What is loudness and how do we measure it?

KEY POINTS

Loudness is Perceptual: Loudness isn't just about volume; it’s a mix of physical sound pressure and how your brain interprets factors like pitch and duration.

The Decibel (dB): We use decibels to measure sound intensity. Because the scale is logarithmic, a 10 dB increase usually sounds about twice as loud to the human ear.

LUFS vs. Raw Volume: LUFS (Loudness Units Full Scale) is the industry standard because it measures how loud a song feels, whereas older meters only measure how much energy the signal has.

Human Ear Sensitivity: Our ears are much more sensitive to midrange frequencies than deep bass. LUFS accounts for this "K-weighting" so your meters actually match what you hear.

True Peak (dBTP): This measures the absolute limit of your digital signal. If this is above 0 when you send a mix to mastering, it can cause unwanted distortion during the conversion process.

When we talk about how loud a sound is, we’re really describing how strong that sound feels to our ears. The most direct physical factor behind loudness is sound intensity, or sound pressure, measured close to the eardrum. In simple terms, it reflects how strongly the air is vibrating when a sound reaches your ear.

However, loudness isn’t determined by intensity alone. Other factors also shape how loud a sound seems, including:

❖ Frequency content (pitch)

❖ Duration (how long the sound lasts)

❖ Context (what sounds came before or after it)

Because of this, loudness is partly physical and partly perceptual. It depends on both the sound itself and how our brains interpret it.

One of the main physical characteristics connected to loudness is the amplitude of a sound wave. Amplitude describes the size of the wave’s vibrations, and in general, larger amplitudes are perceived as louder sounds. As amplitude increases, the sound carries more energy, which leads to a stronger sensation of loudness. This relationship between amplitude and loudness is why changes in sound level can feel dramatic even when the physical change seems small.

To describe these changes more precisely, loudness is measured using decibels (dB), a unit that quantifies sound pressure level (SPL). SPL tells us how intense a sound is by measuring the strength of pressure variations in the air. Decibels are also a relative measure: they compare one sound level to a reference point. The scale begins at 0 dB, which corresponds to the quietest sound a typical human can hear at a frequency of about 1 kHz. As the dB value increases, the sound is perceived as louder, giving us a standardized way to compare everything from a whisper to a rock concert.

Because of the logarithmic nature of the decibel scale, a sound that is 10 dB louder than another is often perceived as roughly twice as loud. While this rule of thumb is useful, it only applies reliably when comparing identical or very similar sounds. When sounds differ in frequency, loudness does not always add up in a simple way, especially if those frequencies fall outside the same critical band of hearing. In these cases, the auditory system processes the sounds separately, and perceived loudness can change in less predictable ways.

These limitations of basic decibel measurements become especially important in digital audio, where measuring loudness is not just about peak levels, but about how loud a signal actually feels to a listener. This is where LUFS, or Loudness Units Full Scale, comes in. Much like dBFS, which measures signal level relative to the maximum possible level in digital audio, LUFS measures loudness relative to full scale. The key difference is that LUFS is designed to reflect perceived loudness, not just raw signal strength. A loudness unit (LU) is essentially equivalent to a decibel but with an important distinction: it accounts for how the ear and brain respond differently to various frequencies. Because of this, if you turn a sound or mix up by 3 dB, its LUFS value will also increase by about 3 LU. However, if you change the tonal balance using EQ without changing the overall level, the LUFS reading may still change. This happens because our ears are more sensitive to certain frequencies than others, and LUFS measurements are designed to reflect that sensitivity. 

Another way to measure loudness is RMS. RMS (root mean square) measures the average power of a signal, providing a more accurate sense of its sustained intensity. Unlike a peak meter, which only shows the maximum amplitude, RMS reflects the overall density and continuous strength of an audio signal. However, this measurement shows the overall energy of a sound but doesn’t account for frequency. That means it treats a low-frequency sub-bass (like 40 Hz) the same as a midrange tone (like 2 kHz), even though our ears hear them very differently. As a result, two songs with the same RMS can sound very different in loudness because human hearing is more sensitive to midrange frequencies than to deep bass. This is why LUFS improves on RMS by incorporating two key concepts:

❖ K-weighting filter: an equalization filter applied before measuring loudness. It reduces low frequencies and emphasizes upper-midrange frequencies, mimicking the sensitivity of the human ear (based on Fletcher-Munson curves);

❖ Multiple Measurements: LUFS offers different readings depending on the context:

❖ Integrated LUFS (I): The average loudness across the entire track. This is the main target when mastering music for streaming.

❖ Short-Term LUFS (S): The average loudness over the last 3 seconds. Useful for checking the consistency of different sections.

❖ Momentary LUFS (M): The average loudness over the last 400 milliseconds. Helps identify the moments of highest intensity.

In addition to LUFS, True Peak (dBTP) meters are important. They measure the peak level that will come out of a digital-to-analog converter, which can cause distortion during playback or conversion. It is fairly commonplace for modern masters to hit true peaks above 0, and the resulting effect has become a part of the sound of modern music in the streaming era. However, the decision to have true peaks over 0 or not is best left in the hands of a capable mastering engineer. Additionally it's important to keep dBTP in mind when delivering a mix to mastering. A 24-bit mix that already clips at the output can yield a less clean master, to avoid this we recommend delivering your mix at 32-bit or making sure your dBTP output is below 0.

Mastering for streaming

LUFS has become the standard for measuring loudness on streaming platforms, TV, and film. Whether it’s Netflix, Amazon, or Dolby Atmos mixes, audio often needs to meet specific targets like integrated loudness, true peak, or loudness range.

Post written by Ruxandra Dedu, edited by Kristian Montano.