In a previous article about sound, we talked about dynamics and dynamic range (the range between softest and loudest sounds) and how important it can be to pay attention to these things.  For this installment, we will look at the actual measurement of sound levels and how to think about them when mixing for music, podcasts or film sound.


One important concept to keep in mind is that an audio signal could often exist in or pass through 4 different mediums:
The acoustic or physical medium where minute and rapid changes in air pressure move in various ways in a room, and made the eardrum of the ear vibrate in sympathy.  This is the medium of the ACOUSTIC.

The physiological/psychological perception of sound, where those vibrations of the eardrum were translated into the mental perception of a sound.  This is the medium of PERCEPTION.

Early audio was concerned with ways to take sound and convert it to an electrical signal and manipulate it in various ways, before converting back into acoustic sound.  The electrical signal waves were seen as analogous to the sound waves in the air, and so audio based on electrical conversions and processing was called the ANALOG medium.

When computers and digital sampling technology was developed, it became possible to encode sound as a series of numbers.  Mathematical processes done by the digital computer could manipulate that series of numbers towards various ends.  Audio signal that is encoded/decoded and manipulated by computers is called the DIGITAL medium.

Often sounds pass through a chain of these mediums.  For example, a word spoken by a voice starts out in the ACOUSTIC.  A microphone and a preamp converts sound waves into electricity, in the ANALOG medium.  An audio interface’s analog-to-digital convertor samples that electrical wave into the DIGITAL medium as a series of numbers, where mathematical processes by the computer alter the sound according to the directions of the sound operator.  It is then converted from a series of numbers back into an electrical signal, into the ANALOG medium.  The speakers in my headphones convert those electrical impulses into ACOUSTIC sound.  The acoustic sound tickles my eardrums and the mechanisms in my ear and the processes in my brain give me the PERCEPTION of sound.

Because a signal can pass through several mediums, one can look at sound from subjective perspectives (your personal experience of sound or even a generalized model of how an ideal concept of the average human ear and brain perceives sound) or objective perspectives (the physical and scientific measurements) - both perspectives are valuable when working with sound.


The decibel is the most common unit of measurement of sound.  It is a logarithmic concept, not linear - an increase of 10 decibels represents a subjective doubling of sound volume.  This logarithmic quality is actually similar to how the human ear perceives changes in volume and allows for an easier scale of volume to work with. As well, it is technically a ratio or comparison of a given sound pressure level to reference pressure, such as the threshold of hearing.  While there are other units of measurement for loudness, such as the sone or the phon, the most commonly used will be the decibel, abbreviated as some variation of dB.


Levels are an objective measurement of the intensity of a sound and can be a measurement of the physical acoustic sound (the Sound Pressure Level) or a measurement of the analogous electrical or digital signal when working with audio electronics or computer processing.  Levels are usually measured in decibels.  Note that there are different scales used in different mediums.

Sound (as we usually think of it, in the acoustic medium) is usually transmitted when there are rapid changes in air pressure…the bigger the change of air pressure, the louder the sound. Sound Pressure Levels are the measurement of physical/acoustic sound wave pressure - are often measured by a Sound Pressure Level Meter (or SPL Meter). In the real world physical medium, 0dB indicates the absence of sound, and measurements are positive integers (such as 20dB or 102dB).

When working in software (ie the digital medium), there is an absolute limit to how high a level can get.  Digital audio has set that limit at 0dB - sounds that attempt to go over that limit won’t get any louder, they’ll just distort in a harsh and unpleasant way and so going above that 0dB limit in the digital medium of your software is to be avoided.  Levels below or leading up to that point are measured as negative integers, such as -10db.  An absence of any signal would be measured as -infinity dB.  Sometimes levels in the digital medium are referred to as dBFS, or “decibels (relative) to full scale”, with “full scale” being that 0db limit, so -20dBFS is 20dB less than the full scale that the software is capable of.

So sounds can be measured by the same unit (the decibel) but using different scales:

Acoustic sound uses measurements in dB SPL, sometimes weighted to approximate the innate sensitivities of the human ear to specific frequency ranges.

Analog audio uses measurements in dBv, dBu and others, which are based on standards involving electricity - these are often only encountered in professional sound studios.

Digital audio uses measurements in dBFS (often just called dB).


Any path of audio signal can be seen as a chain of components, whether it be physical devices or stages of digital signal processing in a computer’s software- at any component in that chain, the level of the sound can be increased or decreased.  Gain generally refers to the level of a signal going IN to a component.  Volume is generally seen as the level of a signal coming OUT of a component.  It is important to set levels correctly at all component stages of the chain, to make sure each component is working with an optimum level of sound - not so loud that the sound gets distorted making sound too loud for the components to handle cleanly, and not so soft that it means having to crank the gain at the input of the next component in the chain, and possibly introducing unnecessary noise into the signal.  This setting of levels along the path is often called gain-staging.  Much of the time, gain-staging is automatic in the digital medium, but not always, and its best to pay attention to your levels at every stage of the game that you can.  For example, audio professionals often calibrate the volume of their speakers so that they put out a specific Sound Pressure Level  (for example, 85 dBA) when the meter in their audio software indicates that they are putting out sound at 0dB or -3dB.


Early attempts to measure sound used physical and electricity-based components to measure sound levels.  Volume Units, often associated with VU meters, are a holdover from these earlier days of audio, when things were measured by electrical components, such as the traditional needle-and-dial volume meters.  These volume meters were relatively slow to respond to sudden peaks and so could provide an averaged measurement of sound.  There were several variations of these meters and standards were developed by various audio and broadcasting agencies in different regions.  Several criteria were developed to try to standardize things more broadly that related to precision and speed of measurement when a sound was sensed, the level of electrical activity and other things.  When digital audio was developed, attempts were made to imitate some of these measurements and criteria; sometimes these attempts were imperfectly adjusted and adapted - indeed some were impossible to carry over and so new methods and standards were created.  There is constant work being done to figure out how to correct for imperfections of adaptation and to best measure sound in the digital medium in such a way as to most closely approximate what happens when audio is converted from a digital signal into acoustic sound.

Most audio software and video software will have a meter somewhere that measures the level of audio coming out of your project.  Usually, it is a constantly-adjusting bar graph that measures peak (instant by instant) or RMS (average) levels or both.  Some also register the maximum peak of a sound as a number and/or a red light to indicate overload (ie when sound volume has tried to go higher than the 0dB limit in digital audio.


Volume levels can be measured in a few different ways, with RMS and Peak being the most common.

RMS or Root Mean Square measurements are a sort of averaging of sound levels over a short period of time.  This is an effective way to get a better sense of the overall volume of a sound but won’t really register sudden short jumps in sound level.

Peak measurement takes a very short measuring of the sound level at any given instant and so can detect sudden short jumps or dips in sound that may need to be addressed.  Note that there are a couple of different types of peaks.  Sample peaks are a concept of digital audio and are more of mathematical concepts or ideals.  True peak relates more to how sound is actually handled by real-world physical components, such as speakers, and so is considered a more useful way of measuring peak signals.  It may be useful to find out if your meters work with sample peak or true peak.  Some meters will record the highest peak level as a number that you can use to reference how strong the level gets for a particular piece of audio

There has been a recent and gradual shift from using audio levels to measure sound, to using tools and scales that measure loudness, particularly in the broadcasting field.


Loudness is the subjective perception of sound pressure.  There is much research still being done in this area as loudness not only involves physical properties such as Sound Pressure Levels, but also the complexities of physiological responses to sound and psychological perceptions of sound.  There are a lot of standards that are being developed to provide more objective measurements of loudness and these standards and even their definitions are constantly being refined.  Different regions and even different broadcasters and cinema distribution companies will have their own standards as to how loud various things should be and even how to measure that loudness.

For example there is a unit of measurement that is gaining popularity called LUFS.  LU stands for loudness units, which are somewhat similar to decibels but are a relative unit of loudness; the “FS” part refer to the full scale of the digital medium.  There is also the LKFS scale of measurement, which stands for “Loudness Units, K-Weighted, relative to Full Scale”.  K-weighting refers a special filtering system to approximate the way the human ear perceives the loudness of various pitches, so that lower, bassier pitches register less on the meter than higher pitched frequencies.  These loudness units are meant to more closely match how the ear perceives loudness.


Note that there are other things that can affect the perception of sound.  One important one is that the human ear’s sensitivity to different pitches and frequencies (low-pitched bass noises to higher-pitched shrill noises) actually varies depending on the intensity of the sound.  For example, at lower volume, the ear is less sensitive to lower, bassier pitches and higher shrill pitches, making the mid-range (for example, where most of the intelligibility of human speech lies) seem more prominent at these lower volumes. Usually one uses one of two weighted scales of sound pressure level measurement (SPL A and SPL C) to help simulate and accommodate for the ear’s different sensitivities at various volumes.  Often, it is helpful to listen to one’s audio at various levels to get an overall sense of things - a mix of different audio elements might sound perfect at louder volumes, but might mask your dialogue at lower volumes.

Another thing to watch for is that the ear will often perceive something at higher volume as sounding “better” than the same mix of sound at a lower volume.  If you compare two sounds, the higher volume one will usually seem “better”, which may have a negative effect on being able to listen and judge the quality of sound objectively.


Most audio and video editing software will include some form of audio metering.  However, they may be limited in scope, especially when measuring in LUFS.  There is, however, a free version of a measuring tool, a plugin meant for use within audio or video software, called the Youlean Loudness Meter II.  It can be found here:


So, is it best to rely on meters or on your own ears with regards to loudness?

Ideally, you should use both.  Your ears will provide a subjective experience of the sound, and ultimately, the sound you put out will be destined for human ears.  However, you headphones, speakers, speaker positions, room acoustics, the position of your head and the state of your own hearing will colour the experience of sound in ways that may not apply to other people who will be listening to things.  Our ears can tire during sessions of critical listening - sometimes professionals will limit themselves to only a few hours of such listening at one time.  And too, you need to make sure that you’ve calibrated your metering to the actual sound pressure levels in your listening space.

Visual meters can provide an objective reference to what is going on with your sound.


You can use the Peak Meter to check to see if there are a lot of jumps in your sound levels.  If so, you may need to use something like a limiter or compressor to tame your peaks/jumps to make your audio more manageable.  We will look at using these components in a later article.  Even if there are only a couple of peaks/jumps, you should make sure that they don’t try to go above 0dB on your meter, as the sound will be distorted at those points.

Use your RMS level Meter to get a sense of how loud a piece of audio is overall, as an ongoing average.  If you know that your piece of music is going to have a lot of sound going on, perhaps at the chorus, you can see how loud things are in the chorus, vs how soft things might be during the intro part of the first verse.  Maybe the range between those soft and loud parts is too big and you need to find ways to bring up the volume a bit more on the soft parts or the volume down a bit for the loud parts.

Different mediums and even different destinations for those mediums will need different sound levels.  For example, if you are intending to send a film project for consideration by a streaming service, such as Netflix in Canada, you will need to look at their standards and then shape and adjust the audio elements of your project to fall within those guidelines.  You may also need to make sure you have the proper tools for measuring your levels appropriately.

Throughout history, there have been various ways to view and measure sound.  Things are still being developed to more closely approximate how we perceive loudness.  It is a complex field with standards still being developed, so research will be necessary to find out the most recent and relevant standards one will need to follow for any given project.