Audio normalization – Audio normalization

Audio normalization is the application of a constant amount of amplification of a sound recording to bring the amplitude of a target level (standard). Because the same amount of gain over the entire recording, the signal-to-noise ratio and relative dynamics are unchanged.

Two basic types of audio normalization exist. Peak normalization adjusts the recording based on the highest signal level present in the recording. Loudness normalization adjusts the recording based on perceived loudness.

Normalization differs from dynamics compression, which applies varying levels of gain across a recording to fit the level within a minimum and maximum range. Normalization adjusts the gain with a constant value over the entire recording.

Normalization is one of the functions usually provided by a digital audio workstation.

Peak normalization

One type of normalization is peak normalization, where the gain is changed to bring the highest PCM sample value or analog signal peak to a certain level – usually 0 dBFS the loudest level allowed in a digital system.

Since it only goes to the highest level, only peak normalization does not take into account the apparent loudness of the content. As such, peak normalization is commonly used to change the volume so as to ensure optimal use of the available dynamic range during the mastering phase of a digital recording. In combination with compression / restriction, however, peak normalization becomes a feature that can provide a volume advantage over off-peak normalized material. This feature of digital recording systems, compression and limiting followed by peak normalization, sets contemporary trends in program loudness.

Loudness normalization

Another type of normalization is based on a measurement of loudness, where the gain is changed to bring the average amplitude to a target level. This average can be a simple measurement of average power, such as the RMS value, or it can be a measure of human perceived loudness, such as that offered by ReplayGain, Soundcheck and EBU R128.

For example, YouTube reference level -14 LUFS, so if a program analyzed at -10 LUFS, YouTube will decrease the level 4 dB to the reference of -14 LUFS.

Loudness normalization was made in different volume combat when listening to different music in a series. Before loudness normalization, one song in a playlist would be quieter than the rest, so the end listener would have to put a volume knob to adjust the playback volume.

Depending on the dynamic range of the content and the target level, loudness normalization may result in peaks that exceed the storage medium. Software offering such normalization usually offers the option of using dynamic range compression to avoid clipping when this happens. In this situation, signal-to-noise ratio and relative dynamics changed.

Audio normalization is a process in which the amplitude (volume) of an audio recording is increased or decreased in a constant relationship over time, so that the maximum amplitude or the maximum effective value or the perceived volume (volume) reaches a predetermined level, the standard. If the signal has multiple tracks, they all undergo the same correction.

Example: normalization of peaks to -3 dB:
A collection of digital recordings is made with a peak modulation standard of -3dB FS.
A new stereo recording is measured. The highest maximum level is -5.5 dB FS on the left track, -5.7 dB FS on the right track.
Normalization consists of applying a constant gain of 5.5 – 3 = 2.5 dB.
Standardization requires two passes. The first determines the maximum level, the second applies the correction to the entire recording.

Maximum normalization changes the level, but not the dynamics of the sound.
Volume normalization or perception of loudness often includes compression that changes the dynamics of sound.

Peak normalization

Peak normalization applies a constant gain to a recording to bring the highest peak to a target level, 89% professional audio (-1 dBFS true peak (True Peak)).

The sound dynamics of the recording are more or less preserved, except that maintaining a low distortion level after multiplication of all samples may involve the application of a known quantization error decorrelation noise. under the name redithering (tingling of the least significant bit) 2, which slightly increases the background noise level.

Volume normalization

The purpose of volume normalization is to bring all sound elements in a collection to the same sound volume level, so you can hear them without having to adjust the volume. In fact, the normalization of the maximum level in no way guarantees a homogeneity of the perceived sound volume (Loudness).

A simple approach to volume normalization, which is provided by various software programs, is to normalize the RMS value of the integrated signal within a few tenths of a second. The most advanced machines use extensive algorithms for more accurate evaluation of the perceived noise level. The European Broadcasting Union published a recommendation 1 in 2011, which provides a relatively simple method for this evaluation.

If the standard is not low enough, volume normalization involves compression for recordings whose sound dynamics would be higher than implied when setting the standard from the maximum level. If not, the signal peaks would exceed the quantization limits.

In the simplest implementation, volume normalization collects volume data during the first pass, determines the gain or attenuation necessary for the maximum volume to reach the norm, and applies this correction to the second pass. If the elements of the collection have the same characteristics, from form factor to top factor and dynamics, as is the case with popular music collections or recorded speech, this approach produces satisfactory results.

Extensive implementations use a standard that includes not only the volume of the sound, but also the maximum maximum values ​​and dynamics of the sound. They collect loudness levels and maximum values