technology

How the signal is built

Fluctara generates its audio rather than replaying a recording. Every session is assembled from named parameters, and those parameters are the same ones described on this page. Nothing below is a claim about what the sound does to a listener; it is a description of what the software produces.

Three ways to construct a beat

A rhythmic pulse can be built in more than one way, and the methods are not interchangeable. They differ in what reaches each ear, which is why one of them needs headphones and the others do not.

Binaural: Two carriers, one per ear, offset by a small difference. The pulse is a product of hearing both at once, so the construction depends on separate delivery to each ear. Parameters: carrier frequency, beat frequency, amplitude, phase offset.
Monaural: The two tones are summed before they reach the ear, so the pulse exists in the waveform itself. It survives a single speaker. Parameters: carrier frequency, beat frequency, amplitude, phase offset.
Isochronic: One carrier switched on and off at a pulse rate. Parameters: carrier frequency, pulse rate, amplitude, duty cycle, and an edge length in samples.

Why the edges are a parameter

Switching a tone on and off abruptly produces a click, and a click is a wide band of frequencies nobody asked for. The isochronic generator therefore takes an explicit edge length: the number of samples over which each pulse rises and falls. The duty cycle sets how much of each period the tone is present.

This is the level at which the system is specified throughout. Where a parameter changes the output, the parameter is named and exposed rather than tuned behind a preset.

Noise is generated, not sampled

Noise layers are synthesised by colour and carry a seed. The same seed and the same colour produce the same samples, which is what makes a rendered file reproducible rather than merely repeatable. Environmental layers — rain, forest, stream, ocean, wind, fire, night — are generated on the same basis.

Placement and mixing

Sources can be positioned by azimuth, elevation, and distance. A mix combines several carriers under a master amplitude with an explicit ramp, so a session does not begin at full level.

The output is measured against a published standard

Rendered audio is analysed rather than assumed. The quality report is bound to a schema and carries a receipt digest, so a report that has been edited after the fact stops verifying.

Five source rates are supported, each with its own true-peak oversampling factor, because a true-peak reading is only meaningful at a declared interpolated rate.

44.1 kHz: Compact-disc family, four-times oversampling to 176.4 kHz.
48 kHz: Broadcast family, four-times oversampling to 192 kHz. This is the rate for which the EBU Loudness Test Set v5.0 minimum-requirement vectors apply.
88.2 kHz: Two-times oversampling to 176.4 kHz.
96 kHz: Two-times oversampling to 192 kHz.
192 kHz: Already at the interpolated rate; no further oversampling.