S+S actually stands for Sample and Synthesis, but I prefer that it’s really based on old Hercules movies. The first S+S device was the Roland D50 in 1987, and it’s back in a super-deluxe form in the ‘new’ ZenCore of 2020. Quite the evolution!
The reason I became interested in S+S was reading people bickering over this versus that keyboard and wondering what made them different. Over a few months of idle reading it became obvious that the majority of Roland’s synthesisers are nearly the same guts over and over again, simply tossed in different ways. And to me that’s like being backstage at a magic show.
Hercules: The D50
Each D50 voice is two waveforms. In most cases the first waveform is the initial transient of a sampled sound, and the second is a sustaining sample loop. For example, for a piano the first waveform is a very short hammer strike and the second is a loop of piano strings. The idea was to fit as much expression as possible into the tiny expensive sample memory of the time. It works because the initial transient sound is the strongest identifier of the sound, and the listener accepts that it’s the familiar noise.
There is also the option to have one or both of the waveforms be a synthesised square wave and filter. The two tones (waveforms and treatments) are then added, or multiplied as ring modulation. Cryptically, the manual says a saw wave is made by processing the square wave at the filter, which could mean that the saw is really a rounded square – it does sound like it. The maths for the filter pretends to be a low pass analogue with resonance although I’m told it’s something to the left of that.
At the time Roland were busy kicking Yamaha’s DX7 in the nuts, and so the emphasis is all about adding and multiplying tones in linear arithmetic synthesis with subtraction kept hidden up the back of the warehouse. How times change.
Two of these voices can then be stacked or split across the keyboard. In many cases you’ll now have four tones in two voices sounding on each note, although they have yet to be wedded, as they will in the later models. Four tones becomes the Roland way of life.
Son of Hercules: The D70
Not long after Roland tried to do something tricky with the D70 called super linear arithmetic synthesis (the word ‘super’ in Rolandese just means ‘upgraded’). The four tones are now all samples, and there is a filter for each. Although there’s upper and lower pairs they really act like 4 separate signals.
The ‘super’ bit is Differential Loop Modulation, which is a weird thing.
DLM seems to have been shit, dooming the D70 to instant oblivion. Along with an LFO that actually slows down as you play more keys – a prime example of Roland, Spend The Five Bucks you will see throughout the years as petty cost cutting screws up. Roland stopped trying to be tricky.
Hercules Versus Rompler: everything for years running.
By the time we hit the JV’s in the late 80’s/early 90’s there’s been a lot of changes under the hood. ‘LA synthesis’ is dead. Now the four tones are arranged in pairs. Each pair can be used as it was in the D50, but memory has come down in price and the samples are longer, so that you’re less in need of a chiff up your front. You can control the voice load by turning individual tones on or off.
The JV80 introduced plugin sound cards such as the very desirable ‘Vintage Synth” that adds a good variety of Moog, Oberheim and other oscillators to the mix. But it was the JV1080 rack with multiple cards installed inside that gave Roland a sugar hit they found hard to escape for the next decade – the JV1080 to the XV5080 (at the time we didn’t use 4’s in Japan) were the business for years on end. The one box could make synthesiser sounds and near natural acoustic sounds and does both rather well.
The tone pairs are joined in a structure. The basic structure is each tone with a sample, multi mode filter, and amp. The second adds both waveforms into both filters in series. The third structure has the first waveform overdriving the second into a hard clip which causes a nasty kind of wave shaping. The last structure uses ring modulation to multiply the tones. Understanding how to use these is all a bit trial and error. Here’s a sensible guide.
The filters are now lowpass, highpass, bandpass or peak, which is the opposite of bandpass. The filter on the Fantom XR is an odd one – it wants to be vowel more than any other Roland I’ve used.
Hercules, Space Man: SH-32
We’re left wondering – just which Roland keyboards are the same guts used in different recipes? From my own listening I’m pretty sure the JP8080 isn’t sword and sandals, but I’m very sure the SH-32 is. Even though you see twiddly knobs, and Roland touts some fancy-pants new synthesises technique, the same oscillator noises are passed through the same filters with the same modulations and effects. For some fool reason Roland has identified the samples on the SH-32 by numbers, but they are obviously the old samples from the D50, JP8 et al. that were there on the JV80 – I am a nerd and I have proved it. I’ve the sad suspicion that the piano and oboe samples are in there somewhere, tied up and blindfolded.
The Fantoms were S+S in their infancy but later took on Supernatural
and now ZenCore. This story to be continued…
