At first an alphabetical listing of all synthesis forms and devices. A short description gives an initial insight.
01 Additive Synthesis (Fourier Synthesis)
The spectrum has to be built overtone by overtone. Every part of the sound possesses amplitude, frequency and phase.
02 AFM >14
03 AI >35
04 AM (Amplitude modulation) synthesis >24
The velocity of a basic frequency is processed from outside, e.g. through another frequency. No synthesis, rather a procedure.
05 Analogue Synthesis >38
A hardware classifikation manipulating analogue data. Sound parameters are processed flowingly and quickly.
Chaotical prozesses und aberrances during the manufacturering give the analogue sound a “human touch”
Nothing has to be calculated since changes through the controler devices immediately affectuate a change in the sounds.
DSP (= Digital Sound Processing) imitates (simulates) the behavior of analogue circuits. Digital synthesizers are „physical models“ of analogue synthesizers!!
06 AWM >35
07 AWM 2 >35
08 Composite Synthesis
Artificial and sampled sounds are mixed .
Generic term of am mathematical procedure of which can be derived modulation, filtering, temporary processes (delay etc.).
10 Cross Modulation >34
11 Digital Signal Processing (> DSP) >13
12 Direct Digital Synthese
Digital data of waves are to be transformed by certain rules of handling (calculating)d
13 DSP (= Digital Signal Processing);
Works with digital data of sounds, e.g. samples. Essence of Software Synthesis.
14 FM (Frequency modulation); see too: AFM (Advanced FM).
A carrier wave can be manipulated by a governor (modulator). Take an oscillating modulator an you will get very complex overtones
very hard to prognosticate (> Bessel equations).
15 FMX >34
16 Fourier Synthesis >01
17 Formant Shaping >01
18 Formant Synthesis
FM Synthesis, a few algorithms and some filters.
19 Granular Synthesis
Many little samples are composed in a new manner. Differing to > 44 (Wave Sequencing) here are the samples that short that they are heard as sound colors, not as impulses.
20 iPD >28
21 Karplus Strong Synthesis >29
A short sound (impulse) or an explosiv noise initiates a delay unit with feed back. Given certain circumstances one can get a sound with defined pitch. That kind of synthesis is used to form plucked strings’ sound or other sounds with percussive quality.
22 Linear Arithmic Synthesis
Uses the assumption that the most important part of the sound is the A (attack) part.
23 Linear Predictive Coding (LPC)
Used in speech synthesis/analysis. Some re-synthesis stressing filtering to get the final sound.
24 Modulation Synthesis ;
coined by Curtis Roads in his „Computer Music Tutorial“ . See 02, 03, 10, 14 und 15.
25 Multiwave Synthesis
Has crept in, has nothing to do with our theme, nevertheless existing, even using laser technology. The most interesting thing was a device changing the frequency of the AC-supply just to improve the HiFi sound quality!
26 Nekro synthesis
(Just for recreation, the headword is my invention) Members of the German „Verein für Transkommunikationsforschung“ ( = Club for Research of Transcommunication) receive broadcast data. Using the software „Cool Edit“ they produce speech pieces of 50 milliseconds, mix them up in an aleatoric manner getting a medley of syllables. This mixture of words and sentences is meant to come from the deceased and gives replies to questions posed during the recording sessions. No comment.
27 PCM (= Pulse Code Modulation) >35
Changes analogue data to digital.ones. No synthesis.
28 Phase Distortion Synthese; (see also iPD )
Similar to FM Synthesis ; though the carrier has been modified by a wave-shaper-modul.
29 Physical Modeling (also: modelling)
P. M. tries to describe exactly classical instruments. By mathematical means has been created a world of arches, strings, resonating bodies and tubes, even the procedure of generating sound is a matter of equations. P. M. uses e. g. Karplus-Strong or waveguide.
30 Pulse Width Modulation
PWM works in an analogue manner. Waveforms of an oscillators’ pulse can be altered in width by another oscillator (LFO) or an envelope.
31 RCM (Realtime Convolution)
Modulation Synthesis, refined FM Synthesis
32 Resonance Modulation
Works with alterations of the resonance of the filter.
Conditions a former analysis. Additive Synthesis, done digitally, based on data from a Fourier analysis. Why do this? Because now you can change the envelope parameters (pitch, velocity) of every partial without change the other sound characteristics! This combination of analysis and following resynthesis is e.g. important for “time stretching” , the phase vocoder and LPC (Linear Predictive Coding)
Fits also „cross modulation“ and FMX („frequency cross modulation“).
Coined after the blueprint of the original analogue wiring; like AM, though missing the carrier. No synthesis, just a procedure. Two basic frequencies are mixed with their difference and sum frequencies. Thus one will get non harmonic overtones ( = metallic sound).
35 ROM sample synthesis (sample playback)
PCM (Pulse Code Modulation) supplies data. Here also belong to: AWM (Advanced Wave Memory), AWM2 (Advanced Wave Memory Version 2) and AI (Advanced Integrated); stored Samples are to be played back.
36 Sample Playback >35
37 Scanned Synthesis
Scanning of an object leads to a shape. That shape means to be the transforming element of a sample. Everything could be the shape, e.g. a drawing, a photograph, even a statistical graph!
38 Subtractive Synthesis
Waveforms full of overtones (sawtooth e. g.) and generated of one or more oscillators are in their origin very cold and hard. Processed through filters and thus loosing some partials they become more “dark”, smooth. The filters effects can be governed. So it is possible to modulate and generate envelopes by certain controlers.
39 Time Slice Synthese
Derived from the Fourier synthesis. Sound has been generated of spectra settled on the time axis one after the other. The oscillator interpolates between them; each spectrum has its “time slice”.
40 Transitional Synthesis
The transition is done by interpolating of two similar events. > 47 deals with the transition between two filter shapes.
41 Vector Synthesis
Extension of the Wavetable Synthesis. Four Oscillators are controlled by a joystick, an envelope or another controler. In a general manner VF. S. is a procedure allowing to fade in real time from one sound source to another.
42 Virtual Synthesis >29
43 Wavelet Analysis;
Variation of Fourier Analysis in dissection of sound events. Wavelets are equations which can split sound events in arbitrary components.
The resolution could be unrestricted, not possible in Fourier Analysis. So even graphs with peaks and non steady areas can be handled.
44 Wave Sequencing
Little parts of different samples are composed linear to a new one.
Mutation of the original sound by a non-linear algorithm. Leads to distortion or fuzz. Can be used creatively because of emerging resonance.
46 Wavetable Synthesis
1) Ready made digital waveforms are processed either by additive or subtractive methods to get new sounds. The ROM storage of the digital waveform data in tables ( = wavetables) gives the name.
2) Some manufacture use the word to indicate the soundcard’s ability to store their playback samples in RAM. Has nothing to do with synthesis.
47 Z - Plane Synthesis
Transition between two filter shapes.
A new calibration of terms, more severe than in the page “Overview”
1) record of a certain sound event
2) digital description of 1), done bei PCM (Pulse Code Modulation). PCM is not a modulation! The scan is been handled by an A/D converter.
wave, wave form, sometimes mistaken for oscillation
Oscillator reading samples (> table-lookup).
Produced by PCM; table of digital data describing asound in time.
Curtis Roads („Computer Music Tutorial“) has nominated it more precisely “fixed-wave- form table-lookup Synthesis“. The elemental work of an oscillator; it reads exactly one period of a basic wave form (e.g. sawtooth) and sends it as long as desired to th A/D converter
(has nothing to do with Wavetable Synthesis!).
Early synthesizers processed physical data such as voltage to get changes in sounds. The sounds’ features changed analogue to the physical signals. Because of the character of this
Analogue Synthesis to affect the sounds by taking away (e.g. through filters or envelope generators) one speaks of Subtraktive Synthesis.
What are the materials of Sound Synthesis?
1) analogue or digital samples
2) elemental waveformes
3) none but data and mathematical procedures.
1) Working with samples:
ROM-Sample-Synthesis (Sample Playback), AWM, AWM 2, AI, Direct Digital Synthesis, Granular Synthesis, PWM, Scanned Synthesis und Transitional Synthesis (Wave Sequencing, Wave Table Synthesis, Vector Synthesis)
2) Working with elemental Waveforms:
Additive Synthese (to puristes the only true kind of synthesis), also known as Fourier Synthesis , Formant Shaping, Time Sclice Synthesis and different forms of Modulation Synthesis.
Modulation Synthesis has several techniques derived from Additive or Subtraktive Synthesis followed by Resynthesis.Included are:
AM, FM (AFM), Ringmodulation (FMX, cross modulation), Phase Distortion (iPD), RCM, Resonance Modulation (Filter), Wave Shaping und Formant Synthesis. Resynthesis covers LPC and Phase Vocoder.
3) Digital Signal Processing, DSP deals with digital material (exactly: with “0” or “1”). Binary data kann be handled easily and become audible by D/A converters. Convolution (cross modulation) stand almost for everything: modulation, filter, reverberation und Cross Synthesis. Physical Modelling (sometimes Modeling) or Virtual Synthesis, Karplus-Strong and waveguide have their places here, too.
Mixed forms of synthesis are Composite Synthesis and Linear Arithmic Synthesis