Loudness Range

History

In 2006, the ITU published the initial version of their BS.1770 recommendation, which grew to shape the whole media industry. The recommendation specifies a method for measuring the loudness of broadcast material.

Upon that recommendation, the EBU, which stands for the European Broadcasting Union, implemented their recommendation, R128, which sets targets for program loudness in broadcast context. The standard recommends normalizing content to an integrated loudness of -23 LUFS, and further recommends that the true peak shall be kept at or below -1 dB TP. Effectively this recommendation had the power to stop the loudness wars, and in broadcast, it did.

However, when the integrated loudness measures the loudness across the whole track, it doesn’t measure consistency. That’s where the LRA comes in.

The unit

LRA uses a unit called LU. It stands for Loudness Unit, which maps directly to decibels when measuring a change in level. LUFS on the other hand, maps to dB FS, which represents the Full Scale in the digital domain. For anything other than a change in level, the units measure completely separate things.

The equal loudness curve

The Loudness Unit uses an equal loudness curve, which is usually called the K-weighting. It represents how sensitive our ears are towards certain frequencies. The K-weighting is somewhat comparable to the Fletcher-Munson equal loudness curve, but is a simplification of it.

The frequency-weighting is needed because when measuring only an overall level of the material, you could get equivalent results for sounds that are perceived as louder or softer because of their frequency content. The K-weighting enables us to get comparable results for different-sounding materials.

Consistency

Normally, an audio track has some variance in its loudness between louder and softer passages. In a music file that’s very compressed, the variance might be as low as a few decibels. In an audio drama or a film, the variance between the loudest and softest passages might be over 15 decibels. In classical music, the variance might be even higher.

LRA presents this variance with a single number: for example, the music track has an LRA of 3, while the film has an LRA of 15. Neither of these are inherently right or wrong, but ear fatigue tends to creep in as the Loudness Range gets narrower. On the other hand, the listener might feel the need to turn the volume up and down if the Loudness Range is very wide.

Targets

When deciding what LRA to target, it’s worth remembering how the majority of the audience consumes your content. For movies it’s a bad idea to target an audience who watch it on their mobile devices (though it seems to be a growing trend, movies should still be produced for a proper viewing environment), and for spoken content, such as news, it would be a bad idea to have a very wide Loudness Range. Most audio content nowadays is probably consumed using AirPods, or a phone speaker.

At Castcore, we target an LRA of 7 for spoken content. LRA 7 still has variance between the loudest and softest passages, but is consistent enough to allow listeners to listen to it in a car, on a bus, from a phone speaker, or from a Homepod while they’re doing their dishes.

While we aim to achieve a 7 LU Loudness Range, we never expand the loudness range. If you create content with very consistent dynamics, we respect your choice and only apply the linear loudness normalization, which simply affects the overall volume of the entire track.

We set these targets because your content deserves to sound best.