Granular synthesis

Granular synthesis is a method used in some synthesizers and programs to create artificial sounds.

Similar to a movie pretending through individual images a liquid flow in the granular synthesis a continuous sound is simulated, which is in fact composed of many individual parts. For these items, the grains, is very short, digital sound fragments whose length is usually less than 50 milliseconds. This limit is exceeded, the listener can identify the fragment as an independent sound event.

It is a widespread application of granular synthesis for resynthesis sampled material. Here, the recorded sound is analyzed and broken down into those very Grains them subsequently reassemble.

Compared with the conventional sample playback method, this provides the advantage that the speed of the playback operation can be changed independently of the pitch. This is achieved by the multiplication of Grains in stretching of the material or the individual grains ( time-stretching ). Also, the formants and the pitch can be processed independently on granular synthesis ( pitch shifting ).

• 3.2.1 Pitch - synchronous granular synthesis
• 3.2.2 Synchronous and quasi-synchronous granular synthesis
• 3.2.3 Asynchronous granular synthesis

Theoretical basis

Theoretical basis of the granular synthesis are the Gabor transform and any related theory of quantum sound of Dennis Gabor. Here is an analogy between quantum phenomena in particle physics and the acoustic properties of short sound sections ( Gabor grains) is exploited.

Naturally occurring acoustic signals typically have a change in its spectrum over time. Thus, in order to be able to describe exactly one would be able to determine the spectrum of the signal at a certain period of time over an arbitrarily short period of time. However, the Fourier analysis using the continuous Fourier Transform is unsuitable for the examination of the short Frequenzsprektren Signalabschitte as occur at the required extrapolation to an infinite time period to large errors.

The resulting inaccuracy in the determination of the Partialfrequenzen signal is inversely proportional to the duration of the signal portion. Time and frequency (spectrum) is thus complementary characteristics of the signal segment which can not be accurately determined at the same time. A similar situation is known in the form of Heisenberg's uncertainty principle in quantum physics.

Gabor takes advantage of this by using mathematical tools in his theory for the description of acoustic signals, which were originally developed for the description of quantum physical phenomena. He considers this elementary signals containing exactly one quantum of information on spectrum and time. These have the form of a harmonic wave which is modulated by an envelope in the time domain, which indicates the probability of the corresponding spectrum to a portion of the time axis.

History

The composer Iannis Xenakis developed early 70s based on Gabor's theory for sound analysis, a compositional theory. Several of his compositions from that period describe the generation of musical events from a number of elementary sound particles, whose composition changes over a time course. Technically, this was initially very expensive realized by cutting of tapes into tiny sections and re- assembly according to the specifications of the composition.

The mid-70s made ​​the composer Curtis Roads further studies on the subject and experimented with computer-generated granular compositions, their calculation time still took a very long time. In order to simplify the control of the synthesis process, he developed for graphic description of the control parameters.

Barry Truax developed in the 80s for the first time a method for generating granular sounds in real time.

Operation

Grain parameters

Waveform

The basis for granular synthesis can basically any waveform can be used, so for example with additive synthesis and FM synthesis sounds produced any kind of audio sample, or even white noise.

Grain - duration

The length of individual grains can theoretically be selected from 1 to 100 milliseconds. Usually a length of about 10 to 50 ms is used because including the pitch is hardly noticeable and above the individual samples can be heard. The grain duration can be chosen to be constant or variable for all grains.

Amplitude and envelope

On each grain, an envelope is placed, which usually controls the sound volume on the basis of the following parameters: The approved peak amplitude, time to reach the ( attack), the holding time ( sustain ) and the time to return to a level of zero (release ). Such an envelope may be realized by a trapezoidal shape in the simplest case, but it may also be more complex curves such as the Gaussian curve can be used.

Grain density

The grain density is the number of grains to be played per second, is the factor that influences the texture of the generated sound, in effect the strongest. A low density of less than 30 grains per second produces a slightly rhythmic effect. The closer the grains are placed, the more they overlap, and the sooner the perception of rhythm is lost. At the same time increase the amplitude and the pitch and the result is a complex superimposition of sound texture with more different harmonics and formants.

Arrangement of the grains

Pitch - synchronous granular synthesis

The pitch- synchronous granular synthesis is a method for the analysis and resynthesis of a given sound with the goal of data reduction while preserving the formants. Put simply, the time-frequency plane is to split into many small cells, which at the end of each is represented by a grain.

Synchronous and quasi-synchronous granular synthesis

In the quasi-synchronous granular synthesis, a stream of grains is formed, which are attached to each other at certain time intervals. Usually this also a random deviation is chosen adjacent to a fixed time interval, whereby the individual grains to be played total at irregular intervals, however, a predetermined density ( grains per second) is maintained. A special case of this type of synthesis is the synchronous granular, in which the intervals between the reproduced grains are all of equal length.

Asynchronous granular synthesis

The principle of asynchronous granular synthesis detached from the concept of a linear stream of grains. Instead, the grains of a stochastic algorithm or chaotic in the so-called " clouds" of the time-frequency plane to be distributed. The composer may determine various properties of the " cloud", as the duration that allowed frequency band used waveforms for the Grains or a spatial distribution of sound.

Systems that can work with granular

• Kyma
• CrusherX Live
• CrusherX studio studio
• Malström in Reason
• Reactor (software) from Native Instruments
• Native Instruments Absynth
• Max / MSP from Cycling '74
• Ableton Live
• AudioMulch
• Super Collider
• Csound
• Pure Data
• TubeOhm Pure-D16/24Grain
• Camel Audio Alchemy
• FL Studio
• Kenaxis
• Padshop Steinberg