The Origins & Evolution of Sampling and Synthesis
Sampling & Synthesis
Section One
Section Two
The Origins & Evolution of Sampling and Synthesis
A synthesizer, built especially for sound production or "synthesis" and modification, is essentially a device that merges sound generators and sound modifiers in one package with an integrated control system. The first and most elaborate of these devices was the RCA Electronic Music Synthesizer, first exposed in 1955. But the story goes back far before that.
Radio engineers experimenting with radio vacuum tubes discovered the principles of beat frequency or heterodyning oscillators by chance during the first decades of the twentieth century. The heterodyning effect is created by two high radio frequency sound waves of similar but varying frequency combining and creating a lower audible frequency, equal to the difference between the two radio frequencies. The musical potential of the effect was noted by several engineers and designers including Maurice Martenot, Nikolay Obukhov, Armand Givelet and Leon (or Lev) Sergeivitch Termen the Russian Cellist and electronic engineer.
One problem with utilizing the heterodyning effect for musical purposes was that as the body came near the vacuum tubes the capacitance of the body caused variations in frequency. Leon Termen realized that rather than being a problem, body capacitance could be used as a control mechanism for an instrument and finally freeing the performer from the keyboard and fixed intonation.
Termen's first machine (shown above left with it's inventor), built in the USSR in 1917 was christened the "Theremin" (after himself) and was the first instrument to make use of the heterodyning principle. The original Theremin used a foot pedal to control the volume and a switch mechanism to alter the pitch. This prototype evolved into a production model Theremin in 1920, this was a unique design, resembling a gramophone cabinet on 4 legs with a protruding metal antennae and a metal loop. The instrument was played by moving the hands around the metal loop for volume and around the antennae for pitch. The output was a monophonic continuous tone controlled by the performer. The resonance of the instrument was fixed and resembled a violin string sound. The sound was produced directly by the heterodyning combination of two radio-frequency oscillators: one operating at a fixed frequency of 170,000 Hz, the other with a variable frequency between 168,000 and 170,000 Hz. The frequency of the second oscillator being governed by the closeness of the musician's hand to the pitch antenna.
This Theremin model was first shown to the public at the Moscow Industrial Fair in 1920 and was witnessed by Lenin who requested lessons on the instrument. Lenin later commissioned 600 models of the Theremin to be built and toured around the Soviet Union.
Maurice Martenot a Cellist and radio Telegraphist, met the Russian designer of the Theremin, Leon Termen in 1923, this meeting lead him to design an instrument based on Termen's ideas, the first model, the "Ondes-Martenot" was patented on the 2nd of April 1928 under the name "Perfectionnements aux instruments de musique électriques" (improvements to electronic music instruments).
His aim was to produce a versatile electronic instrument that was immediately familiar to orchestral musicians. The first versions looked nothing like the later production models: consisting of two table-mounted units controlled by a performer who manipulated a string attached ...
This is a preview of the whole essay
Maurice Martenot a Cellist and radio Telegraphist, met the Russian designer of the Theremin, Leon Termen in 1923, this meeting lead him to design an instrument based on Termen's ideas, the first model, the "Ondes-Martenot" was patented on the 2nd of April 1928 under the name "Perfectionnements aux instruments de musique électriques" (improvements to electronic music instruments).
His aim was to produce a versatile electronic instrument that was immediately familiar to orchestral musicians. The first versions looked nothing like the later production models: consisting of two table-mounted units controlled by a performer who manipulated a string attached to a finger ring. This device was later incorporated as a fingerboard strip above the keyboard. Later versions used a standard keyboard.
The Ondes-Martenot became the first successful electronic instrument and the only one of its generation that is still used by orchestras today. Maurice Martenot himself became, 20 years after its invention, a professor at the Paris Conservatoire teaching lessons in the Ondes-Martenot.
Termen left the Soviet Union in 1927 for the United States where he was granted a patent for the Theremin in 1928. The Theremin was marketed and distributed in the USA by RCA during the 1930's and continues, in a transistorized form, to be manufactured by Robert Moog's 'Big Briar' company.
R.A. Moog, Inc.
While getting his doctorate in physics at Cornell University in 1963, Robert A. Moog earned money by selling do-it-yourself transistorized theremins. One of Moog's thereminists was Herbert Deutsch, a composer. Moog and Deutsch met in 1963 at a conference. This led to conversation about the need for new electronic instruments. In 1964, they spent weeks trying electronic circuitry. Moog had his very first prototype built in August 1964.
The principle behind Moog's prototype is the use of "Voltage-Controlled Electronic Music Modules," (i.e. the use of controlled generators and controlled filters). This allows the performer to create his own sounds by connecting modules with electric wires and by turning knobs.
There were two main models produced by Moog at this time.
Model:
Modular Moogs (3C, 2C, 1C, 3P, 2P, 1P, 10, 12, 15, 35, 55, C.E.M.S)
A list of modules for the Modular Moog
Production period:
3C/2C/1C: 1967-72
3P/2P/1P: 1970-72
0/12: 1971-73
5/25/55: 1972-81
Model:
Minimoog
Production period:
970-81
Quantity produced:
over 12,000
Other companies also began bringing out more synthesizer models. Roland brought out the Jupiter-8 and after that Sequential Circuits brought out the Six Trak (model 610). Both are shown below respectively.
A more sophisticated model, the RCA Mark II, was installed in 1959 at the Columbia University Studio in New York City, and still resides there today. This immense machine is capable of generating any imaginable sound or combination of sounds, with an infinite variety of pitches, durations, timbres, dynamics, and rhythmic patterns far beyond the capabilities of conventional instruments. The synthesizer represented an enormous step forward, and the developments of smaller, portable synthesizers that can be played directly have made live electronic performance possible. Electronic music studios are now common in universities and colleges across the world.
Today, just about any synth that you can find will incorporate sampling and synthesis, although eventually it, too, will be superseded. Already, Virtual Synthesis (a.k.a. Physical Modelling) is becoming more widespread as it tries to get closer and closer to those original analogue sounds, and also to more standard ones. It may not have replaced S&S yet, but once someone can accurately model each instrument of the orchestra, who knows what will happen?
The Principle Types of Synthesis
Sound is a series of airborne vibrations (compressions and decompressions of the air). The brain turns these vibrations into a recognisable sound, with the ear therefore acting as a transducer (something that turns one form of energy into another). The higher the frequency of the sound, the higher its musical pitch. However, there is more to a sound than just its pitch - consider the difference between a flute and a saxophone, even when they both play the same note. Sound synthesis is therefore about recreating these different qualities of sound, and also about creating new ones.
This is known as a Sine Wave, because it can be represented by the equation y = Sin (x). It is a boring sound, as its single frequency means that there is no change in tone. It cannot be filtered, as there is nothing to filter out. It sounds just like a whistle or a test tone. Most sounds are much more complex, due to the presence of harmonics.
Each harmonic in a harmonic series is an integer multiple of the fundamental. The harmonic series for A is 110Hz.
Harmonic
F (A)
2
3
4
5
6
7
8
9
0
Frequency (Hz)
10
220
330
440
550
660
770
880
990
100
The Character Of Sound
The character of a sound is affected by two factors. These are the harmonic content, and any changes in level and pitch whilst the sound is heard. Envelopes determine such changes.
An envelope is generally seen in the ADSR form, standing for Attack, Decay, Sustain and Release. Some examples should help to clarify this:
General Envelope:
Percussive Envelope:
Bowed Envelope:
Just as an envelope alters the level of a sound, it can alter the pitch. A synthesiser needs to know the all the harmonics and overtones that make up a sound, to set them at their relative values. Separate pitch/level envelopes can then be applied to each of them individually.
Analogue (Subtractive) Synthesis
The classic analogue synths relied upon oscillators, filters and envelope generators in order to create their sounds, all of which were voltage-controlled. The circuits inside them, however, were not entirely stable, leading to slight fluctuations in the voltage. This resulted in the full, fat sound that is now associated with analogue synths.
Subtractive synthesis is a lot like sculpture: you start with a large lump of stone (or a harmonically rich waveform) and carve away what you don't want to leave the desired result. The starting point can be any waveform, but the best one to use is noise, as it is the most harmonically rich sound of all. This is generated by an oscillator, and is then passed through a filter to remove any unwanted harmonics and overtones. The new sound is then subjected to an envelope generator.
The VCOs (Voltage Controlled Oscillators) found in analogue synths are capable of creating a variety of waveforms:
> Sine waves There are no harmonics present in a sine wave - only its level can be altered. It is therefore useless for subtractive synthesis.
> Square and pulse waves The square wave is rich in odd harmonics (3, 5, 7, etc.) It can be modified by filtering and pulse width modulation. It fluctuates between constant positive and negative amplitudes:
> Triangle waves The triangle wave is rich in odd harmonics, but to a lesser extent than the square wave.
> Sawtooth waves This wave is rich in even harmonics, and is therefore very good for filtering.
> Noise Noise contains all frequencies of sound, and is just a hiss.
Each waveform has its own unique shape and characteristics, and some are more useful than others.
Digital (Additive) Synthesis
As the name implies, additive synthesis creates sounds by adding harmonics and overtones, not taking them away as you do in subtractive synthesis. This is achieved by summing together sine waves whose frequencies are related to each other in exact correspondence with the harmonic series. The absolute minimum number of harmonics needed is 32, with many synths offering 64, or even 128. Most of the circuitry is the same as for subtractive synths, with the sine waves being produced by oscillators, and the sounds being adapted by envelope generators and LFOs (Low frequency oscillators). Of course, as this is not subtractive synthesis, the filters are not needed, although many additive synths include filters for those who like to fiddle with them whilst playing.
Other Forms of Digital Synthesis
> FM Synthesis FM synthesis was developed in the 1980s and used exclusively by Yamaha. It is a great deal more complicated than subtractive synthesis, because it works the other way around - instead of chopping bits of your sound off, you add them on.
> PD Synthesis PD (standing for Phase Distortion) synthesis was developed by Casio in the late 1980s. It is similar to FM in that it uses simple waveforms (although more complex than FM's sine wave) and mathematically alters them.
> VPM Synthesis VPM stands for Variable Phase Modulation, and it is really very similar to FM - the only major difference is that it has been developed by Korg instead of by Yamaha.
The Principle Types of Sampling
A waveform, when it is recorded, is done so in samples:
The samples are equidistant in time, and the distance either above or below the x-axis records the position of each sample. That is how we can transmit a waveform using binary.
Samplers are used to record samples of sound, and load them into their RAM (Random Access Memory). It is therefore necessary to save the samples to a disk before the sampler is switched off. Any sound can be sampled, from a single note to a full orchestra, and either short patterns or loops can be sampled, too.
The sampler can also be used to alter the sample. Its pitch, for example, can be altered by changing the speed at which it is played (faster playback = higher pitch).
When looping, the first 2 or 3 seconds of a long sound are the best. The section to be looped must be taken from the sustain section of the envelope, and not include any part of the attack, decay or release phases, as this would cause audible glitches every time the loop returned to the beginning.
It is important to include the original attack of the sound at the beginning of the playback, before the loop begins, as it is this attack that gives the character of the sound. A new release envelope should also be applied, so that the sound does not end abruptly. Your looped sample would then look something like this: