When broadcasting media that contains audio to the masses, certain guidelines have been put in place to ensure the quality of audio content is safe and maintained to a high standard. There are a multitude of different frameworks used around the world all with differing standards.
The global standard for broadcasting audio is ITU.BS1770-3 which was developed by the International Telecommunications Union, hence the abbreviation. There are other standards which build off ITU.BS1770 (the original recommendation, currently at its third revision)
ATSC A/85 which is used by the United States (target loudness -24 LKFS, only digital)
EBU R128 which is currently in use by Europe (target loudness -23 LKFS uses a relative gate)
TR-B32 (based off of ITU.BS1770-2) which is used by Japan (target loudness -24 LKFS also uses a gate)
OP-59 which is used here in Australia. (recommends a speech based as well as universal approach to audio normalisation.)
Because these standards are all derivatives of the same principle they only differ slightly. All use the same techniques and elements, such as
- LKFS, LUFS & LU
- Loudness Range, Program Loudness & Descriptors
- Target levels
K-Weighting is a method used to determine the relative loudness of an audio source. Two different sources can be the same technical loudness but the way we as humans perceive them could be very different. Essentially a filter curve is applied to all available audio channels to normalise the difference in perceived volume.
Loudness K-weighted Full Scale (LKFS) and Loudness Units Full Scale (LUFS) are the exact same thing just named by different standards (EBU R128 uses LUFS, most other standards use LKFS). Loudness Units (LU) is a relative measure created for ease of use when setting the broadcast target level. All three of these units equate to one dB.
Loudness Range (LRA) is a measurement of the entire audio clips loudness measured in LU. The way this is calculated is by measuring the softest and loudest parts of the file in question. To avoid discrepancies due to extreme occurrences in the audio, the lowest 10% and top 5% of the total LRA is excluded from the final measurement.
Program loudness is the average loudness of the entire audio in question. This is portrayed using LKFS or LUFS.
Gating is a tool used to further increase the accuracy of the Program loudness measurement. When the source material drops below a level of -10 LU (relative to an ungated measurement of the same material) the measuring of loudness is paused until the level returns above -10 LU. This allows for more dynamic audio to be presented at the same level as other less dynamically rich media such as movies compared to pop music.
Target levels are the specific loudness levels that the standards are trying to achieve. For example the target level of EBU R128 is -23 LUFS whereas the target level for ATSC A/85 is -24 LKFS.
True-Peak metering is a more refined and detailed metering approach. A regular peak meter detects the level of audio by looking at the sample in question. A True-Peak meter looks between the samples for inter-sample peaks. Essentially it is peak meter with a magnify glass. The discrepancy between a regular peak meter and True-Peak meter can be quite significant, for example a -0.2 dB peak could go as far as +3 dB peak on a true-peak meter reading.
These frameworks are what keep the level of audio for broadcasting as high as they currently are. If these standards were not in employ the difference in levels between programs and advertisements would be insufferable.
Covers LO 10