Common Midpoint: The Foundation of Modern Seismic Imaging

What Is a Common Midpoint?

Common midpoint (CMP), also called common depth point (CDP) in older literature, is the surface location exactly halfway between a seismic source and a receiver such that multiple source-receiver pairs sharing the same midpoint will, under the assumption of horizontal layering, all record reflections from the same subsurface point. Collecting all traces that share a common midpoint into a CMP gather enables normal moveout correction, velocity analysis, and trace stacking to dramatically improve the signal-to-noise ratio and form the basis of a migrated seismic section used to image subsurface structure and stratigraphy.

CMP Geometry and the CMP Gather

In a conventional 2D seismic survey, a single shot generates signals recorded by a linear array of geophones at various distances (offsets) from the source. Every shot-receiver pair defines a unique midpoint. When the acquisition geometry is designed so that multiple shots and receivers sample the same midpoint, these traces can be extracted from the raw field data and assembled into a CMP gather: a collection of traces ordered by source-receiver offset, all sharing the same surface midpoint. For a 2D survey with receiver group interval of 25 m and shot interval of 25 m, each midpoint is sampled at 25/2 = 12.5 m spacing. The fold at each CMP equals half the number of active channels in the receiver spread; a 240-channel spread with 50% overlap generates 60-fold data at interior CMP locations.

The key diagnostic use of the CMP gather is velocity analysis. On an unmoved CMP gather, reflections appear as hyperbolas curving upward with increasing offset because energy traveling longer source-receiver paths takes more time to arrive. The shape of each hyperbola encodes the root-mean-square (RMS) velocity of the rocks above the reflector. By scanning a range of trial velocities and computing the semblance (a normalized measure of trace-to-trace coherence after NMO correction at each trial velocity), the processor identifies the velocity that best flattens each reflection hyperbola. The resulting velocity spectrum is a semblance panel plotted in time-velocity space, with peaks at the correct stacking velocity for each reflector. Velocity analysis is performed at CMP locations spaced every 500-1,000 m along a 2D line, and the velocity field is interpolated between analysis locations.

After NMO correction flattens the reflection hyperbolas using the picked stacking velocities, the corrected traces in the CMP gather are summed (stacked) to produce a single output trace at that midpoint location. This process attenuates random noise by a factor equal to the square root of the fold while preserving signal that is coherent across offsets. The stacked section, a 2D display of stacked traces plotted against CMP number on the horizontal axis and two-way travel time on the vertical axis, is the primary product delivered to the geologist or interpreter after basic processing. It represents a zero-offset approximation of the subsurface, assuming all reflections come from directly below the midpoint, which is only accurate when layers are horizontal.

Common Midpoint Synonyms and Related Terminology

Common midpoint is also referred to as:

• Common depth point (CDP) — the original term coined by W. Harry Mayne in his 1956 patent; technically less accurate because the reflection point is at depth, not at the surface midpoint, but still widely used interchangeably with CMP
• Common reflection point (CRP) — the preferred term when discussing the subsurface reflection location specifically; acknowledges that surface midpoint and reflection point coincide only for flat, horizontal layers
• Common image point (CIP) — used in angle-domain processing and prestack depth migration, where gathers are organized by reflection angle rather than surface offset
• CMP bin — in 3D seismic, the rectangular surface cell within which all midpoints are grouped to form the 3D equivalent of a 2D CMP gather

Related terms: normal moveout, seismic stacking, velocity analysis, seismic migration, seismic fold

Frequently Asked Questions About Common Midpoints

Why does CMP not equal common reflection point when layers are dipping?

When a reflector is dipping, the ray from the source reflects off the dipping surface and travels to the receiver along a path governed by Snell's law and the geometry of the dip. The actual reflection point on the dipping interface shifts updip from the surface midpoint by an amount that depends on dip angle, reflector depth, and offset. For a 15-degree dip at 2 km depth, the reflection point displacement from the midpoint can exceed 200 m, meaning that traces nominally in the same CMP gather are actually sampling different points on the reflector. This is why stacked sections produced without migration show dipping reflectors as "smeared" or displaced from their true structural position. Seismic migration collapses these diffraction patterns and repositions dipping reflectors to their correct subsurface locations, transforming the CMP stack into a geometrically accurate image.

How does 3D CMP bin geometry differ from 2D CMP?

In 2D acquisition, midpoints fall along a single line and are sorted into 1D bins (each bin being a point on the line with a defined bin width, typically half the receiver group interval). In 3D acquisition, sources and receivers are deployed in a 2D areal pattern (e.g., parallel shot lines perpendicular to receiver lines), so midpoints scatter across a 2D surface. The 3D equivalent of the CMP is a rectangular bin covering typically 12.5 x 12.5 m to 25 x 25 m. All traces whose midpoints fall within a bin are gathered together for velocity analysis and stacking. Because traces in a 3D bin come from a range of azimuths as well as offsets, 3D CMP gathers contain azimuthal information that 2D gathers lack, enabling azimuthal anisotropy analysis for fracture characterization.

What is NMO stretch and why does it matter?

NMO correction is a time-shift applied to each trace in a CMP gather to flatten reflection hyperbolas. At large source-receiver offsets, the required time shift is large and varies rapidly with time, causing the waveform to be stretched or compressed in ways that change its frequency content. This distortion is called NMO stretch: a wavelet that originally had a dominant frequency of 40 Hz may appear as a 25 Hz wavelet after NMO correction at long offsets. Mixing stretched and unstretched waveforms during stacking degrades the stacked wavelet quality and can create frequency-dependent amplitude artifacts. To prevent this, processors apply a stretch mute that zeros out (mutes) samples where the NMO stretch exceeds a threshold, typically 20-30% frequency change. The stretch mute preferentially removes near-surface, long-offset traces, effectively reducing fold at shallow times.

Why Common Midpoints Matter in Oil and Gas

The CMP method transformed seismic exploration from a low-quality single-fold technique into a high-fidelity imaging tool capable of resolving reservoir-scale features. By gathering and stacking multiple independent measurements of the same subsurface reflection point, the CMP method suppresses noise, enables quantitative velocity analysis, and produces the image quality required to identify structural traps, stratigraphic pinchouts, and amplitude anomalies associated with hydrocarbons. Every modern reflection seismic survey, whether 2D or 3D, land or marine, uses CMP geometry as its fundamental organizing principle. The stacking velocities derived from CMP velocity analysis also provide the primary input to depth conversion and pore pressure prediction, making the CMP gather not just an imaging tool but a source of quantitative subsurface information that directly informs drilling decisions.