@gsknnft/fft-ts

Production-facing FFT and spectral utility toolkit for TypeScript and Node.js.

fft-ts is the public FFT lane in SigilNet. It exposes:

• high-level FFT helpers
• lower-level transform engines
• image-frequency helpers
• deinterleave utilities for interleaved channel data
• event-oriented frequency helpers for quick signal inspection

Install

pnpm add @gsknnft/fft-ts

Public Surface

Core transforms

import {
  computeFFT,
  computeFFTSpectrum,
  fft,
  FFTProcessor,
  FourierTransform,
} from "@gsknnft/fft-ts";

const samples = new Float64Array([0, 1, 0, -1, 0, 1, 0, -1]);

const bins = computeFFT(samples);
const spectrum = computeFFTSpectrum(samples);

const engine = new fft(samples);
const out = engine.createComplexArray();
const input = engine.toComplexArray(samples);
engine.transform(out, input);

Deinterleave utilities

import { deinterleave, deinterleaveChannel } from "@gsknnft/fft-ts";

const interleaved = new Float32Array([
  0.1, 0.9,
  0.2, 0.8,
  0.3, 0.7,
]);

const [left, right] = deinterleave(interleaved, { channels: 2 });
const justLeft = deinterleaveChannel(interleaved, {
  channels: 2,
  channel: 0,
});

Event / frequency helpers

import {
  fourriouoorAny,
  fourriouoorFreq,
  fourriouoorFreqInv,
  fourriouoorTick,
  padToPowerOfTwo,
} from "@gsknnft/fft-ts";

const events = [
  { amplitude: 0.2 },
  { amplitude: 0.5 },
  { amplitude: -0.1 },
];

const padded = padToPowerOfTwo(events.map((event) => event.amplitude));
const quickSpectrum = fourriouoorAny(events);

Image-frequency helper

import { FFTImageDataRGBA } from "@gsknnft/fft-ts";

const complexImage = FFTImageDataRGBA(imageData.data, imageData.width, imageData.height);

This helper is useful for experimental image-analysis pipelines, QA metrics, and frequency-domain image inspection. It is not an image reconstruction engine by itself.

Exported Types

The package also exports shared types such as:

• ComplexLike
• FFTDirection
• FFTResult
• SpectrumBin
• AmplitudeEvent
• SignalEvent
• AnyEvent
• DeinterleaveResult

It also exports utility aliases for the legacy complex-array helper path:

• FFTUtility
• InvFFTUtility
• frequencyMapUtility

Notes

• The top-level helper APIs now avoid mutating caller-owned ComplexArray inputs unexpectedly.
• fftimage is exported as a real runnable surface.
• deinterleave is production-safe and no longer tied to benchmark-only code.

Suggested Use In Media Pipelines

fft-ts is a good fit for:

• frequency fingerprints
• edge/detail scoring
• texture/noise analysis
• synthetic capture QA
• interleaved audio/signal channel splitting

It is not intended to replace domain-specific reconstruction or mesh-generation pipelines.

Release Checks

pnpm --filter @gsknnft/fft-ts typecheck
pnpm --filter @gsknnft/fft-ts build
npm pack --dry-run --prefix packages/fft-ts

License

MIT