Questions: Granular Synthesis and Composition

Question 1

A composer wants to transform a string quartet recording into a shimmering, slowly evolving cloud that retains harmonic content but dissolves all rhythmic structure. Which granular synthesis approach best achieves this?

• A. Set grain size to 2ms to extract maximum frequency resolution from the source
• B. Use dense grain clouds (high grains/second) with randomized timing scatter, grain sizes around 40-80ms, and slight pitch spread around source pitches
• C. Apply a sharp rectangular grain envelope to preserve the spectral clarity of each grain
• D. Increase grain density above 1000 grains/second to completely eliminate any source characteristics

Answer: (B)

Moderate grain sizes (40-80ms) preserve enough of the original timbre for harmonic content to be recognizable. High density with randomized scatter breaks up rhythmic patterns and creates smooth clouds. Pitch spread adds shimmer while staying close to source harmony. Very short grains (2ms) would smear pitch; rectangular envelopes cause audible clicks at grain boundaries; extreme density can destroy source identity entirely — all working against the stated goal.

Question 2

When grain size is set very short (2-5ms), what is the primary acoustic consequence?

• A. The source material's pitch becomes cleaner and more focused because each grain is more spectrally pure
• B. Pitch information is smeared because grains shorter than the period of most pitched sounds cannot represent pitch accurately
• C. Rhythmic structure of the source is preserved more clearly because each grain is a discrete time-point
• D. The output sounds identical to the original but at reduced amplitude

Answer: (B)

Pitched sounds have periods typically ranging from about 2ms (high frequencies) to 50ms or more (bass). Grains shorter than the period of a pitch cannot complete enough cycles to establish that frequency — the result is a noisy, pitched smear rather than a clear tone. This is directly analogous to the time-frequency uncertainty principle: shorter time windows mean less frequency resolution. Longer grains (30ms+) preserve pitch at the cost of smearing rapid time-domain changes.

Question 3

Granular synthesis can produce pitched melodic content by transposing individual grains to target frequencies, not just abstract textural clouds.

• T. True
• F. False

Answer: True

True. Pitch transposition is a core grain parameter. By setting each grain's playback rate relative to its original sample rate, you can target specific pitches. Spreading grains across a range of transpositions creates chords; moving the transposition center over time creates glissandi and melodic lines. The misconception that granular synthesis is 'only textural' misses that grain-level pitch control enables precise melodic and harmonic organization.

Question 4

Granular synthesis is essentially the same process as digital time-stretching — both work by manipulating small segments of audio.

• T. True
• F. False

Answer: False

False. Time-stretching adjusts the duration of a sound while preserving its pitch (or vice versa), but it maintains the original time order of segments. Granular synthesis allows arbitrary reordering, scattering, transposing, and re-enveloping of grains — the original temporal sequence can be completely abandoned. A granular composer can loop a single grain indefinitely, scatter grains randomly, or play the source backwards at varying densities. These operations produce results qualitatively different from anything time-stretching can achieve.

Question 5

How does granular synthesis transform source material, and what does shifting the compositional unit from note/phrase down to grain allow composers to do that conventional synthesis or recording techniques cannot?

Model answer: Granular synthesis transforms source material by atomizing it into micro-fragments (grains) and reassembling them under compositional control of density, pitch, envelope, scatter, and playback order — completely independent of the original time sequence. This allows composers to treat timbre and texture as sculptable parameters rather than fixed properties. A single recorded sound becomes plastic: it can be frozen (dense looping), scattered across time, spread across pitches as a chord, or dissolved into a cloud with no recognizable connection to the source. Conventional synthesis builds sounds from oscillators; recording captures sounds whole. Granular synthesis operates at the level between — manipulating the microscopic time-structure of existing sound.

The conceptual shift is from 'sound as object' to 'sound as material.' Composers like Curtis Roads and Xenakis developed entire formal languages around grain-level control, making texture itself a structural element rather than surface decoration. This is only possible because granular synthesis decouples the time-ordering of sound from its timbre, allowing independent control of both.