Demystifying Classic Synth Character
Fat, Lush, Warm and Full of Character
Subjective Terms vs Objective Observations
There is a general consensus that older, classic synthesizers from the 1960s, 1970s and 1980s have a certain "mystical character" to their sound. We're talking about synths like the MiniMoog, MemoryMoog, Yamaha CS80, Prophet 5, Oberheim OBX, Roland Jupiter 8, and others from the golden and silver ages of hardware synthesizers.
People use a variety of terms to describe this so called character, including "organic", "warmth", "lushness", "fatness", or other somewhat subjective terms. I set out on a mission to try and identify specific, objective factors that give classic synths the desireable "character" that they have. This article describes the four aspects that I have identified as the most important.
Four Key Factors of Classic Synth Character:
1. Voice-by-Voice Imperfections - The most underrated, and most important factor
2. VCF Filter Types - The factor that is most talked about and definitely has a substantial effect
3. Oscillator Waveshape - Unique characteristics of classic VCO waveshapes
4. VCO Harmonic Jitter - Instability in VCO circuitry that washes out upper harmonics
Part 1: Voice Imperfections
The most important (and underrated) factor of classic poly synth character is voice-by-voice imperfections and offsets. If you take any classic VCO poly synth from the 70s or 80s, and take measurements of electrical imperfections on a voice-by-voice basis and circuit-by-circuit basis, you will find that the voices vary with tuning offsets, filter and ADSR envelope performance characteristics. These variances / imperfections my be small, or they may be very large (looking at you MemoryMoog ;).
When you play a chord on one of these classic synths, you're not hearing each note perfectly in tune. Rather, each note in a chord will have small offsets, sometimes just a few cents each, but when playing a multiple oscillator, multiple voice chord, these small offsets "add up" to create a lot of subtle, natural phasing and detuning motion. These imperfections make the sound less sterile and perfect, and more organic. In addition, variances in the filter and envelope circuitry create additional variance on a voice-by-voice basis. This is perhaps the most consequential factor that differentiates classic poly synths from their more modern counterparts.
As I researched the information for this article, it became apparent that that voice-by-voice imperfections was a subject that could take up a whole article on itself. For more information on this topic, and how to acheive "Voice Component Modeling" with modern synthesizers, check out this separate article, which includes in-depth analysis on the topic, and specific tricks to model classic synths using advanced modulation techniques with modern synths.
More info: Voice Component Modeling
Part 2: VCF Filter Types
The most commonly debated topic in terms of classic synth character is which kind of filters are used. If you search through the trenches of internet forums, you'll find hundreds of discussions about VCF filter types and implementations.
The filter circuitry is absolutely a huge part of the sound of a synthesizer - it is the core circuitry that shapes the raw sound generation of the oscillators, and spectral audio output. There are many different filter types that each have their own signature sound.
Some of the most popular analog filter types are: Moog-style Transistor Ladder, Prophet-style OTA Filter, Oberheim SEM-style State Variable Filter, Roland-style Acid TB Diode Ladder, and Korg-style Sallen Key Filter.
Part 3: Oscillator Waveshape
If you compare the sawtooth waves from a variety of different synths, you may notice that there is some variation in waveshapes. You can ignore the vertical polarity, as the sawtooth will sound the same with either polarity. And to compare apples to apples, a reverse saw wave is not the same as a forward saw wave.
Now, if you compare the actual waveshape of two saws, you may notice differences in the ramp shape (it may be more concave or more convex), and near the wavefront, you may notice a perfect sharp transition, or there may be rounding, smoothing, or extra transient spikes at the transition. These variances may be small and nearly indecipherable at the fundamental frequency, but sawtooth waves produce a rich series of harmonics, and the shape of the fundamental sawtooth form affects the rendering of the harmonic series. This can have a significant effect on overall tone, especially if using a resonant filter and sweeping through the harmonic range.
More info: Synth Oscillator Waveshape Variance
Part 4: VCO Harmonic Jitter
The second area that I focused on while researching this article was taking measurements of raw sawtooth waves from a variety of synthesizers - VCO synths, DCO synths, and digital/software synths. There was one striking thing that I noticed: Classic synths, which all feature VCO oscillator circuits, exhibit a harmonic frequency jitter in the upper harmonics of their timbre.
If you use spectral analysis to measure a raw sawtooth oscillator, with its rich series of harmonics, you'll notice that at the fundamental frequency, the frequency jitter is nearly indecipherable, but once you get up into the 2khz, 5khz, 10khz range and above, there is a noticeable amound of frequency modulation happening, that has the effect of washing out some of that higher harmonic content. The effect is subtle, because of amplitude dropoff in higher harmonics, but I believe this effect may be a key factor into why many people say classic synths have a "warmer character" than their modern DCO and Digital Synth counterparts.
DCO, Digital and Software Synths generally do not exhibit this type of harmonic jitter, however it can be modeled with an LFO.
More info: Synth VCO Harmonic Jitter / Phase Jitter
These subjective to objective conversions are my general take -- how I think of things, or try to interpret others descriptions:
Warm/Dark = 4-Pole LP Filter, less upper harmonics
Fizzy/Bright = 2-pole LP Filter, more upper harmonics
Harsh/Cold/Metallic/Sterile = Little to no voice/component variances. More technically accurate tunings with perfect envelope and lfo timings.
Lush/Organic/Alive = Per voice/component variances with natural phasing / temporal offsets
Full = Saw osc
Hollow = Square osc
Pure = Tri/Sine osc