ASA Demos — User Guide

Outputs

Each run creates one mono Sound at 44.1 kHz, scales its peak to 0.99, optionally draws the AudioTools visualization, and optionally plays the result. No input Sound or external audio file is required.

Slow versus fast repetition of the same six-tone cycle

Cycle: H1–L1–H2–L2–H3–L3, with high tones at 2500, 2000, and 1600 Hz and low tones at 350, 430, and 550 Hz.

Level relation: low tones are generated 6 dB stronger than high tones before final peak scaling.

Structure: four cycles with 400 ms tones, 1 s silence, then sixteen cycles with 100 ms tones. Total output duration: 20.200 s.

Listening idea: compare integrated sequential hearing at the slower rate with the stronger possibility of high/low stream segregation at the faster rate.

Output: Demo_1_StreamSegregation

Three-tone standard inside a six-tone family

Standard: three high tones separated by equal 100 ms silences. The six-tone comparison alternates high and low members of the same frequency family.

Timing: 100 ms per tone or silent slot. The standard is repeated 15 times, followed by 1 s silence, then the full cycle repeated 15 times.

Level relation: low tones are 1.5 dB stronger than high tones before final peak scaling.

Purpose: illustrates how a pattern can remain associated with members of one putative stream when interleaved with another frequency range.

Output: Demo_2_Within; total duration 19.000 s.

Pattern membership crosses the high/low grouping

The standard and full six-tone cycle use the same 100 ms event grid and the same high/low level relation as Demo 2a, but the target pattern assigns one of its members to the low-frequency range. The comparison is designed to contrast within-stream and across-stream pattern organization.

Output: Demo_2_Across; total duration 19.000 s.

Accelerating HLH-gap cycle

Frequencies: H = 1400 Hz, L = 500 Hz. The separation is about 17.8 semitones.

Cycle: H–L–H–silence, with all four units having the same duration inside each cycle.

Tempo law: 20 cycles. Unit duration follows an exponential progression from 287 ms on cycle 1 to 88 ms on cycle 20, so the requested endpoint is actually reached.

Listening idea: at faster rates, frequency-based grouping can compete with the four-unit galloping pattern.

Output: Demo_3_LargeSeparation; duration approximately 13.548 s.

Same acceleration with close frequencies

Frequencies: H = 1400 Hz, L = 1320 Hz, about one semitone apart. The cycle and 287→88 ms exponential timing law are identical to Demo 3a.

Listening idea: compare the persistence of the HLH-gap rhythm when the two pitch ranges are much closer.

Output: Demo_3_SmallSeparation; duration approximately 13.548 s.

Build-up across longer uninterrupted HLH blocks

Frequencies: high = 2000 Hz; low = 700 Hz. Low tones are generated 6 dB stronger.

Cycle: H (113 ms), 12 ms gap, L (113 ms), 12 ms gap, H (113 ms), then 125 ms silence.

Blocks: 2, 4, 8, 16, and 32 uninterrupted cycles, separated by 4 s silence.

Listening idea: provides progressively longer exposure to the same repeating pattern so that possible build-up of streaming can be compared across blocks.

Output: Demo_4_Cumulative; duration 46.256 s.

Five transposed melody passes against unchanged distractors

Melody: a fixed 16-note semitone pattern relative to C4 = 256 Hz.

Presentations: the melody is shifted by +0, +2, +4, +6, and +8 semitones over five passes.

Distractors: a fixed random-like 16-value offset pattern is generated once relative to the untransposed melody and reused unchanged in every pass. There is no runtime randomness.

Timing: each melody or distractor event contains 120 ms tone + 23 ms silence, giving an event interval of 143 ms (about 7 events/s). Melody and distractor events alternate. Passes are separated by 500 ms silence.

Listening idea: as the melody moves away from the stationary distractor region, pitch proximity and continuation cues change.

Output: Demo_5_Melody; duration 24.880 s.

Synthetic fixed-upper / changing-lower alternation

The 16-note synthetic measure alternates a repeated G5 with changing lower notes: G5–E5–G5–C5–G5–D5–G5–B4–G5–C5–G5–A4–G5–B4–G5–G4.

Structure: two slow measures at 250 ms/note, 1 s silence, then four fast measures at 100 ms/note.

Listening idea: at the faster rate, the repeated upper note can support a separate-stream interpretation from the moving lower line.

Output: Demo_6_TelemannModel; duration 15.400 s.

Synthetic two-player equipentatonic interlock

The model uses a five-step equal division of the octave with 240-cent spacing from a 350 Hz base. The current pattern uses f1, f3, f4, and f5 from that grid.

Player A: f1 – silence – f3 – silence. Player B: silence – f4 – silence – f5. Each part is heard alone for eight cycles, followed by the other part, then a 16-cycle combined interlock f1–f4–f3–f5.

Timbre: decaying sinusoidal xylophone model sin(2πft) × exp(-40t) with a 2 ms attack ramp. Each slot lasts 120 ms.

Event rates: each player contributes about 4.17 events/s; the combined sequence is about 8.33 events/s.

Output: Demo_7_AmadindaModel; duration 17.360 s.

Same interlock with Part B shifted upward

The timing, player ownership, and synthesis are identical to Demo 7, but f4 and f5 used by Part B are multiplied by 2. This introduces an octave displacement of Part B while preserving its pattern.

Listening idea: compare player-based grouping when the two parts occupy overlapping versus more separated pitch ranges.

Output: Demo_8_PitchSepModel; duration 17.360 s.