Volume, Levels, and Loudness

Summary:

Volume is what the user controls for output loudness, signal level is the measured strength of the audio signal (peak/RMS), and loudness is how loud the ear perceives it (influenced by frequency and psychoacoustics) . Understanding these differences and how they interplay is key in audio production, especially today with loudness normalization on most platforms

Volume, Levels, and Loudness: A Primer

Volume in audio typically refers to the user-controlled loudness of the output. In other words, it is the knob or slider that makes the system sound louder or quieter. It directly adjusts the output amplitude sent to speakers or headphones . For example, when you turn up the volume on your speakers or in your media player, you increase the electrical signal driving the speakers, making the sound louder. (Audio engineers sometimes use “volume” informally to mean perceived loudness, but strictly speaking it’s the output control the listener adjusts .)

Audio level refers to the strength or amplitude of the signal at any point in the audio system. It’s a measurement of how “big” the waveform is in electrical or digital terms. Levels are usually expressed in decibels (dB) relative to some reference (for example, dBFS in digital audio or dBu in analog). In practice, engineers monitor levels using meters in a DAW or mixer. Peak meters report the instantaneous maximum signal, while RMS meters report the average energy over time. The peak level shows the loudest instantaneous sample in a waveform (important for avoiding clipping), whereas the RMS level (root-mean-square) corresponds to the signal’s average power and correlates more closely with perceived loudness. In summary, “level” is the technical measure of signal amplitude in the system.

Loudness and Perception

Loudness is the perceived intensity of sound - how loud it seems to the human ear. Crucially, loudness issubjective and depends on frequency. Humans do not hear all frequencies equally. We are most sensitive inthe midrange (around 2-5 kHz) and much less sensitive to very low or very high tones . This means a 100Hz tone must be played at a higher sound pressure than a 1 kHz tone to sound equally loud. Therelationship between frequency and perceived loudness is captured by equal-loudness contours (also knownas Fletcher-Munson curves) . These curves show, for each frequency, the SPL (sound pressure level)needed for an average listener to perceive a given loudness. For example, at low volume levels the ear isless sensitive to bass and treble, which is why some audio systems include a “loudness” compensationbutton that boosts bass and treble at low volume . (These contours are measured in phons; e.g. a 1 kHztone at 60 dB SPL is defined as 60 phons.)

In short, loudness is not a fixed physical quantity but a psychoacoustic one: it depends on the sound’sfrequency content and on how the human ear-brain system responds. Equal-loudness research (Fletcher &Munson, ISO 226, etc.) underpins modern loudness metering. As one source explains, “Equal-loudnesscontours [are] sound-pressure levels over the frequency spectrum for which a listener perceives a constantloudness”.

Why Loudness Matters

Loudness consistency is very important in production and broadcasting. If one song or ad is much louderthan the last, listeners will instinctively grab the volume knob, which is a poor listening experience. To avoidthis, broadcast and streaming standards have adopted loudness normalization. For example, the European12 13456 5761R128 standard calls for a target loudness (integrated over programs) of around -23 LUFS for TV/radio, sothat all content plays at a similar perceived level. Similarly, streaming services like Spotify and YouTubenormalize tracks to around -14 LUFS. By enforcing these targets, loudness normalization means thatoverly “hot” (extremely loud) masters will be turned down, and very quiet masters may be boosted, leadingto a more uniform listening experience across songs or shows.

Consistent loudness also prevents listener fatigue. Over-compressed, maximally-loud music can sound flatand tiring. Excessive loudness (with little dynamic range) may briefly grab attention, but studies andengineers have noted that vibrant recordings with a healthy dynamic range are more pleasing in the longrun. In fact, the extreme loudness levels of past recordings (the “loudness war”) often reduce clarity andlistener comfort.

Normalizing loudness makes quieter details more audible and avoids the earstrain caused by constant maximum levels. In broadcasting, consistent levels also protect listeners’ hearing(no sudden blasts of volume) and honor regulatory requirements on loudness levels.

The Loudness War and Normalization

Starting in the 1980s-90s, the music industry engaged in a “loudness war,” mastering tracks ever louder tostand out on radio and in stores . This meant heavy compression and limiting that squasheddynamics so that the softest and loudest parts became closer together. As one audio expert summarizes:“Overcompression squeezes the life out of music, reducing its dynamic range—the difference between thesoftest and loudest parts. The result is music that might be loud but lacks vibrancy and depth” .Famously, albums like Metallica’s Death Magnetic (2009) were criticized as suffering from extreme limitingand distortion as a result of this push for loudness.

However, with modern loudness normalization, the incentive for such extreme loudness has greatlydiminished. Streaming platforms and broadcasters automatically adjust playback loudness to meet targets.For instance, Spotify plays every track at roughly -14 LUFS, so if you upload a track mastered to -4 LUFS(extremely loud), Spotify will simply turn it down. Conversely, a very quiet track might be turned up(via a limiter) to reach the target. In effect, this means sacrificing dynamic range to be loud has no benefit:a highly compressed track will sound the same average level as a more dynamic one, but with less punchafter normalization . Many artists and engineers now tailor masters for streaming targets (e.g. around -14 to -16 LUFS) instead of aiming for 0 dBFS peak level. In short, loudness normalization has helped end theloudness war by removing the “loudest is best” advantage.

For Advanced Readers

LUFS (Loudness Units relative to Full Scale): LUFS is the standardized unit for measuring perceivedloudness in digital audio. It is essentially the same as LKFS (Loudness, K-weighted, relative to FullScale) defined by ITU-R BS.1770 . A LUFS measurement takes into account the human ear’sresponse (via a K-weighted filter) and averages level over time. Unlike a simple dBFS peakmeasurement, LUFS attempts to quantify how loud a piece of audio will subjectively appear. Forexample, many broadcasters now use -23 LUFS (integrated over a program) as a target , andstreaming platforms may target around -14 to -16 LUFS per track . In LUFS, 0 LUFS is themaximum digital level; typical content will be negative (quieter) numbers. The difference betweentwo LUFS values (in LU) is equivalent to decibels.

Peak vs. RMS vs. True Peak levels:

Peak Level is the highest instantaneous amplitude in the signal. In digital systems this is measured per sample as dBFS (0 dBFS being the maximum). Peak meters are crucial to avoid clipping, but they do not tell you about perceived loudness.

RMS Level is the root-mean-square average of the signal over a time window. RMS correlates more closely with loudness because it reflects the average power. In practice, we often talk about RMS or “average” level when considering loudness. For example, a DAW meter might show both peak and RMS; the RMS meter moves more slowly and indicates how loud the signal feels on average .

True Peak Level goes a step further by estimating the actual continuous waveform peak after digital- to-analog conversion. Because digital audio is sampled, the true analog peak can exceed the sample values if the waveform is between samples. True Peak meters oversample the signal to catch these inter-sample peaks. It’s measured in dBTP and is used in standards to ensure no clipping occurs even after filtering. In short, a track may show -1 dBFS peak sample level, but its true peak could be slightly above 0, which is why some recommend leaving additional headroom for true peaks.

Perceptual loudness models: Modern loudness meters (like those in BS.1770 or EBU R128) combine RMS averaging with frequency weighting to mimic human hearing. Specifically, the audio is passed through a K-weighting filter (a defined curve that boosts low frequencies slightly and attenuates very low and very high frequencies) . Then the filtered signal’s energy is measured (RMS) in short blocks. Very quiet blocks can be gated out so that silence doesn’t drag down the average. The remaining block values are averaged to give an overall LUFS number . In practice, this means loudness meters simulate the Fletcher-Munson effect and our ear’s frequency sensitivity. The result (often called “Integrated Loudness” in LUFS) gives one number representing how loud the whole track sounds. (See ITU-R BS.1770 for details; it’s the basis of most LUFS measurements .)

Analog signal measurements: In the analog domain, audio is a voltage waveform. Levels are measured in decibels relative to a reference voltage. Common units are dBu and dBV. By definition, 0 dBu ≈ 0.775 volts RMS, and 0 dBV = 1.0 volt RMS. (Pro audio conventionally uses +4 dBu as its nominal reference level, while consumer gear often uses -10 dBV .) These references mean, for example, that +4 dBu (about 1.23 V) is roughly 12 dB hotter than -10 dBV (about 0.316 V). dBu/dBV describe average (RMS) voltages, so analog line levels are inherently tied to RMS power.

VU meters vs. digital peak meters: The classic analog VU (Volume Unit) meter responds relatively slowly (on the order of 300 ms) and shows average level. It was calibrated so that 0 VU corresponds to a standard reference level (often +4 dBu) . VUs were designed to mimic perceived loudness. In contrast, modern digital peak meters read instantaneous sample peaks (dBFS). A peak meter in a DAW can jump on very brief transients, whereas a VU meter would barely budge. As a result, two signals might both peak at -1 dBFS but have very different loudness: the one with heavier compression will have a higher RMS and sound louder . Thus, peak meters are essential for catching overloads, but VU/RMS metering is needed to gauge how loud something feels.

Analog voltage references (dBu/dBV): Recall that analog gear often specifies levels by decibels relative to 0.775 V (dBu) or 1 V (dBV) . For example, professional line-level gear expects around +4 dBu. When connecting analog and digital, engineers often set the converter so that 0 dBFS corresponds to a chosen analog dBu level. (For instance, many interfaces use +4 dBu = 0 dBFS.) Getting these reference levels right is crucial in hybrid studios to avoid clipping or noise.

Metering standards (EBU R128, ITU-R BS.1770, etc.): In broadcasting, specific loudness standards are enforced. In Europe, EBU R128 (2010) mandates that programs be normalized to -23 LUFS, with short-term and momentary tolerances, and limits true peaks (often -1 dBTP) . The international ITU-R BS.1770 series defines the algorithm (K-weighting + gating) for measuring LUFS. The US ATSC A/85 standard is similar (-24 LKFS, essentially the same). These standards ensure that all content (voice, music, ads) on TV/radio has a consistent perceived loudness. (For streaming and music services, various recommended targets exist: Spotify/YouTube ~-14 LUFS, Apple Music uses - 16 LUFS, etc. .) In mastering for broadcast or streaming, engineers routinely use LUFS meters and true-peak meters to comply with these standards.