Why you should normalize your sample library
How to balance the loudness of your sample library for better quality mixes
There is no shortage of great samples on the internet! You can easily find an array of premium services selling preprocessed sample packs as well as free sounds made available by producers and members of the music production community. This is perfect if you are looking for some inspiration or a kick-start for a new song, but it does come with some technical challenges when it comes time to mix. Likely, your personal library of samples will contain files with different sample rates, bit depths, file formats, and importantly — different loudness. It can be a pain to blend sounds together from various sources. Luckily, there are some steps you can take to simplify the task. This blog post will explain some of the basics of loudness vs. level and provide a guide to overcome mixing challenges by normalizing your sample library before you start creating.
Level vs. Loudness
Level and loudness are two concepts that are related but not the same when used in a professional audio context. Level is an objective measure of an audio signal whereas loudness is a subjective measure. Mixing engineers must accommodate both level and loudness in their choices so it is an important topic to understand.
- Level
In an acoustic setting, the level is measured in SPL, which means Sound Pressure Level. When you make a sound, molecules vibrate in the air, causing changes in pressure. SPL is what the police will measure outside of your house when your neighbors have called in for a noise complaint. In an electrical setting, the level of an audio signal is measured in voltage. When you change the settings on your preamp, you are amplifying or attenuating the voltage of the signal. If you then convert the electrical signal into a digital signal via an audio interface, the signal is represented as a binary number on a computer. This is the level you are changing when you move that fader in your DAW. No matter which state your audio is in (acoustic, electric, or digital), the level is always something that can be precisely measured.
- Loudness
For something to be loud something must experience it as loud. For this reason, loudness can be approximated but not precisely measured. Younger people tend to hear high frequencies better than older people and dogs can hear much higher frequencies than any human. Thus, there is no true objective “loudness.” This is a challenge for mixing engineers since their jobs are very much centered around how musical elements sound in relation to one another. Still, even with variance in the way people hear, human hearing does follow patterns that can be used to help get at the fuzzy concept that is loudness.
Psychoacoustics
Psychoacoustics is the field of study that deals with the perception of sound, including loudness. Researchers have discovered that humans hear higher range frequencies better than low frequencies. This means low-frequency sounds require more energy for humans to perceive them to be the same loudness as high-frequency sounds. With this information, audio researchers have been able to develop measures of loudness such as RMS. RMS stands for Root-Mean-Square — the mathematical operation performed on the signal to calculate a loudness.
Peak normalization
One way to deal with level without paying anything is to peak-normalize your sample library in your DAW. This procedure finds the loudest moment in your audio file and amplifies the signal to the maximum level that will not cause distortion. In a digital setting, the maximum level is 0 dB. This is a straightforward operation, but caution is required when using files that have been peak normalized. Since files are boosted to the maximum allowable level, any increase in the gain will introduce digital distortion into your signal. Another issue with this approach is that loudness is ignored. Imagine one file with a short snare hit and another file with a long 808 kick. The snare signal will have a sharp spike, but the average loudness of the signal will be low. The kick drum file will be more constant in loudness, so boosting the signal up to what the loudest moment will allow will also boost the average loudness much more than the snare drum. This is not ideal for mixing as you would much prefer to have sounds that are even in loudness as opposed to level.
How can I adjust my samples to have an even loudness?
The traditional way to do this would be to load your files in and manually adjust the faders on your mixing console until things sound “about the same”. You would want to watch where the peak level comes to, ensuring that it never goes above 0 dB. Also, ensure that the level is centered around –18 dB for the duration of the sound. Conventional wisdom says an audio signal should peak around 3 dB and be centered around –12 dB to –18 dB. This approach works well since you make all the decisions about what sounds right. However, it is tedious with hundreds of samples and virtually impossible with thousands or tens of thousands of samples.
Is there an automated way to adjust level and loudness in many samples simultaneously?
Yes (!), and you have come to the right place! Origin Audio offers a tool called Batch Normalizer that can automate this process for you. This way, you can balance all your samples without going through the lengthy task of manually checking each file. Batch Normalizer automates the manual process described in the previous chapter by examining peak level and RMS level at the same time.
Try Batch Normalizer for free when you sign up for an Origin Audio trial and save yourself a whole lot of time on audio “housekeeping” so you can focus on the fun parts of making music.
In conclusion…
…no matter which method you choose, balancing the level and loudness of your sample library will make mixing more intuitive, more enjoyable and help improve the overall quality of your mixes! Try one of the recommended methods and see how it improves your workflow.