Indexing the archive…
Your Universe of Digital Possibilities
A pure tone has a frequency but no moment; a sharp click has a moment but no frequency. Every real signal lives between — and Fourier, which assumes a wave repeats forever, smears all of when into a single timeless spectrum. The wavelet is the honest reading: slide a little ripple across the signal at every scale and weave a picture of which frequency lives when. The price is a wall — the Gabor limit, the classical twin of Heisenberg: pin the moment and the pitch blurs, pin the pitch and the moment does. You choose the trade. You can’t escape it.
Correlate the signal against one little wave ψ, scaled by a (≈ 1/frequency) and slid to time b. |W|² is the scalogram — energy laid across time and frequency at once.
A complex sinusoid wrapped in a Gaussian. The width ω₀ is its wiggle-count: more of them sharpens frequency and blurs time, fewer does the reverse — the one knob on the trade.
No signal is sharp in both time and frequency: their spreads multiply to a fixed floor. The Gaussian alone meets it — the classical twin of Heisenberg’s Δx·Δp ≥ ℏ/2.
For the transform to invert, the wavelet must average to zero — a true little wave, not a lump. That’s why ω₀ can’t fall too low: too few wiggles and the mean stops vanishing.
The whole signal folded onto the frequency axis: perfect pitch resolution, and zero sense of when. The wavelet trades a sliver of this for all of the missing time.
This is the modern successor to The Spectrum (INST·01): Fourier folds a signal onto the frequency axis and loses time entirely; the wavelet keeps both, at a resolution the Gabor limit caps. It reads the same single wire as The Shadow (INST·31) and lives in the sampled-time world of The Strobe (INST·08). And it is the twin wall to The Pact (INST·38): the Gabor limit is a hard floor on a signalthat holds, where Bell’s bound is a hard wall on reality that breaks. For a builder whose thesis is that everything is a time-series (วิเคราะห์ → จำลอง → ทำนาย, AxionCore), the wavelet is the tool that answers the question Fourier can’t: not just which behaviours a stream contains, but when each one arrived.