The Resolution Wars
Every survey is a fight for resolution and the earth fights back: wavelength sets the floor, tuning blurs thin beds into amplitude, aliasing punishes coarse sampling, Q quietly confiscates the high frequencies with every kilometer, and dispersion makes the lab, the log, and the survey measure three different rocks. Learn what can be seen at each scale, what cannot, and the honest arts, spectral decomposition, careful bandwidth, scale reconciliation, that claw detail back.
You can state a survey's vertical and lateral resolution and defend the numbers, work the tuning wedge and read thickness from amplitude below it, spot aliasing in time and space before it forges geology, budget the bandwidth Q will take and what compensation restores, explain why lab, log, and seismic velocities disagree by physics rather than error, choose bin size against dip, and use spectral decomposition to see stratigraphy the broadband stack hides.
The floor
A quarter wavelength is the vertical floor and the Fresnel zone the lateral one; every resolution promise starts by computing both and most disappointments start by skipping it.
Sharpness lives in octaves, not just high frequencies: the low end sets the envelope and the high end the detail, and the wavelet's bandwidth is the budget the whole image spends.
Below tuning, thickness stops being geometry and becomes amplitude; the wedge model is the decoder ring, and every thin-bed map in industry is secretly one of its slices.
Sample too coarsely and high frequencies return wearing low-frequency masks; aliasing in time or space manufactures signal that never existed, the resolution failure that lies rather than blurs.
Bin size is sampling in space, and steep dip is its enemy: the acquisition-side resolution decision that no processing can undo afterward.
The thieves
Attenuation taxes the high frequencies exponentially with depth, and the deep section arrives poorer for it; Q compensation can refund some of the loss, at the price of amplified noise.
Velocity depends on frequency, so the lab, the log, and the survey measure the same rock at three different speeds; dispersion is physics, not error, and forgetting it corrupts every cross-scale calibration.
A wave longer than the layering averages it: Backus theory is the exact mathematics of what the seismic wavelength refuses to resolve, the resolution floor derived instead of lamented.
Every processing stage either protected the spectrum or spent it; the spectral QC is the ledger that catches decon overreach and filter damage while the damage is still reversible.
Clawing it back
The broadband stack averages away what individual frequencies see: decomposed, each frequency tunes to its own thickness, and channels invisible in full band light up at thirty hertz.
Where targets are shallow the resolution war is winnable: ultra-high-frequency surveys image meters-scale detail, proof that the limits are physics, not habit, and that knowing the physics tells you when to demand more.
The final discipline: one rock, three measurement scales, three velocities, and a workflow that reconciles them by dispersion and averaging instead of splitting the difference; the capstone of resolution literacy.