haha - I have to say that I don’t have any great tips. Use a filter that distorts a bit, in a “mini moog” kind of way (Voyetra-8 and CAT SRM-2 are two of the very few to use the SSM 2044, which was cheap and great sounding). After the waveforms are mixed they go to a VCA that was close to saturating. I think due to a math error the ADSR (first digital adsr in an analog polysynth?) they might be inverted exponential instead of exponential like they should have been (not sure). Also the relatively low level of linear FM between the two VCOs was uses a lot. It wasn’t TZFM, so it’s only in there in small amounts.
BTW, “back in the day” we repaired all kinds of synthesizers. But whenever a CS-80 came in everyone wanted to play it. That and The Strangler’s Oberhiem 8 voice (before the OB-X series, the one that had 4 SEMs in it.). All the other synths - no one cared.
One thing I would change in this patch is the oscillator section. The CS-80 Saw has a peculiar shape. There is a notch at each start of the Saw cycle. Sometimes this is described and implemented as a downward notch, sometimes as a upward notch. And also of different width, from just a very narrow spike to width of several percents. On the web you can find various descriptions and some oscilloscope wave shape pics.
I never found the definitive ‘reference’ shape. Might be that the Saw shape differed per actual physical machine.
Anyway, such a notch can be implemented by mixing a very narrow Pulse with the Saw. Relative amplitude and the width of the pulse can be varied to get the desired shape.
The Surge Wave table VCO has a CS-80 Saw wavetable. In that wavetable the amplitude of the notch (Pulse) is lower then that of the Saw.
As said, others show the notch (Pulse) at a higher amplitude then the Saw.
The CS series VCO is a sawtooth core oscillator which is the most common variety of VCO. One oddity of this particular VCO is that it isn’t a perfect sawtooth, it has a small square pulse at the start of each cycle. Above the IC you can see a silver cylinder – a polystyrene capacitor, it’s the VCO’s integrating capacitor, effectively the electrical “bucket” that fills with voltage until it reaches a certain capacity at which time a transistor short circuits the capacitor, dumping out the voltage and the filling begins again, creating the next cycle of the sawtooth waveform. Yamaha uses a linear CV standard (Hz/Volt) as do Korg who reportedly invented and patented this CV system. Most other synths use a logarithmic system known as “Volt/octave”.
There are plenty of examples of Yamaha CS-80 excelling under the amazing hands of Vangelis.
Here’s a pretty unique and brilliant example of the CS-80 performed by…Eric Persing.
Wikipedia on Eric Persing
Eric Persing Persing started working for the Roland Corporation as Chief Sound Designer[3] from 1984 to 2004,[4][5] where he worked on many influential synthesizers and music-related products such as the Roland D-50,[6] the JD-800, the Roland JX, JV, JP, XP series synthesizers and many others. Even today, his sounds can be heard in many productions.
Eric Persing has had has had a huge impact on sound (and synth) design. Both with Roland and Spectrasonics. Huge numbers of his presets live on in many synths and music.
Short tangential sidestep…
Among these presets is arguably one of the greatest (and overused) patches/presets of all time. The Roland D50 Preset 37 Soundtrack.
Two examples of the D50 Soundtrack preset.
Dr. Mix - D50 Soundtrack demo
Woody from the Piano Shack - D50 Soundtrack demo
Back on track…to the mighty Yamaha CS-80.
Here Eric Persing is collaborating with Pedro Eustache. Eric is playing several synths, but notably also a Yamaha CS-80 (e.g. from 6:45 onward). It’s the huge beast with the brightly colored buttons…
EP-PE One is an extraordinary musical collaboration between keyboardist Eric Persing and multi-winds maestro Pedro Eustache. The two acclaimed artists have never played together before and this video documents their first musical encounter.
One of the defining CS80 sounds is it’s filters. Each oscillator has one filter, which has a gentle “prophet-like” resonance. I’d try a Tangents filter by Vult for each oscillator (also Vult, try the Bleak oscillator). Use a gentle envelope setting and try to patch in expression control.
Sure, as with all subtrative synths, the filters are an important factor is it sonic character. As are other circuit components like the vca’s and ‘noisy’ mixer.
BTW. Each layer (there are 2 per voice) has 2 filters. A High Pass and a Low Pass. Both 12 dB/oct.
For geeky technical nitty gritty…check out this ModWiggler thread. Probably enough info to give an apt c++/DSP developer a sporting chance to emulate it (in VCV rack?).
Interesting tidbit. I’ll look into it because that could make make a difference in the sound. Getting the right mix will require a bit of scope time. I mean, it has the feel of it, having only Cherry Audio’s version as reference and having heard the real version in recordings. I in fact, used the Cherry Audio manual and the CS-80 manual to check out how things worked.
I was going to use Tangents, my goto filter, but I couldn’t figure out if it was 12dB/Octave. Turns out I think it is. I might try swapping it out in a copy and see if it’s an improvement.
I won’t be swapping out the VCO, though, as the EvenVCO has the ability to directly modulate the frequency and outputs all the waveforms I need. There is a limit to how many modules I can get in a patch before VCV gets cranky. Adding the Patchmaster UI was a noticeable increase on DSP usage.
There are actually 4 filters per voice - a HP on the pulse and a BP on the Saw before it goes to the VCF’s two filters. (Update: no, the 4 filters were on the GX1. CS80 had two per voice)
I’ve started playing around with some ideas in PureData and have actually created a working IL AL ADR filter envelope. I may or may not continue, but I’m using a version of PureData that runs standalone or as a VST. Works in Bitwig, but struggling with the fact that PureData expects midi note numbers, not v/oct. So it’s fun to play around with at a more nuts and bolts level.
I do have one suggestion for a change – the velocity filter modulation – VCV sets the velocity and leaves it at the level of the last note. I realized by putting an EG on the velocity before it goes into the filter modulation chain that it would eliminate that. Sharp attack, no decay, full sustain, very short to no release time.
FWIW I accidentally made a really Bladerunner style sound in v1.6 using the WTO wavetable VCO through Vult Tangents. Sadly that’s one of the modules that never made it to v2!
