my hands on a DS-8 or a 707, but I don't actually need another keyboard...

By the way, I like your Cobalt synth; looks like I'm gonna have to pull $25 bucks together from somewhere...(yes, I miss my DW-8000)javascript:emoticon('')

I'm glad you like Cobalt.

delgan · Post by **delgan** » Sat Jun 20, 2009 7:14 pm

I have fond memories of the Korg DS-8. The keyboard keys had a very good action (imho) and some of the sounds were very good. I've owned other FM instruments such as the Yamaha DX7, TX802, TX816, DX21, DX27, TX81Z, CX5M, but the Korg DS-8 still had some sounds that I could never achieve with the Yamahas.

bstageboss · Post by **bstageboss** » Sat Jun 20, 2009 7:51 pm

Leslie Sanford wrote:Maybe if I want to investigate this further, I'd be better off grabbing a decent FB-01 and using an editor to play with it? Basically, limiting myself to a subset of its parameters to simulate a DS-8.

A quick way to play with something very close would be to try out VOPM.

http://www.kvraudio.com/get/228.html

Not multi-timbral AFAIK, and of course no FX, but for basic sound design it should do the trick.

STV

Leslie Sanford · Post by **Leslie Sanford** » Sat Jun 20, 2009 8:32 pm

bstageboss wrote: A quick way to play with something very close would be to try out VOPM.

http://www.kvraudio.com/get/228.html

Thanks, I'm having fun playing with this.

Do you recall if there was a way with the DS-8 for the carrier to be greater than 1, i.e. a way to set up the c/m ration to something like 3/2, for example?

EDIT: Looking back over what you've written, I guess it would depend on whether the DS-8 scales up the pitch of the modulators when you adjust the pitch of the carriers. For example, if setting the pitch of OSC1 up an octave didn't affect the modulator and the ratio was originally 1:1, it would then become 2:1.