AGC

AGC, or Automatic Gain Control, is a signal-processing technique in which the amplification applied to a recorded waveform is continuously and automatically adjusted so that the output amplitude remains within a defined target range regardless of variations in the true input amplitude. In seismic data processing, AGC normalises seismic trace amplitudes by computing the root-mean-square (RMS) energy within a sliding time window centred on each sample and scaling the sample by a gain factor that keeps windowed RMS constant throughout the trace. This equalization corrects for geometric spreading, intrinsic attenuation, and instrument drift, improving the visual detectability of weak late-time reflections that would otherwise be invisible on a display calibrated to large early-time amplitudes. However, AGC simultaneously destroys the true relative amplitude information that carries lithology and fluid content signals: if a sand body is bright because it contains gas, AGC suppresses that brightness to the same level as adjacent shales, obliterating the direct hydrocarbon indicator (DHI). For amplitude-versus-offset (AVO) analysis, any standard AGC application is therefore incompatible because it removes the near-to-far amplitude gradient that defines AVO class. In wireline logging and MWD/LWD tool electronics, AGC refers to an electronic gain-control circuit that adjusts detector amplification as formation properties change during logging, keeping digitised signals within the dynamic range of the analogue-to-digital converter. In production automation contexts, AGC occasionally stands for Automatic Gas Control, a surface well-control system that adjusts choke position or backpressure to maintain a target gas production rate; this usage is distinct from the signal-processing meaning and confined primarily to gas compression and sales metering facilities.

Key Takeaways

Seismic AGC equalises amplitude across a trace to improve visual display quality but is incompatible with any amplitude-sensitive interpretation: The RMS gain applied to each sample equals the ratio of the target RMS level to the measured windowed RMS. For a sliding window of gate length T centred at time t₀, RMS_window = sqrt(1/N × sum(x_i²)) where N is the number of samples in T. Dividing each sample by RMS_window and multiplying by the target value forces all parts of the trace to have equal energy, making weak reflections at depth as visible as strong shallow reflections. On a seismic section, this produces a display where every geologic horizon appears with similar amplitude, ideal for structural interpretation but meaningless for any quantitative amplitude analysis including DHI mapping, gross rock volume estimation from amplitude maps, or AVO crossplot analysis.
Gate length is the single most important parameter in AGC design, governing the trade-off between noise suppression and amplitude preservation: A short gate of 100 to 200 ms computes RMS over a window spanning only 5 to 10 reflections at typical seismic frequencies (25 to 50 Hz), producing rapid gain variation that aggressively equalises amplitude at the cost of destroying lateral amplitude continuity. Even a 50 ms lateral variation in a bright sand would be suppressed by a 100 ms gate, making the anomaly invisible. A long gate of 500 ms to 2 seconds spans 25 to 100 reflection events, preserving more of the gross amplitude trend while still correcting for deep-shallow energy contrast. Neither approach preserves amplitude correctly for DHI work; only true relative amplitude (TRA) processing without normalisation is acceptable for AVO and DHI applications.
True relative amplitude (TRA) processing is the required alternative to AGC for AVO, DHI, and 4D seismic workflows: TRA corrects for geometric spreading using the theoretical 1/r² law and for intrinsic attenuation using a Q-compensation operator, but does not normalise out genuine amplitude variations caused by lithology, fluid content, or impedance contrasts. On a TRA-processed section, a gas-saturated sandstone appears brighter than a brine-saturated sandstone, and the amplitude gradient from near-offset to far-offset traces (the AVO response) is preserved. CSEG Geoconvention processing workshops and the Canadian Society of Exploration Geophysicists specification documents mandate TRA processing as a prerequisite for all AVO and DHI work in WCSB plays including Montney, Cardium, and Duvernay seismic programmes. The processing parameter flag indicating whether a dataset was processed with TRA or AGC must be documented in the SEG-Y processing report attached to all submitted seismic packages.
Wireline logging tool AGC is an electronic circuit mechanism completely unrelated to seismic processing AGC despite sharing the same acronym: In natural gamma ray (NGR) detectors, resistivity tools, and density-porosity tools, the detector count rate or voltage varies by orders of magnitude across different formation types: a high-gamma radioactive shale may generate 10 times the detector pulses of a clean carbonate. Without electronic AGC, the analogue-to-digital converter would be saturated in high-count-rate formations and buried in noise in low-count-rate formations. The logging tool's AGC circuit adjusts the high-voltage supply to the photomultiplier tube or the gain of the pulse-shaping amplifier so that the digitised output always occupies a useful portion of the ADC dynamic range. This electronic AGC has no effect on the interpretation of resistivity or density values; it is purely a measurement-quality-control mechanism and does not suppress formation amplitude information in the way that seismic AGC does.
AGC parameters must be held consistent across all vintages in a 4D seismic programme or the time-lapse amplitude difference map becomes meaningless: In a 4D (time-lapse) seismic survey designed to monitor fluid movement in a producing reservoir, the goal is to measure the small amplitude change between the baseline survey (acquired before production) and the monitor survey (acquired after production). If different AGC gate lengths or target RMS levels were applied to baseline and monitor datasets, the amplitude difference map will contain a synthetic artefact that is indistinguishable from genuine fluid-substitution amplitude change. For Pembina Cardium and Weyburn Midale 4D programmes in Alberta and Saskatchewan, processing contractors are contractually required to apply matched-filter processing and identical gain functions to all vintages, with AGC explicitly prohibited unless applied identically to every trace in all vintages using the exact same gate parameters. Even then, TRA or minimal-gain processing is strongly preferred over AGC for 4D, because gain matching of AGC is impractical when geological noise levels differ between vintages.

Seismic AGC: How the Gain Function Is Applied

In practice, seismic processing software applies AGC in two variants: instantaneous AGC and RMS-window AGC. Instantaneous AGC scales each sample by the reciprocal of a smoothed absolute-value trace computed by convolution with a normalisation filter, equivalent to a very short effective gate. RMS-window AGC, described above, is more common in modern processing because the gate length is an explicit parameter that appears in the processing report and is therefore traceable. Both variants produce a gain function G(t) that is applied multiplicatively to the raw trace: output(t) = input(t) × G(t). The stored gain function can in principle be used to recover the original amplitude by dividing by G(t), but only if G(t) was recorded precisely at the time of application, which is not always the case in legacy seismic datasets processed before digital parameter logging was standard.

On legacy 1990s and early 2000s seismic lines in the WCSB, short-gate AGC was the default display parameter applied to field tapes before archival. When these vintages are retrieved for reprocessing to support AVO-compliant Cardium or Mannville mapping, the AGC must be removed before any amplitude analysis can begin. This is only possible if the raw pre-stack shot records (field tapes) are available without the AGC applied. If only the AGC-processed stacks survive, amplitude information is irretrievably lost and the data cannot be used for quantitative AVO or DHI work regardless of subsequent processing steps.

AGC in Production Automation: Automatic Gas Control

In gas production facilities and compression stations, AGC sometimes refers to Automatic Gas Control, a programmable logic controller (PLC) or distributed control system (DCS) loop that regulates gas throughput. A typical Automatic Gas Control loop measures gas flow rate at a fiscal metering point and compares it to a setpoint defined by the gas sales contract, then adjusts choke valve position, compressor speed, or separator pressure to maintain the contracted delivery rate. In Alberta gas plants operated under NOVA Gas Transmission or Alliance Pipeline interconnect agreements, Automatic Gas Control systems must respond within 30 seconds to demand signals from the pipeline operator, and are audited annually for response time compliance under AER Directive 017 (Measurement Requirements for Oil and Gas Operations).

Fast Facts

Seismic AGC was first applied to analogue seismic recording systems in the 1950s to prevent tape saturation on strong shallow reflections, and has been a standard seismic processing parameter since the shift to digital recording in the 1960s. The Society of Exploration Geophysicists (SEG) SEG-Y Revision 2 (2017) specification reserves header bytes 169 and 170 for storing the amplitude scalar applied to each trace, allowing gain values to be tracked through processing workflows. The Canadian Society of Exploration Geophysicists (CSEG) Model-Based Interpretation (MBI) committee guidelines, last updated in 2019, explicitly prohibit AGC-processed data from being used as input to Bayesian inversion workflows, acoustic impedance modelling, or any AVO cross-plot analysis submitted as part of a WCSB Crown land licence application technical report. Schlumberger (now SLB) Omega and CGG Geovation, the two dominant seismic processing platforms used by Calgary-based contractors, include AGC as a selectable module in their processing flow builders, and both flag AGC application in the trace-history record embedded in SEG-Y headers so that downstream interpreters can identify whether a given dataset has been amplitude-normalised. The dominant WCSB play fairways requiring TRA-processed seismic for AVO work are the Spirit River Formation (deep-basin tight gas), the Cardium Formation (tight oil), and the Montney Formation (tight gas-condensate), all of which have active AGC-reprocessing campaigns underway with Calgary-based seismic contractors as of 2025.

AGC in the seismic processing sense is closely related to true relative amplitude (the amplitude-preserving processing approach that corrects only for geometric spreading and attenuation without normalising trace amplitudes, and is the required alternative to AGC for any quantitative interpretation including AVO, DHI mapping, 4D monitoring, or acoustic impedance inversion), amplitude versus offset (the seismic attribute that measures how reflection amplitude changes with source-receiver distance, requiring TRA processing because AGC destroys the near-to-far amplitude gradient that defines AVO class I, II, or III behaviour), direct hydrocarbon indicator (amplitude anomaly on a seismic section caused by the acoustic impedance contrast at a gas-saturated sand, detectable only on TRA-processed sections because AGC equalises all amplitudes to the same display level and suppresses the bright spot), gain (the general term for any amplitude scaling applied to seismic data, encompassing spherical divergence correction, absorption compensation, and AGC, with AGC being the only gain function that is not physically motivated and therefore the only one that destroys interpretable amplitude information), and seismic processing (the sequence of computational steps applied to raw field seismic records to produce a final migrated image, within which AGC is one of the most commonly misapplied steps when processing objectives shift from structural display to quantitative amplitude analysis).

AGC-Suppressed DHI Anomaly on a Cardium Tight Oil Prospect

A small Calgary-based independent holds a Crown licence on a Cardium tight oil target in west-central Alberta north of Rocky Mountain House. The available vintage seismic, acquired in 1994 and processed with 150 ms RMS-window AGC, shows no obvious amplitude anomaly over the mapped structural closure, and the company is uncertain whether the trap contains oil or is wet. A land broker advises that an identical structural closure 8 km to the south, which was drilled in 2008 and found 22 m of oil-saturated Cardium sand, showed no amplitude anomaly on the same vintage AGC-processed lines prior to the well discovery. The broker argues that the AGC processing may be masking a genuine DHI on the licence block as well.