EBU Recommendation R 128 establishes a predictable and well-defined method to measure
the loudness1 level for news, sports, advertisements, drama, music, promotions, films etc., throughout the broadcast chain, and thereby helps professionals to create robust specifications for ingest, production, playout and distribution to a multitude of platforms.
In addition to Programme Loudness, R 128 introduces two more audio descriptors, Loudness Range and Maximum True Peak Level. All three are designed to work together,
forming a set of essential descriptors that characterise an audio signal.
Loudness metering and loudness normalisation signify a true audio levelling revolution. Furthermore, this loudness-levelling paradigm affects all stages of an audio broadcast signal, from acquisition to distribution and transmission. Thus, the ultimate goal is to harmonise audio loudness levels within broadcast channels as well as between channels to achieve an equal universal loudness level for the benefit of the listeners. To be clear: the loudness level can (and should!) still vary according to artistic and technical needs within a programme. Rather, the loudness normalisation method uses the average loudness of a programme, to make the level consistent between programmes and between channels.
The experience of several EBU Members has shown that working with the loudness paradigm is both liberating and satisfactory. The fight for “Who is the loudest” disappears, mixes can be more dynamic, there are fewer dynamic compression artefacts, such as “pumping”, and thus there is an overall increase of audio quality! Programme makers who favoured dynamic mixes in the past are now relieved from potential compromises because their programme no longer sounds softer than more-compressed ones. With loudness normalisation, this compromise is gone.
The origin of the problem – the “loudness war”
Audio metering in broadcasting today is based typically on quasi-peak programme meters (QPPMs– more usually known as just PPMs). It is “quasi” because of its finite reaction time of 10 ms (although 5 ms is also found). In practice, this means that signal peaks shorter than this reaction time won’t be displayed correctly, if at all (for
example, transients such as those created when keys are jangled). In order to provide headroom for these transients, which one wouldn’t see on the meter, but which should nevertheless be there so as to contribute to the “openness” of the audio
signal, the agreed Permitted Maximum Level (PML) was set at –9 dBFS.
This value was based on the familiar – and inmany places, still extant – method of delivering sound to the home on an FM carrier. The carrier's maximum deviation for TV was standardized in many countries at 50 kHz and the PML at 30 kHz deviation (equating to –9 dBFS), which thus allowed 20 kHz, or 4.4 dB of headroom.
However, commercial pressures have taken a hold and the response to the pressure to stand
out from the competition has been to be louder than it. Modern peak metering and powerful
dynamic-range processing allowed organizations to realign the PML to equate to the maximum deviation (50 kHz) of the FM carrier. All the transients have to be chopped at the PML to avoid distortion, but that has been thought an acceptable compromise by those who have implemented it. When someone at home switches between one of these stations and a station which has not joined the “war” or is taken to a loud advertisement in a well-balanced programme with a wider dynamic range, the audio level jumps and there is a grab for the remote control to adjust the sound to a more acceptable level. In the case of loud advertisements, the sound then has to be re-adjusted when the main programme returns. It is no wonder that so many complaints are received by the broadcasters.
Other people solve the problem by muting the sound during the advertisements and so the message is at least greatly diluted.
A Standard Emerges – and the EBU develops it
The International Telecommunication Union (ITU) recognized the problem and its work gave rise to ITU-R BS.1770 . The purpose of that standard was to establish an agreed algorithm for the measurement of loudness and the true peak levels of programmes. It is a robust standard which has the benefit of a simple implementation. In brief, it defines a “K-weighting” curve (a modified second-order high-pass filter) which forms the basis for
matching an inherently subjective impression with an objective measurement.
This curve is applied to all the channels (except the Low-Frequency Effects (LFE) channel
which is discarded from the measurement), the total mean square energy is then calcu-
lated (with different gain factors for the front and surround
channels) and the result is displayed as “LKFS”
(Loudness, K-Weighting, referenced to digital Full Scale).
For relative measurements, Loudness Units (LU) are used,
where 1 LU is equivalent to 1dB.
BS.1770 also defines and recommends the use of a true peak meter for measuring peaks. Such a meter runs at a higher sampling rate than the audio signal (usually 4x oversampling) to catch inter-sample peaks which might otherwise exceed 0 dBFS and thus cause distortion later in the chain.
ITU-R BS.1770 provides the basis for EBU Recommendation R 128 which
extends the ITU standard by actually defining a specific Target Level for loudness normalisation as well as a gating method to improve the loudness matching of programmes which contain longer periods of silence or isolated utterances.
The EBU’s development was required to accommodate the needs of programme makers, with particular regard to having a means to measure complete mixes (rather than just one component, such as the dialogue) and the loudness range of the programme.
To do this, the EBU has specified three new parameters:
- Programme Loudness
- Loudness Range
- True Peak Level