01The perception isn't in your head — it's built in at the transmitter
Spin the dial and you will notice it almost immediately: one station hits harder than its neighbours. The music feels closer, the presenter's voice arrives with more presence, and when you switch back to a quieter station you reach for the volume knob without thinking. The difference is real, measurable, and entirely deliberate. It is the product of a decades-long arms race in audio processing — an industry-wide pursuit of perceived loudness that transformed the technology sitting between a microphone and a transmitter, and in doing so changed what radio sounds like.
02What loudness actually means
Before the engineering, the physics. Loudness is not the same as level. Level is a measurement — the actual voltage or digital value representing the audio signal at a given moment. Loudness is what the human ear perceives, and the two diverge in important ways. A sustained, dense signal at a given peak level will sound louder than a signal with the same peak level but a wider dynamic range, because the ear integrates over time. The difference between average level and peak level is often described as crest factor or peak-to-loudness ratio, while dynamic range is the span between quietest and loudest passages, and it is the central battleground of broadcast audio processing.
A live orchestra has enormous dynamic range — the difference between the quietest passage and the loudest might span forty or fifty decibels. A heavily processed radio station might run its music at an average level only a few decibels below its peaks. That is not how the music was made; it is how the station chose to transmit it. The technical measure that captures this relationship is called the loudness range, and the tool that shrinks it is the compressor.
03The toolkit: compression, limiting and beyond
A compressor is a device — hardware or software — that automatically reduces the gain of a signal when its level exceeds a set threshold. Turn the signal up a lot and the compressor turns it back down a little, or a lot, depending on how it is set. The ratio control determines how aggressively it does this: a gentle 2:1 ratio means that for every two decibels above the threshold, only one gets through; an extreme 20:1 ratio is close to a brick wall. The attack and release times determine how quickly the compressor responds to a transient and how quickly it lets go — settings that have profound consequences for the feel of the sound. Slow attack times let the initial punch of a drum hit through before clamping down; fast attack times catch everything and make the signal feel dense and forward.
A limiter is a compressor with a very high ratio and a very fast attack — effectively a ceiling that the signal cannot exceed. Broadcast limiters ensure a station never overshoots the maximum permitted deviation for its transmission format, which would cause distortion or regulatory trouble. But limiting is not just a safety catch. When the ceiling is set carefully, a limiter also raises the average level toward that ceiling, because the peaks that would otherwise demand headroom have been caught and flattened.
Broadcast stations stack these tools. The signal chain might include a multi-band compressor, which splits the audio into separate frequency ranges — typically four or five bands — and compresses each independently. This means a heavy bassline does not cause the high frequencies to duck, which is what a single-band compressor would do. The result is a more consistently dense sound across the whole spectrum. After the multi-band stage, a final brick-wall limiter catches anything that sneaks through. Then there is the clipper — a harder, simpler process that simply cuts the tops off waveform peaks. Clipping introduces harmonic distortion, but at low levels that distortion is not obviously audible, and it allows the average level to be pushed higher still without triggering the limiter.
That is not how the music was made; it is how the station chose to transmit it.
The software and hardware that does all of this in a single outboard unit is called a broadcast audio processor. The manufacturers who defined the category — companies such as Orban (maker of the Optimod line) and later Omnia — became names spoken with reverence in engineering rooms. An Optimod box in a studio rack is a signal that a station takes its sound seriously, for better or worse.
04The arms race
The competitive logic is almost too simple. In the 1960s and 1970s, FM broadcasters discovered that heavier processing made a station feel more present to a listener with the radio on in a noisy kitchen or a moving car. A louder, denser sound cut through ambient noise more reliably. Audience research appeared to confirm that listeners preferred the louder station when scanning, even if they could not articulate why. So stations pushed their processors harder. Neighbouring stations, losing listeners at the margins, pushed harder too. Within a generation, FM audio processing had escalated far beyond what the engineers who designed the original equipment intended.
The same logic gripped digital audio in the 1990s and 2000s — not only in broadcasting but in record mastering, where the phenomenon became known as the loudness war. Albums were mastered progressively louder with each cycle of competition, the waveforms increasingly resembling solid blocks rather than dynamic performances. Radio processors then received that pre-compressed material and compressed it further still.
The cost is real. Heavy processing smears transients — the sharp initial attack of a snare drum, the pluck of a guitar string, the consonants in speech. It creates pumping and breathing artefacts when the compressors' release times interact with the music. It can cause listening fatigue over extended periods, though the evidence on this is more anecdotal than clinical. And it imposes a particular character on everything the station plays, regardless of whether the material suits it. A sparse folk recording and a dense electronic track get treated to the same density target, and neither necessarily benefits.
05Loudness normalisation and the new landscape
Streaming services and podcast platforms introduced a partial correction. Spotify, Apple Music and others adopted loudness normalisation, measuring each track to a target loudness level — specified in LUFS, or Loudness Units relative to Full Scale, a standard defined in the EBU R128 recommendation — and adjusting playback gain accordingly. A heavily mastered track is turned down to match a quieter, more dynamic one. The incentive to over-compress is removed, and some mastering engineers have responded by pulling back.
Broadcast radio has moved more slowly. Regulatory bodies in several countries set limits on audio levels, and the EBU R128 standard has been adopted by public broadcasters across Europe as a production and delivery target. But commercial radio in competitive markets still tends to process aggressively, because the logic of the dial — the listener scanning between stations — has not gone away, and because habits built over decades are not abandoned quickly.