Automatic Gain Control: Seismic AGC, Amplitude Destruction, and True-Amplitude Alternatives

Key Takeaways

• AGC gain computation and application mechanics: The AGC operator works in a two-pass fashion. In the first pass, the RMS amplitude within a window of length T (centered at each sample) is computed across the entire trace length. The instantaneous gain G(t) is then calculated at each sample as the ratio of the desired output target level to the measured RMS(t). In the second pass, G(t) is multiplied sample-by-sample into the input trace. The gain function G(t) is the inverse of the trace's energy envelope: wherever the trace is high-energy (shallow reflections, hard interfaces), G(t) is small, scaling those amplitudes down; wherever the trace is low-energy (deep reflections, attenuated zones), G(t) is large, boosting those amplitudes up. The result is a trace whose local RMS amplitude is nearly constant throughout its entire length, regardless of the true physical amplitude variation caused by geometric spreading, earth attenuation, or lithology contrasts. The time gate T and the target level are the two user-controlled parameters. Most processing software applies a smoothing filter to the gain function G(t) to prevent abrupt gain transitions at the edges of strong reflectors, which would appear as ringing artifacts in the output. The gain trace G(t) is typically saved separately from the processed data to allow the AGC to be reversed for comparison or QC purposes, though once applied, the original trace amplitude information is effectively lost for practical purposes.
• Window length and its effect on amplitude equalization aggressiveness: The time window length T is the single most important parameter controlling AGC behavior. A short window (50-100 ms) computes local RMS over only a few cycles of the dominant seismic frequency, meaning the gain responds very rapidly to changes in local amplitude. This aggressive equalization produces a section where even closely spaced strong and weak reflectors are brought to similar display amplitude, maximizing the visual continuity of every reflection event but destroying nearly all amplitude information. A short-window AGC is useful for structural mapping in poorly imaged areas where a faint event needs to be brought up to display level. An intermediate window (200-300 ms) produces a moderate equalization that preserves broad amplitude trends (deep-to-shallow differences) while equalizing within intervals of comparable frequency content. A long window (500-1,000 ms) applies a gentle gain recovery that compensates primarily for geometric spreading and bulk attenuation without equalizing individual reflection events relative to one another. A very long window (equal to the full trace length) approaches a scalar gain applied to the whole trace, which preserves all relative amplitude relationships within the trace even though it normalizes different traces to the same trace energy. Choosing the correct AGC window for a given application is a QC step that requires the processor to examine the gain traces alongside the output amplitudes to ensure the chosen equalization is appropriate for the interpretation goal.
• Why AGC must never be applied before quantitative amplitude analysis: The destructive effect of AGC on amplitude information is the most critical practical point for geoscientists interpreting seismic data in the context of reservoir characterization. AVO analysis requires that the relative amplitude of a reflection event changes predictably from near to far offset according to Shuey's approximation of the Zoeppritz equations: a gas sand at top produces a negative intercept A and a negative gradient B, giving a far-offset amplitude that is more negative than the near-offset amplitude. If AGC has been applied, the near-offset and far-offset amplitude stacks are independently scaled within their own windows, and the gradient information is contaminated or destroyed. Similarly, DHI mapping on stack data requires that amplitude anomalies (bright spots, dim spots, flat spots) represent genuine contrasts in acoustic impedance, not artifacts of varying local gain. If a bright-spot candidate sits adjacent to a noise burst or a high-amplitude shallow reflector that drove the AGC gain down in its window, the bright-spot amplitude will be suppressed by the low gain in that window, making it appear fainter than its true reflectivity warrants and potentially causing it to be discarded as non-anomalous. Acoustic impedance inversion and seismic-to-well ties are equally sensitive: the inversion attempts to recover absolute reflectivity from the seismic amplitudes, but if those amplitudes have been distorted by a varying AGC gain, the recovered impedance will be systematically biased in ways that vary unpredictably from one part of the section to another. The rule is absolute: amplitude-sensitive workflows must use true-amplitude processed data, not AGC-processed data.
• True-amplitude processing alternatives to AGC: The geophysical community developed true-amplitude processing to meet the need for amplitude-preserved seismic data suitable for quantitative analysis. True-amplitude processing applies physically based gain corrections that restore the seismic trace amplitudes to a state representative of the earth's actual reflectivity, rather than equalizing them to a visual target level. The primary components of true-amplitude processing are spherical divergence correction (applying a gain proportional to t², the square of two-way travel time, to compensate for the geometric spreading of wavefront energy as it expands away from the source), Q compensation (applying a frequency- and time-dependent inverse-Q filter to recover the high-frequency energy absorbed by the earth's anelastic attenuation, using Q estimates from VSP or spectral ratio methods), and surface-consistent amplitude corrections (normalizing receiver and source coupling variations by solving for source and receiver scalar terms in a surface-consistent factorization of the trace amplitude). After these physically based corrections, the residual amplitude variation in the data represents genuine geological contrasts in acoustic impedance, and the data can be used for AVO, DHI mapping, inversion, or any other quantitative amplitude workflow. The critical distinction from AGC is that true-amplitude processing applies a single, physically motivated gain function per trace or per surface-consistent factor, preserving the relative amplitude relationships between reflections while removing the effects of wave propagation physics. True-amplitude data still require care in interpretation because noise, multiple contamination, and processing artifacts can mimic amplitude anomalies, but the amplitude information is not deliberately destroyed.
• AGC in electronics, seismic recording instruments, and MWD telemetry: The term automatic gain control originated in radio communications and electronics, where it refers to a feedback circuit that adjusts the gain of an amplifier to maintain a constant output signal level despite variations in input signal strength. This electronic AGC is a standard feature of AM radio receivers, radar systems, and audio amplifiers. In seismic data acquisition, the recording instruments (geophones connected to 24-bit analog-to-digital converters via field acquisition units) use electronic AGC-like dynamic range management to handle the large amplitude variation from the direct wave at close offsets to the faint deep reflections at far offsets. Modern seismic recording systems with 24-bit digitizers have sufficient dynamic range (approximately 144 dB) to record both the strongest and weakest signals in a single record without electronic gain adjustment, but older 16-bit systems with only 96 dB of dynamic range required electronic AGC in the amplifier chain to prevent clipping on the strong early arrivals while still recording the weak late arrivals. In MWD (measurement while drilling) mud-pulse telemetry, the surface decoder uses an electronic AGC circuit to manage the variable amplitude of the pressure pulses arriving at the surface transducers: pump noise, pipe resonance, and variable mud flow create a highly variable noise floor, and the AGC keeps the decoder sensitive to the incoming pulse signal regardless of the ambient noise amplitude. The seismic-processing and electronics definitions of AGC are conceptually identical: both adjust gain to maintain constant output amplitude in the face of highly variable input signal levels, and both are appropriate for signal reception and display but inappropriate for measurements that depend on the true relative amplitude of the incoming signal.