# Automatic gain control **Automatic gain control** (AGC), sometimes called automatic volume control, is a closed-loop feedback circuit inside an amplifier or a chain of amplifiers that adjusts its own gain, moment to moment, to hold the amplitude at its output within a workable range despite changes in the amplitude arriving at its input. This article carries no interactive sketch of its own; its Microsims section below points to the [[Radar]] sketch, whose echoes vary in amplitude by many orders of magnitude, as the clearest illustration of the dynamic-range problem AGC exists to solve. A [[Radio_receiver|radio receiver]] without AGC would swing from a bare whisper on a distant station to an uncomfortably loud blast on a strong local one, and would swing again within a single station as its signal faded and returned; AGC instead measures the strength of what the detector is currently receiving and feeds that measurement back to turn the amplification down when the signal is strong and up when it is weak, so a listener sets the volume once and the receiver holds it roughly there regardless of which station, or how much fading, follows. The same feedback idea, sampling an output and correcting a gain, reappears far beyond broadcast radio: in radar receivers, in recording equipment, and even in biological senses that work across a similarly enormous range of input strengths, all covered in Example use cases below. Every version shares the same trade-off, covered in Recovery times, between reacting quickly enough to a sudden change and reacting so quickly that the correction itself becomes an audible or visible fault of its own. ## How it works Every automatic-gain-control loop needs three parts wired as a loop rather than as a one-way chain: a variable-gain amplifier, or an attenuator, placed ahead of or among the stages whose combined gain is being controlled; a detector that turns whatever signal is currently present into a slowly varying measure of its strength, often the same [[Envelope_detector|envelope detector]] already demodulating an amplitude-modulated signal, or a dedicated peak or average detector where no such stage already exists; and a feedback path, usually little more than a resistor-capacitor smoothing network, that turns the detector's output into a control voltage the variable-gain stage responds to. Turning that control voltage up lowers the gain and turning it down raises the gain, so a stronger input ends up commanded to less amplification and a weaker one to more, opposing the very change in output level that provoked the correction; this is [[Negative_feedback|negative feedback]] applied to gain itself rather than to a voltage or a position. A loop built this way behaves, in miniature, like any other proportional feedback controller: the more strongly a departure from the target output level is fed back as a gain correction, the smaller the residual departure that survives once the loop settles, though some residual departure survives at any finite loop gain, and only an integrating term added to the feedback, rather than raw gain alone, closes it away entirely. Because an ordinary AGC loop is exactly this kind of proportional controller, it never holds a receiver's output perfectly constant either; it only narrows the swing a listener notices to something the ear, or the next stage of [[Signal_processing|signal processing]], can tolerate. How wide a swing the loop must tame in the first place is set at the bottom by the receiver's own noise floor, since a signal weaker than the thermal noise a front end adds on its own carries no information worth preserving regardless of how much the loop amplifies it, and the same noise-figure accounting that bounds any receiver's sensitivity bounds the useful bottom of an AGC loop's working range.[^agc-noisefloor] ## Example use cases ### AM radio receivers Amplitude modulation carries its information directly in the carrier's own amplitude, so an [[Amplitude_modulation|AM]] receiver's envelope detector is already measuring exactly the quantity AGC needs: the same diode that recovers the audio also yields a direct-current level proportional to how strong the tuned station currently is, and that level is fed back, with little extra circuitry, to control the gain of the intermediate-frequency or radio-frequency amplifier stages of a [[Superheterodyne_receiver|superheterodyne]] front end ahead of it. This is the original and still the clearest use case, since without it an AM receiver's volume would vary drastically between a strong local station and a distant, fading one, and even within one station as its own signal faded and returned. ### FM radio receivers A [[Frequency_modulation|frequency-modulated]] signal already carries its information in the carrier's frequency rather than its amplitude, and the limiter stage ahead of an FM discriminator deliberately clips away amplitude variation before demodulation, so FM reception is far less dependent on AGC for its audio quality than AM reception is. A gain-control loop is nonetheless still used ahead of that limiter, both to keep earlier stages from overloading on a very strong local signal and to hold the limiter's own input within the range it needs to limit effectively, so AGC and limiting work together rather than one replacing the other. ### Radar A [[Radar|radar]] receiver faces the widest amplitude range of any of these examples: echo power falls with the fourth power of a target's range, so a distant aircraft can return a signal many orders of magnitude weaker than a nearby one on the same sweep, exactly the range compression the neighbouring Radar article's own sketch shows directly through its faint, distant blips and its bright, near ones. Left unmanaged, that range would either saturate the receiver on strong, close returns or bury weak, distant ones below the display's own resolution, so radar AGC is normally paired with, or replaced by, faster instantaneous processing such as [[Constant_false_alarm_rate|constant-false-alarm-rate]] normalization, which adjusts a detection threshold locally against nearby [[Clutter_(radar)|clutter]] rather than adjusting one gain slowly across an entire sweep. ### Audio/video Broadcast and recording audio chains use AGC to keep a programme's average level within the range tape, transmission or a digital format can carry without either clipping on the loudest passages or falling into background noise on the quietest ones, and analogue television receivers use a related loop, keyed to the picture signal's own synchronization pulses, to hold the sync level constant so the picture stays locked regardless of the received signal's overall strength. ### Vogad A voice-operated gain-adjusting device, or Vogad, is a variant built specifically for speech rather than for a broadcast carrier: it raises the gain during pauses so a following word is not clipped by a slow-reacting loop still adjusted for the previous, louder syllable, and it is intended to maximize a listener's ability to make out speech over a noisy or fading channel rather than to preserve the original signal's own natural loudness variation the way a broadcast AGC circuit tries to.[^cn-vogad] ### Telephone recording Recording equipment built for dictation or for logging telephone calls uses AGC for a related reason: speakers vary widely in loudness and in distance from a microphone, and a fixed recording level would either clip a nearby, loud speaker or record a quiet, distant one too faintly to transcribe, so the recorder's own gain is instead adjusted continuously to keep speech within its recording medium's usable range regardless of who is speaking or from how far away. ### Biological Sensory systems face the same problem AGC solves, over a range engineered AGC would find extreme: the human eye adjusts to light levels spanning many orders of magnitude between night and full daylight, and the ear's own protective reflexes and neural adaptation similarly compress an enormous range of sound pressure into a workable operating range for the receptors and nerves that follow, an analogy long drawn between engineered gain control and the adaptation built into biological sense organs.[^cn-biological] ## Recovery times Every AGC loop must choose how fast to react in each direction, and the two directions are rarely symmetric. Attack, the speed at which the loop reduces gain when a signal suddenly grows stronger, is normally made fast, often a few milliseconds or less, so a sudden strong signal does not overload the following stages or clip before the loop can respond. Release, or decay, the speed at which gain recovers once a strong signal fades or ends, is normally made much slower, tens or hundreds of milliseconds to a few seconds, because a release fast enough to track individual cycles of the signal itself would modulate the gain at an audible or visible rate of its own, called pumping, rather than simply following the signal's slower overall envelope. Both times are ordinary RC time constants in most implementations, exactly the same exponential charge-and-discharge behaviour that governs any single capacitor charging through a resistor: a capacitor's voltage covers roughly 63 percent of the remaining distance to its final value in one time constant, `t = RC`, regardless of how large that final value is,[^agc-rc] which is why designers reach for two different resistors, one for the fast charging (attack) path and a much larger one for the slow discharging (release) path, rather than a single shared value that could only ever compromise between the two. A control voltage built this way also keeps its average level essentially unaffected by exactly which time constants are chosen, since a low-pass network's response at zero frequency does not depend on where its cutoff sits;[^agc-rc2] only the speed of the correction, not the long-run gain the loop settles on, is what the choice of attack and release times actually trades away. ## Microsims This article carries no p5.js sketch of its own, and no three.js companion is planned for it in this run. The dynamic-range problem described throughout this article, and the trade-off in Recovery times between a fast correction and a stable one, are both easiest to see in a system built to push amplitude toward its extremes rather than to hold it constant. The sketch carried by the neighbouring Radar article, itself one of the Example use cases described above, is exactly such a system: it renders the fourth-power falloff of echo strength with range that AGC exists to compress, letting the same many-orders-of-magnitude swing this article describes in words be watched directly, blip by blip, as a target sweeps toward and away from the receiver. *Try:* in the [[Radar]] sketch, compare a near target's echo against a distant one on the same sweep - the same many-orders-of-magnitude range an AM receiver's AGC loop exists to compress before it ever reaches a listener's speaker. *Try:* in the [[Sonar]] sketch, lengthen the range to a target and watch how much fainter its echo becomes against the background - the same shrinking signal-to-noise margin an AGC loop's lower limit runs into once a signal sinks toward the receiver's own noise floor. ## Wikipedia : Wikitube **Strict pair:** [Wikipedia](https://en.wikipedia.org/wiki/Automatic_gain_control) : [Wikitube](https://en.wikitube.io/wiki/Automatic_gain_control) Skeleton mirrored at revision 1344456175. Prose, emphasis and the microsims are Wikitube's own. ## See also - [[Radio_receiver]] - [[Envelope_detector]] - [[Radar]] - [[Dynamic_range_compression]] - [[Companding]] - [[Negative_feedback]] ## References The RC charge-and-discharge law and the low-pass network's unchanged response at zero frequency are standard textbook results and are not separately footnoted beyond the citations below, per the Wikitube style guide §6.1. Page numbers are PDF pages of the open editions linked below. [^agc-noisefloor]: Ellingson, S. *Radio Systems Engineering, Revised First Edition*. 2023, pp. 96-101 (PDF pages). Open Textbook Library: https://open.umn.edu/opentextbooks/textbooks/radio-systems-engineering . CC BY-NC. [^agc-rc]: Tiberius, C.; Mulder, M. *Engineering Signal Analysis: From Fourier to filtering: Theory*. 2026, p. 95 (PDF page). Open Textbook Library: https://open.umn.edu/opentextbooks/textbooks/engineering-signal-analysis-from-fourier-to-filtering-theory . CC BY. [^agc-rc2]: Johnson, D. *Fundamentals of Electrical Engineering I*. 2014, p. 146 (PDF page). Open Textbook Library: https://open.umn.edu/opentextbooks/textbooks/fundamentals-of-electrical-engineering-1 . CC BY. [^cn-vogad]: Citation needed: a primary specification or standards document naming the Vogad and describing its intended telecommunications use has not been pinned down in this pass. [^cn-biological]: Citation needed: a primary physiological source for the specific adaptation mechanisms in the eye and ear referenced here would confirm the range each covers and how each is achieved. **Further reading** — the open textbooks this article draws on, since the pair carries no separate Further reading heading of its own: - Steven Ellingson. *Radio Systems Engineering, Revised First Edition*. Open Textbook Library: https://open.umn.edu/opentextbooks/textbooks/radio-systems-engineering - Christian Tiberius; Max Mulder. *Engineering Signal Analysis: From Fourier to filtering: Theory*. Open Textbook Library: https://open.umn.edu/opentextbooks/textbooks/engineering-signal-analysis-from-fourier-to-filtering-theory - Don Johnson. *Fundamentals of Electrical Engineering I*. Open Textbook Library: https://open.umn.edu/opentextbooks/textbooks/fundamentals-of-electrical-engineering-1 <!-- Hubs: Signal_processing. Portals: PORTAL_Radio. Radio portal wave 1 · 2026-09-17 · drafted. -->