Tensions in Practice: Time localization in tension with frequency localization¶
Two ideal Gaussian pulses
A pulse concentrated near one instant needs a wider range of frequencies. For the particular Gaussian pulses shown here, doubling the spread in time halves the spread in angular frequency. Their width product remains 0.5. One choice separates nearby events in time more readily; the other concentrates energy into a narrower frequency range. The difference persists even with a perfect instrument.
Localize the event in time
Choose a smaller time spread.
Concentrate the frequency content
Choose a smaller frequency spread.
Why these aims pull against each other
The two displayed widths belong to the same pulse, not independent measurement settings. Within this Gaussian family, sharpening one spreads the other.
Choose an arrangement to see what changes and what remains difficult.
Width means standard deviation of energy spread, not a hard boundary. Time and angular frequency use reciprocal units; the product stays 0.5 under this declared Gaussian and Fourier convention.
What this choice protects
What it costs
When it fits
Compare the arrangements
Concentrate in time
Choose a Gaussian whose time width is 1 in the declared time unit. Its angular-frequency width is 0.5 in the reciprocal unit.
| Time width | Freq. width | Product | |
|---|---|---|---|
| Pulse | 1 | 0.5 | 0.5 |
- What it protects
- The pulse is more concentrated around its central instant.
- What it costs
- Its frequency spread is twice that of the longer pulse.
- When it fits
- Fits a task where temporal separation matters more than a narrow frequency band.
Illustration note: Widths are standard deviations of normalized energy distributions. A Gaussian has infinite tails: width is not a hard start-to-stop duration.
Concentrate in frequency
Choose the same Gaussian shape stretched to time width 2. Its angular-frequency width becomes 0.25.
| Time width | Freq. width | Product | |
|---|---|---|---|
| Pulse | 2 | 0.25 | 0.5 |
- What it protects
- The pulse is more concentrated around its central angular frequency.
- What it costs
- Its time spread is twice as large, making its temporal location less concentrated.
- When it fits
- Fits a task where spectral concentration matters more than temporal separation.
Illustration note: For amplitude exp(−t²/(4s²)) and Fourier kernel exp(−iωt), the energy widths are s and 1/(2s). The two exact rows select s = 1 and s = 2.
What this illustration does—and does not—establish
The source supplies the structural tension. This bounded example makes a particular relation inspectable; the aims, conditions and residual costs are part of the comparison.
- These are exact ideal Gaussian calculations; no instrument noise or measured performance is being compared.
- Freq. width means angular-frequency standard deviation, not ordinary cycles-per-time bandwidth. The Fourier convention fixes the numerical product.
- The example establishes this signal relation only. It does not transfer a quantitative uncertainty product to organizations, learning models or other analogies in the surrounding source.
Source entries
Conjugate-Observable Complementarity
The canonical tension motivates this comparison. The setting and arrangements are declared editorial illustrations, not observed findings.
Which Observable to Sharpen (Application Dependence)
A conjugate pair offers a frontier, and selecting a point means deciding which observable to resolve at the cost of its conjugate — a choice that depends entirely on the downstream use, not on the structure. The tension is that the structure fixes the floor but is silent on *where* to sit.
The source operation
The defining content is not that we lack good enough instruments; it is that the two quantities are *defined in a relationship* that forbids their simultaneous arbitrary determination. There is a well-posed value of either one alone to any precision you like, but the pair cannot both be made arbitrarily sharp at once, because the very mathematics that makes one definite makes the other spread.