Intimate Production Techniques

With the runaway success of Billie Eilish’s debut album (winning 4 Grammys) ‘When we all fall asleep, where do we go?’ it’s clear that ‘intimate’ or ‘mumbling’ vocals are in vogue. In fact, it would be accurate to state that ALL the sounds within a mix are produced with ‘presence’ or ‘intimacy’ in mind and not just the vocals. It might not be to your tastes but this type of production is fast becoming very popular amongst the ear-bud listening generation. If intimate productions are rocking your boat then the following techniques and processes should help you in achieving an up-close and personal mix.

The approach to achieving a close-up and personal sound rests in the use of existing old school technologies coupled with innovative new processes and a daring mindset. As technology moves forward so do production techniques – they are inexorably linked. Sadly, in today’s world of ‘let the software do the work’ ethos producers seem to be reluctant to use tried and tested techniques that engineers have developed for them. There is an unhealthy move towards using this type of ‘analysis and application’ software and we are seeing more and more software developers accommodating this requirement. From track analysis software to ‘one knob magic processes’ the role of the producer has changed to the point of being an assistant to the software rather than the other way around. The same can be said for the mastering market. It seems preset driven software is all you need to master a song nowadays. However, out of all this negativity, some good has surfaced – most notably in the design of multi-function plugins. We are seeing more and more plugins offering all manner of extended functionality for a required process; a good example would be FabFilter’s Pro MB (multiband compressor) which not only provides traditional downward compression but also offers upward compression and downward and upward expansion. FabFilter’s approach to designing plugins that offer all the required functionality inherent within a process is one of the main reasons that professionals use their plugins. Izotope have taken the multi-function ideology a step further and their products now come supplied with all manner of analysis and compensatory processing. Sound Theory’s Gullfoss is another product that offers adaptive processing in that you dial in a set of values for the basic parameters on offer and sit back and let the plugin do the work for you. Mastering the Mix is another company that has grabbed the analysis and application market by the neck and their products are not only useful and helpful guides but most can make intelligent suggestions as to how to alter values to achieve optimum results. In fact, many software developers are going down this route and it would be fair to suggest that almost all mixing and mastering tasks can be achieved by the use of these analysis and application software.

1 UP 1 DOWN

All of this whinging has led me to the first of the processes used in achieving intimate productions – that of combined downward and upward compression also referred to as two-way compression. But before I unleash all that is two-way compression let me gently foray, with you, into the world of compression. Every producer knows that compression is about manipulating the dynamic range of audio and NOT the volume. A long, long time ago in a workshop far, far away an engineer came up with this insane invention; he called it ‘the volume knob’. This crazy creation had the amazing ability to raise or lower the volume of audio. Mad huh? At no point did he, or his mates, pose the question ‘why don’t we call this volume knob thingy a compressor’? All jokes aside, audio compression is actually a very simple process and the subject has been eloquently covered by countless articles at this manor. With regards to intimate production techniques, the important distinction to outline is the difference between upward and downward compression and how these two processes affect the dynamic range of audio, which are critical in achieving ‘presence’, or ‘closeness’, in a mix context.  For this article, and to keep the processes in context, the dynamic range of audio is determined to be the difference between the loudest and quietest parts of the audio signal, and is expressed in decibels (dB). A downward compressor makes the loud bits quieter above the threshold whilst keeping the quiet bits below the threshold unaffected. An upward compressor makes the quiet bits louder below the threshold (by compressing and applying gain makeup) whilst keeping the loud bits above the threshold unaffected. By using downward and upward compression simultaneously, the audio signal’s dynamic range can be reduced from both above and below the threshold which results in a signal that does not have severely compromised peak transients had the same level of dynamic range reduction been attempted with just the downward compressor.

We now know that compression simply alters the dynamic range of the audio being processed but how does that help us with regards to achieving intimate productions? The answer is quite simple but it does require you to use your imagination in envisaging the shape of the audio being processed by using two-way (combined) compression. Imagine the audio is represented as a nice and squishy hotdog shape, no bread of course. Now imagine grabbing the hotdog between your fingers and squeezing its width. This is what happens when you apply downward and upward compression at the same time: you are, in effect, squeezing the audio both above and below the threshold closer to the threshold. If the quiet bits are louder and the loud bits are quieter then the audio is much easier to gain stage to a specified level. Let’s look at this in context; you have a busy mix and you are trying to get the vocals to sit nicely in the mix. Now, each time you drop the level of the loudest part of the audio the quiet bits get masked by other sounds. If you raise the quiet bits to be heard then the loud bits are too loud. Of course, you can use downward compression to narrow the dynamic range and afford a more balanced difference between loud and quiet bits but this comes at a cost; peak transients are compromised, and a better approach might be to use volume automation to control these gain variances, and this is the more traditional approach in managing variable gains within a mix but with intimate productions, we are trying to achieve an even narrower dynamic range without having to use aggressive downward compression. The idea is to use a range that acts as both the upper and lower gain limits for the audio to ‘travel’ in so that once left at the target gain value it can be heard clearly and in detail. Once two-way compression is used and the target value set no automation is required as the overall range the audio travels in is so small that the loud and quiet parts are almost at the same level and can, therefore, be left at one specified target gain value.

You can apply two-way compression as two distinct processes or as a single process. A novel approach in achieving two-way compression as two separate processes is to use the auxiliaries of your DAW to house each process. By disabling the routing for the channel the audio resides in and having only the auxiliaries active you can blend various amounts of each compressor mode to taste. You can also apply further processing to each auxiliary thus allowing for even more processing choices. I like to keep things tidy and use combined processing for two-way compression. The Waves MV2 (fig 1) is a fun tool for achieving both downward and upward compression simultaneously but it does not allow for selecting different threshold values, attack and release times or even offering control over the ratio. Hornet’s Dynamics Control plugin (fig 2) is far more versatile and offers an upward and downward compressor and a compressor/expander which acts to squeeze the sound around the threshold. You can alter threshold and ratio values for each compression mode and there is a global attack and release function to shape all the compressors. However, currently, the plugin has no make-up gain feature which I find a little surprising. This omission forces the user to use independent processing to control the output make-up gain. I use two-way compression on all sounds that exhibit a wide dynamic range. By narrowing the range the quieter elements of a sound are exposed and this translates across as ‘presence’.

waves mv2 downard upward compressor

 

ADAPTIVE PROCESSING

Upward and downward compression are great processes for altering the dynamic range of audio and thanks to the narrowing of the dynamic range quiet bits can be heard alongside the loud bits which means we can work with very narrow gain margins and still hear everything in detail. But this isn’t the only process we use to create intimate textures within a mix context. One very powerful technique that works but in a markedly different way to two-way compression is that of adaptive and compensatory processing and the best example I can think of to express this process is Sound Theory’s Gullfoss. Gullfoss is an intelligent auto eq plugin. It has only five parameters on offer and comes with a single page manual, but don’t let this simple and minimalistic approach fool you, the software is remarkably powerful and yet easy to use. The real power of this software is what goes on behind the scenes and the whole topology and product goal is based on a model of human auditory perception. The user dials in a number of settings for the 5 parameters on offer and the software creates an eq response that adapts itself continuously and dynamically as the audio is played. It is stated that the eq will counter the side-effects of phase problems and most notably with temporal smearing that takes place with the use of minimum phase designed equalisers and multi-mic scenarios. The eq is also constructed to avoid overshoots and ringing. However, with regards to intimate productions, it is what the proprietary perception modelling achieves when trying to expose masked frequencies and taming pronounced frequencies. The two main parameters are Recover and Tame. Recover exposes and processes frequencies that are masked and Tame suppresses and processes frequencies that are dominant and cause the masking problems in the first place. The final three parameters are control parameters for Recover and Tame and to balance the overall eq response. The beauty of Gullfoss is that once you have dialled in the necessary settings the software takes over and applies the desired changes in real-time and continually adapts to achieve the goals set by the parameters. With intimate production techniques, we are constantly trying to unmask and emphasise frequencies and bring them up in the mix closer to the more dominant frequencies that may, at times, be the cause of the masking. In effect, we are trying to make the quieter hidden frequencies louder and louder dominant frequencies quieter. Although this is not quite the same as narrowing the dynamic range of the audio the overall process does result in a similar type of ‘quiet meets loud, loud meets quiet’ scenario which, after all, is exactly what our goal is in trying to achieve an up close and personal sound.

NO SPACE IS GOOD SPACE

Reverb is without a doubt the most important effect used by producers. It is predominantly used to define the overall space that all the sounds within a mix will reside in. Well recorded sounds do not need any additional processing to define their location within a given space as the recordings themselves will expound the characteristics of the space the sounds were recorded in. However, even with well recorded stems, producers will still use reverb to glue all the sounds into a pre-defined space. In addition, reverb is used to express depth. By using the various reverb parameters, most notably diffusion and filtering, we can ‘place’ sounds within a space. By ‘placing’ I am referring to sounds that sit up-front or ‘at the back’ of a mix and not left and right pan locations. Intimate production techniques generally omit the use of reverb almost completely, with the exception of providing a reference (see; Perception), and instead use equalisation and filtering to achieve both depth and front-back placement. This type of approach requires a different mindset to the norm and the real skill of the producer is in achieving space and depth without the use of reverb. If you consider how sound travels then you will note that high frequencies dissipate much faster than low frequencies. We mimic this behaviour by using equalisation to low pass sounds that need to sit further back in a mix and high pass sounds to bring them up-front in a mix. Volume plays a huge role in the depiction of distance and by coupling this with equalisation/filtering we can create both a sense of space and how sounds sit within that given space.

Although mix reverbs are a no-no it doesn’t mean that other effects cannot be used. The aim is to not use effects that denote a given space but, rather, to use them to emphasise and colour a sound, and that includes reverb. I often use delay effects instead of reverbs to extend a sound’s sustain or to add the perception of depth. Delay effects have the advantage of not smearing frequencies like reverbs do as they only concentrate on a single ‘reflection’ as opposed to a multitude of early/late reflections afforded by reverb effects. Distortion is another process I use quite regularly to treat vocals instead of opting for an equaliser and with dry unaffected mixes it adds an extra layer of sparkle and definition. Harmonic exciters are potent alternatives to simple equalisation processes and they can often add sparkle and presence to a sound simply through the auditioning of generated harmonics. Minimum phase designed preamps are often used instead of minimum phase designed equalisers as they do not perform any corrective processing but instead colour a sound in a pleasing way that lends itself to minimalistic and intimate productions. All forms of harmonic distortion are highly useful when it comes to intimate productions as they often provide depth and vibrancy to staid sounds and work extremely well with minimalistic productions as they can be heard in their entirety thanks to the low track/stem counts.

If we use Billy Eilish’s production techniques as a point of discussion then you will note that reverb has been used quite extensively throughout the album. However, it is context that matters and reverb for colouring sounds is a ‘yay’ whereas master mix reverbs are a ‘nay’.

I often use reverbs to add presence to certain sounds but the way I use them is to remove almost all late reflections and work on early reflections, an imperceptible reverb decay value, a very short pre-delay, and tons of diffusion. In effect, I am removing all the parameters that would help in denoting a given space. Instead of the ‘reverb wash’ effect, I am aiming for ‘presence’.

PERCEPTION

A trick that many EDM producers use, and to be honest it is an old school technique, is to utilise an ambient looped or sustaining sound to act as the backdrop or anchor for the mix. All the other sounds in the mix will then be referenced against this ambient backdrop. The psychoacoustic effect of having a sound referenced against an ambient sound is that the sound is then perceived to be more pronounced and clearer. In EDM music you will often hear pink noise sweeps appear at breaks, bridges and prior to the drum beat and bass-line progressing into full flow. The effect is quite noticeable in that the drum beat and bass-line sound stronger, deeper, and clearer. But more importantly, they appear to be sitting in their own given spaces even though no effect has been used to denote the space for the mix. It is the reference that tricks the mind into thinking there is depth and width to the mix.  With intimate productions, the vocals are usually used for all sounds to be referenced against. If you listen to Billy Eilish’s Bury a Friend you will notice the vocals have been ever so slightly affected. Once the mix sounds are referenced against the vocals the perception of ‘space’ takes over the whole mix.

Complimentary processing also works in tricking the brain that one sound is more pronounced than another. Whereas with ambient referencing we are referencing sounds against a single texture/colour, with complimentary processing we are actually highlighting one sound against a like for like sound. A good example of this type of brain trickery is when we come to processing low-frequency sounds that share similar frequency ranges, and the best example I can think of that has cost many a sanity is marrying the bass and the kick within a mix. The traditional approach is to use ducking to control the interchange between the bass and the kick drum but this is not a perfect solution as one sound needs to be, in effect, attenuated to create the space for the other sound to explore and vice versa. With minimalistic mixes the ducking process is obvious and is easily heard and that is not what producers aspire to achieve. The secret to a good production is that processes must go unnoticed and only the results to be heard, and in the case of ducking hearing one sound above another is not the aim. The aim is for BOTH sounds to be heard at the same time and for that to be successfully achieved we need a different set of processes. A combination of compression and expansion yields a far stronger perceptive result than a simple ducking process. The technique involves downward compressing the fundamental of one sound and expanding its harmonics while expanding the fundamental of the other sound and downward compressing its harmonics. This sea-saw process fools the brain into thinking that both sounds are playing at the same time and is a highly effective solution for managing sounds that have shared frequencies and predominantly at the fundamental.

THE AIR BAND TRICK

In audio production, we producers are always looking for ways to trick the brain into thinking that something is there when it isn’t. Using reverb to denote space is one such example. But because we cannot use reverb with intimate productions we head straight for the air band and with the use of some clever trickery, we can fool the brain into thinking that the mix is airy, has acres of presence, and is sparkly. To work the air band, which invariably lies between 10-20 kHz, we need to process frequencies that actually exist within that range. Some producers add huge gain boosts to this range, even if there are no sounds actually present within the range, thinking this will magically create ‘air and space’. In fact, all this achieves is to give the high end a brittle and harsh texture. Once the air band range has been established gentle boosts or applying some form of harmonic excitation can really make this range stand out and sparkle. However, there is a very cool trick that I have been using with intimate productions and it works quite well in providing a sense of space and airiness without the use of reverbs.

The process uses a plugin that has a feature that I wish more manufacturers would adopt. The weapon I am referring to is FabFilter’s Pro Q3 and in particular the ‘split band’ feature. This allows the user to place an eq node anywhere on the audio’s frequency spectrum and to split the node into two further nodes; left and right. These nodes can now be moved to create a wider stereo image for the given frequency range. The trick is to find the perfect air band location. Simply sticking a node around the 10-12 kHz range does not magically create ‘presence’ and ‘air’. You need to locate the frequency range that contains high-frequency information to be able to truthfully use this feature. Let me show you how easy it is to achieve this brain foolery with the following example:

I am using a segment of an instrumental track (below) kindly sent to me by one of my students. The track is called Icarus and I have selected a section that I feel will benefit greatly from this air band trick.

(audio) icarus segment.mp3

I have imported the track into Cubase and inserted the wonderful FabFilter Pro Q3 equaliser plugin on the Icarus audio channel. The GUI’s spectrum analyser displays the frequency response of the audio and it is actually quite easy to locate the air band range just by viewing which frequencies are active at the upper end of the spectrum. Select a midway point between the start and end of the air band and place a node there. By using the band solo feature we can audition only the frequencies in this range and that makes it considerably easier to set the air band range than auditioning the whole spectrum.  I have left the Q factor (band-width) at its default value of 1.00 which affords a wide enough range to capture the air band. If you want to fine-tune the range all you need to do is to change the Q factor. Once you are happy with the air band range split the node into two further nodes; Left and Right. You can now separate and position the Left and Right nodes to taste as I have done below.

air band trick using fabfilter pro q3

(audio) icarus segment air band trick.mp3

You can clearly hear the difference this little trick has made to the overall texture of the mix. The audio sounds as if it has both presence and space even though we have not used reverb to define the space. In some ways, and this is only relevant when it comes to the air band, it sounds as if we have created a faux reverb – fake reverb.

LESS IS MORE

Intimate productions techniques only work if the mix is minimalistic. It is hard defining clarity, space, depth, and presence when many sounds are playing simultaneously. Whereas this is exactly what is required for producing pop music it is the exact appositive for this type of production. With pop music productions, it is not uncommon to have 100+ stem counts as stems/sounds are used for layering as opposed to having each sound sitting in its own space. The interest of the listener is maintained by the multi-layering process. In effect, each sound appears to have deep layers that the brain tries to evaluate and reference and this keeps the brain active and interested which is ultimately the aim of every producer – to keep the listener interested. Hip Hop production is the closest comparable for the type of processes covered in this article as Hip Hop leans towards minimalistic productions both in terms of the number of stems used and in how each stem is presented. The idea is to have a sparse bed for the rap and backing vocals to explore and reside in. Instrument sounds are kept to a minimum and only used to provide a contrast for the driving rap lines and sweet backing vocals. Intimate productions follow the same ethos but take it to a whole new level. Because reverb is regarded as hell-spawn, instrument sounds have to be processed with even more attention to detail. Call it ‘precision processing’. If each sound is not crafted to perfection then the errors jump at you with no consideration for your karma or street cred. It’s amazing how constricted a producer feels when reverb is removed from the equation. But I think it’s a good thing. It forces the producer to think outside the box and experiment with processes that they might not have used had reverb been the go-to effect for all things spatial. It is through these restrictions that new clever and innovative processes are discovered or created. When listening to a busy mix it is hard to isolate specific sounds and evaluate the frequency spectrum and dynamic behaviour each sound boasts. Sounds will overlap and either sum or mask each other. The dynamic motion of individual sounds are difficult to ascertain in a busy mix but with a sparse and minimalistic mix each and every sound is heard in its entirety and this is why it is imperative that each sound is optimised to as near to perfect as possible because errors are easily exaggerated in minimalistic mixes. In busy mixes errors can often go unnoticed but not so with intimate productions.

UP, DOWN AND AWAY

Whereas compression is used to narrow the dynamic range of audio, expansion is used to extend or widen the dynamic range. This may seem to contradict the advice from the 1 Up 1 Down section but it is actually an additional process that is performed on the whole mix as opposed to individual sounds. With mastering, engineers tend to use compression and limiting to homogenise a mix to display a good balance of volume against dynamics. Loudness is not the aim even though it seems to still be a pre-requisite for certain genres. Expansion, for some reason, tends to go unnoticed with the less professional mastering engineers, and yet it is such a powerful process. What is even more perplexing is that upward expansion seems to be used far more often than downward expansion. It seems that making the loud bits louder above a specified threshold is more appealing than making the quiet bits quieter below the threshold. I think this harks back to the loudness issue. We tend to think loudness affords more detail whereas in reality separation is a bigger influence. Although the result is the same when it comes to extending the overall dynamic range of the mix the perception of how the dynamics behave are very different. Which expansion mode you use is dependent on how you want the listener to perceive the ‘up close and personal’ attributes of the mix. If the mix is biased towards a narrow banded response with all the sounds coming across as ‘the same level’ then I might opt for upward expansion. By structuring the threshold such that only the peak transients are highlighted, upward expansion raises the peak transients whilst leaving everything below the threshold unaffected. The overall effect of this process is that the attack portions of the sounds now sit above the bodies which come across as ‘pronounced’. Generally, and I use this word tentatively, most peak transients tend to sit with the attack component of a sound as they define how velocity is applied to the sound. With vocals, for example, it is easy to understand this behaviour as plosives and sibilance tend to carry a lot of peak transients whereas the body of a word that follows the prominent attack will generally be quieter with less pronounced peak transients. However, I might not want to pronounce peak transients and instead aim to quieten the quieter elements of the mix to achieve an even more intimate and controlled response. Staying with the same threshold setting and using downward expansion the quiet parts are made even quieter whilst all the louder peak transients are left untouched and this translates across as more intimate and less distinct. Which mode is selected is really dependent on what you are trying to achieve and because both modes extend the dynamic range it really comes down to how you want the mix to be perceived. My advice is to use both modes, render the mix and listen to the overall responses, and gauge what each mode is doing to the audio using a spectrum analyser.

Let us look at an example that incorporates both compression and expansion but using middle and side. I often use this process when mastering mixes and it really does add another level of detail to a mix that conventional downward compression, static equalisation, and limiting don’t. Using the FabFilter Pro Q3 dynamic equaliser we will add presence and motion to a mix. I will stay with the Icarus audio file for now as we are familiar with it.

(audio) icarus segment.mp3

The aim is to solidify the low end of the mix (mid) using downward compression and add presence and motion to a wide range of upper frequencies (side) using upward expansion.

The process of locating where to place each of the two nodes is agonisingly simple. Use the spectrum analyser to display where the prominent low and high-frequency ranges lie. Create a node somewhere in the low-frequency range and duplicate the process for the high frequency. Solo the low-frequency band and listen to make sure you are capturing only low-end frequencies and not the mid-range frequencies. As a starting point, I generally place a node at 100 Hz and move it around, whilst auditioning in solo mode, until I have caught the most prominent low frequency and from there I create the range using the Q factor (band-width). Once you are happy with the node placement and frequency range click on the node and select Stereo Placement/Mid. Perform the same steps for the high-frequency range and this time select Placement/Side (below).

up and down m/s using fabfilter pro q3

(audio) icarus segment up and down.mp3

Now that we have successfully created the mid and side components we need to define each range’s behaviour. I have opted for downward compression for the mid element and upward expansion for the side element. Adding sparkle and presence to the side element helps to lift the mix and add both space and clarity. The fact that the mid element is compressed at the same time as the side element is expanded further pronounces the sides whilst keeping the low-frequency mid under control. This is a simple yet effective mastering process but please don’t feel you have to use the exact same behaviours as I have. Experiment with different behavioural modes, thresholds, split bands and so on until you achieve the texture you are after.

LOW END

The low end of a mix refers to the low-frequency content present within the mix and how these frequencies are processed. Many believe this is exclusive to the bass and kick but it actually refers to all frequencies that reside in what the producer determines to be the low end of the mix. Invariably, the low end of a mix sits in the 0-800 Hz frequency range, but this is not gospel as the low-frequency range of one mix might be very different to the another, and as vocals, synths, pad sounds etc share this range it is critical that this area is processed correctly.

It might seem strange picking low end as a subject for intimate production techniques but it is this area that reigns supreme for this type of production. With certain genres like Hip Hop, the two critical areas of processing are the low end and the vocals. It is the same with intimate productions. The low end acts as the bed for the mix and all other sounds sit in or hover around and above this bed. In effect, the low end anchors the mix and all other sounds are given free rein to explore the remaining frequencies.

The process I use for managing the low end for intimate productions is the exact opposite of what I do for individual sounds – upward expansion! The aim is not to reduce the dynamic range of the bass and kick sounds but to extend their dynamic range and specifically above the threshold. By setting the threshold to just below the peak transients and using upward expansion we can extend the peak transients further up the dynamic range ladder. The difference from quiet to loud is now extended and that helps to make certain low frequency sounds peak above the mix and then drop to below the average mix level to allow other sounds to explore the vacated, or attenuated, space. This is actually quite an important distinction to make; low-frequency sounds are highly intrusive in that they can completely dominate a mix if left at a constant level nearing the mix level, whereas mid to high-frequency sounds are not as intrusive and benefit from having a constant level as shown in the 1 Up 1 Down section. This is only relevant in the context of intimate productions. The idea is to have all sounds occupy a certain level range for the mix and the low frequency sounds to dip above and below this set range.

The best way to demonstrate this is to use the problematic Roland TR808 bass drum and a dynamic piano line and mix the two using upward expansion on the 808 to raise and drop it above and below the average mix level which in this case is set by the piano line.

(audio) piano riff 90 bpm.mp3

(audio) 808 kick 90 bpm.mp3

(audio) piano and 808 90 bpm.mp3

You can hear that the 808 is struggling to be heard above the piano line. Let us now insert the FabFilter Pro Q 3 on the master bus and use expansion to pronounce the 808 line. I have exaggerated the process so you can hear in detail how the 808 drops and peaks above the average mix level.

using M/S with fabfilter q3

(audio) piano and 808 expanded 90 bpm.mp3

The 808 drops nicely below the average mix level allowing the piano to be heard in full and when the expansion process kicks in the 808 rises above the average mix level and can be heard in its entirety.

This is a very cool approach in achieving separation whilst keeping detail and intimacy in check.

Tracking

Let me end this article by touching on what I feel is the most important aspect in achieving intimate productions; that of tracking. If we use Billy Eilish as an example you will note that all her recordings are sung into the microphone both closely and at a constant level. The overall texture is attributed to the room she sings in, and it might surprise you to know it is a bedroom with no acoustic treatment. The environment is critical when it comes to vocal tracking and we spend more time and money trying to achieve a natural sound that we sometimes forget to use the room’s qualities to our advantage as Eilish and Fiennes have done with their debut album. Microphone choice is as important as mic technique and the recording environment. I have achieved excellent results using both small and large-diaphragm condensers but results can vary depending on a number of factors; the recording environment, the preamp and mic technique. The one thing I have noticed, be it deliberate or by accident, is that Billy Eilish uses the proximity effect to her advantage whereas most producers try to remove it. Whether this is actually the proximity effect caused due to her using a cardioid mic at close quarters or the room’s acoustics is hard to tell but it is a wonderful way to add to the presence factor of her vocal deliveries. On occasion, I will deliberately record with the proximity effect but the trick is to control it, by varying both distance to mic and angle of mic, so as to achieve the right level of low-end presence as opposed to booming low-end mush. At other times I might opt for a large-diaphragm condenser as most microphones of this topology will invariably have what we refer to as a ‘vocal lift’; call it the microphone’s response or colour if you will, and it is the response of the microphone that is as important as the preamp and the recording environment, so choose your mics carefully for the given task, whatever it is.

I could dither (you see what I did there huh?) all day about intimate productions but what I have tried to achieve with this article is to give you an insight into some of the techniques we producers use to achieve intimate mixes. But as is the case with all things audio, experiment and find your own bespoke techniques and when you have share them with the world.

If you found this article to be of use, then these might also interest you:

Mixing Pop Music

Mixing Hip Hop

MixBus Strategies

When we talk about distortion the image, invariably, conjured up is that of a guitarist thrashing his guitar with acres of overdrive. However, I am more interested in covering harmonic and non-harmonic distortion in subtle ways using non-linear systems rather than using a specific overdriven effect like guitar distortion or a fuzz box etc.

In an analog system overdriving is achieved by adding a lot of gain to a part of the circuit path. This form of distortion is more commonly related to overdriving a non-linear device. But it doesn’t end there as any form of alteration made to audio being fed into a non-linear device is regarded as distortion even though the term is quite a loose one and not too helpful. The idea is to create harmonic distortion and this is the area I want to explore in this chapter.

Harmonic distortion means that additional harmonics are added to the original harmonics of the audio being fed. As all sound carries harmonic content, and this is what defines its timbre, then it makes sense that any additional harmonics will alter the sound quite dramatically. Harmonic distortion is musically related to the original signal being treated and the sum of the added and original harmonics make up the resultant harmonics. The level and relative amounts of the added harmonics give the sound its character and for this we need to look at the two main types of harmonic distortion: odd and even order harmonics. The exception to this is digital distortion which sounds unpleasant and the reason for this is that the digital distortion is not harmonically related to the original signal.

Harmonics are simply multiples of the fundamental frequency of a sound and the addition of harmonics within a sound defines the sound’ timbre and character. Even order harmonics are even multiples of the source frequency (2, 4, 6, 8 etc) and odd-order harmonics (3, 5, 7, 9 etc) are multiples of the source frequency (fundamental).
Even order harmonics (2, 4, 6 etc) tend to sound more musical and therefore more natural and pleasing to the ear and higher levels of this can be used as the ear still recognises the ‘musical’ content. Odd order harmonics tend to sound a little grittier, deeper and richer and higher levels of this cannot be used as abundantly as even-order harmonics as the ear recognises the non-harmonic content and it results in an unpleasant effect. But there are uses for both and depending on how the harmonics are treated some wonderful results can be achieved.

Extract taken from the Creative Effects eBook

If you prefer the visual approach then try this video tutorial:

Harmonic Distortion – Odd and Even Harmonics

Whenever I have been called into a studio to assist a producer in managing frequencies for pre-mastering I have always been surprised at the fact that people seem to want to attribute a frequency range for the low end of a track. Every track has its own qualities and criteria that need addressing based on the entire frequency content of the track before a range can be attributed to the low end.

I have come across producers affording insights into some interesting low-end frequency ranges and these ranges are relevant only to the context that the track resides in. If we are talking about a heavy Hip Hop track that uses 808 kicks supplemented with sine waves then the low end of that track will vary dramatically to that of a mainstream EDM (electronic dance music) that will incorporate stronger kicks supplemented with ducked bass tones.

So, working on the premise of a frequency range will not help you at all. What is far more important is to understand both the frequencies required for the low end of a specific track and the interaction of these frequencies within themselves and the other elements/frequencies that share this particular range. This might sound strange: ‘within themselves’ but this is the exact area of the physics of mixing and managing low end that we need to explore. When we come to the chapters that pertain to both the harmonic content of a specific frequency range and the manipulation of those frequencies using advanced techniques then all will become clearer.

To fully understand how to manage low-end frequencies we need to look at frequencies, some of the problems encountered with manipulating frequencies, and some of the terminology related to it, in far more detail.

Timbre

We use the term Timbre to describe the tonal characteristics of a sound. It is simply a phrase to distinguish the differences between different sounds and is not reliant on pitch or volume. In other words, two different sounds at the same frequency and amplitude do not signify that they are the same. It is the timbre that distinguishes the tonal differences between the two sounds. This is how we are able to distinguish a violin from a guitar.

Sinusoids

However, to help you in understanding what this has to do with the low end it’s best to explain the first thing about sound, any sound, and that it is made up of sine waves at different frequencies and amplitudes. If you understand this basic concept then you will understand why some sounds are tonal and others are atonal, why a sampled kick drum might exhibit ‘noise’ as opposed to a discernible pitch and why a pure sine wave has no harmonic content.

To explain the diagrams below: I have drawn a simple sine wave that starts at 0, rises to +1 which we call the positive, drops to 0 and then drops below 0 to -1 which we call the negative. From 0 to +1 to 0 then to -1 and finally back to 0 is considered one complete cycle.

The phase values are expressed in degrees and lie on the x-axis. A cycle, sometimes referred to as a period, of a sine wave is a total motion across all the phase values.

This cycle is measured in Hertz (Hz) over 1 second and represents frequency. A good example of this is the note A4, which you have come across so many times. A4 is 440 Hz: this means that the waveform cycles 440 times per second (repeats itself) and this frequency represents pitch. If I jump to A5, which is one octave higher, I double the frequency 880 Hz. If I halve the A4 I get A3 (220 Hz) which is one octave lower.

Partial and total phase cancellations are critical to understand as I will be showing you how to use some very specific techniques to create new sonic textures using these concepts. Understanding that a sound has a timbre and that timbre can be expressed by partials which form, apart from the fundamental, both overtones and undertones is equally important as we will cover techniques in managing low frequencies without having to use the fundamental frequency of the sound. Additionally, when we come to managing shared frequencies (bass and drums) then the concept of harmonics is very useful as we are continually fighting the battle of clashing frequencies, frequency smearing, gain summing and so on. For example, sine waves have no harmonic content and therefore some dynamic processes yield no useful results and more specialised techniques are required. Whereas saw waveforms are rich in harmonics and therefore we are able to use pretty standard techniques to accent the sweet spots and eradicate artifacts.

I will now copy the same sine wave and phase offset (phase shift and phase angle) so you can see the phase values:

The shift value is set at 90 which denotes a phase shift of 90 degrees. In essence, the two waveforms are now 90 degrees out of phase.

The next step is to phase shift by 180 deg and this will result in total phase cancellation. The two waveforms together, when played and summed, will produce silence as each peak cancels out each trough.

Summing

When two shared (the same) frequencies (from different layers) of the same gain value are layered you invariably get a gain boost at that particular frequency. This form of summing can be good if intended or it can imbalance a layer and make certain frequencies stand out that were not intended to be prominent. A good way around this problem is to leave ample headroom in each waveform file so that when two or more files are summed they do not exceed the ceiling and clip.

If you take two sine waves of the same frequency and amplitude and sum them one on top of the other you will get a resultant gain value of 6dB.

Summing is important when dealing with the low end as any form of layering will have to take into account summed values.

Masking

When two shared frequencies are layered and one has a higher gain value than the other then it can ‘hide’ or ‘mask’ the lower gain value frequency. How many times have you used a sound that on its own sounds excellent, but gets swallowed up when placed alongside another sound? This happens because the two sounds have very similar frequencies and one is at a higher gain; hence one ‘masks’, or hides, the other sound. This results in the masked sound sounding dull, or just simply unheard. As we are dealing with low end this problem is actually very common because we are layering, in one form or another, similar frequencies.

Partials

The individual sinusoids that collectively form an instrument’s Timbre are called Partials also referred to as Components. Partials contain Frequencies and Amplitudes and, more critically, Time (please refer to my book on the subject of EQ – EQ Uncovered). How we perceive the relationships between all three determines the Timbre of a sound.

Fundamental

The Fundamental is determined by the lowest pitched partial. This can be the root note of a sound or what our ears perceive as the ‘primary pitch’ of a sound (the pitch you hear when a note is struck).

Overtones/Undertones

Using the fundamental as our root note, partials pitched above the fundamental are called overtones and partials pitched beneath the fundamental are called undertones, also referred to as Sub Harmonics. These partials are referred to, collectively, as Harmonics. This can be easily represented with a simple formula using positive integers:

f, 2f, 3f, 4f etc..

f denotes the fundamental and is the first harmonic. 2f is the second harmonic and so on.
If we take A4 = 440 Hz then f = 440 Hz (first harmonic and fundamental).
The second harmonic (overtone) would be 2 x 440 Hz (2f) = 880 Hz.

Sub Harmonics are represented by the formula: 1/n x f where n is a positive integer. Using the 440 Hz frequency as our example we can deduce the 2nd subharmonic (undertone) to be ½ x 440 Hz = 220 Hz and so on.

An area that can be very confusing is that of harmonics being overtones. They are not. Even-numbered harmonics are odd-numbered overtones and vice versa. The easiest way of looking at this, or rather, counting is to think of it as follows:

Let’s take the A4 440 Hz example:
If A4 is the fundamental tone then it is also regarded as the 1st Harmonic.
The 1st Overtone would then be the 2nd Harmonic.
The 2nd Overtone would be the 3rd Harmonic and so on…

Inharmonic/Inharmonicity

Most musical sounds consist of a series of closely related harmonics that are simple multiples of each other, but some (such as bells and drums for instance) do contain partials at more unusual frequencies, as well as some partials that may initially seem to bear no relation to the fundamental tone, but we can go into more detail about these later on.

It is important to understand this concept as the area of tuning drum sounds and marrying and complimenting the frequencies with tonal basses, is an area that troubles most producers.

When managing low-end frequencies the phase relationships and harmonic content are more important than any other concept because of the limited frequency range we have to process, the nature of the sounds we are dealing with and the types of processing we need to apply.

I have often found frequency charts excellent for ‘normal’ acoustic instruments but a little hit and miss when it comes to synthetic sounds as these sounds will invariably contain a combination of waveforms and associated attributes that will vary dramatically from the standard pre-defined acoustical frequencies. However, ranges of this type can help as a starting point and some of the following might be helpful to you:

Sub Bass

This is the one frequency range that causes most of the problems when mixing low-end elements and for a number of reasons:

We tend to attribute a range starting from (about) 12 Hz to 60 Hz for this vital area. Although our hearing range has a ballpark figure 20 Hz – 20 kHz we can ‘feel’ energies well below 20 Hz. In fact, you can ‘hear’ the same energies by running a sine wave at high amplitude, but I don’t recommend that at all. In fact, we use these sub frequencies at high amplitudes to test audio systems. It is often said that cutting low end frequencies will brighten your mix. Yes, this is true. It is said that too much low-end energy will muffle and muddy up a track. Yes, this is also true. In fact, I cut out any redundant frequencies before I even start to mix a track. However, this is not the only reason we cut certain frequencies below the frequency we are trying to isolate and enhance and it has to do with the impact the lower end of this range has on using processors like compressors (more on this in later chapters).

Bass

I have seen some wild figures for this range as bass can encompass a huge range of frequencies depending on whether it is acoustic or synthetic. But the ‘going rate’ seems to be anywhere between 60 Hz all the way to 300 Hz. The reason this range is so critical is that most sounds, relevant to this low end, in your mix, will carry fundamentals and undertones in this range and will form the ‘boom’ of a track. This frequency range presents us with some of the most common problems that we will try to resolve in later chapters as so many frequencies reside in this range that their summed amplitudes alone will create metering nightmares.

We will deal with frequencies above these ranges when we come to working through the exercises otherwise it is simply a case of me writing another frequency chart and attributing descriptions for each range. I am only concerned with the relevance of these frequencies in relation to the low end and not for anything else.

Kick Drum

I find kick drum frequency ranges almost useless because in today’s music or the genres this book is concerned with, EDM and Urban, kick drums are multi-layered and in most cases samples as opposed to tuned acoustic kicks. So, remarks like ‘boost between 60 Hz – 100 Hz to add low end’, although a guide, is both misleading and unhelpful. We have a general rule in audio engineering/production: you cannot boost frequencies that are not there. Although most sounds will have a sensible frequency range we can use as a guide the kick drum is an entity on its own, simply because of the move away from using only acoustically tuned drum kits to sample-based content. Tonal synthetic kick drums are a different story entirely as the tone will have a pitch but layer that with other drum sounds and they can amass into one big mess if not handled sensibly. The TR 808, through design, behaves tonally but in quite a specific manner thanks to its clever oscillator Bridged T-network, Triggering and Accent.

To help you, purely as a guide, here is a basic chart outlining fundamental and harmonic ranges.

I have included some of the higher frequency ‘instruments’ like the Soprano voice so you can get an idea of the range of frequencies that we have to consider when mixing one frequency range in a track with another. As I said at the start of this chapter, low-end frequency ranges can only be assigned when the entire frequency content of a track is known otherwise it will be a process in isolation and when it comes to mixing that one frequency range with the rest of the track you will encounter problems in ‘fitting it in’.

 

I have covered the above in most of my eBooks and on my website www.samplecraze.com as part of my ongoing free tutorials. So, if you find the above a little overwhelming please feel free to explore my other books or head on over to my site and read at leisure.

Extract taken from the eBook Low End.

Relevant content:

Low End – what is Low End and how to Analyse it

Sinusodial Creation and Simple Harmonic Motion

Frequency and Period of Sound

Total and Partial Phase cancellation

What defines a good beat? Well, there is a term we use quite extensively when describing the overall ‘drive’ element of a track: ‘The Nod’. If you can nod to the rhythm of a song, then the beat works. The Nod actually refers to the flow of the beat, and the drive element constitutes the drum beat and bassline together. Because this book is about constructing beats, we will eliminate the bass from the equation. Bass, in itself, is a vast topic that I will cover at a later date when dealing with the low end of a track.

Most producers believe that a well-constructed beat, which has the Nod factor, comes down to two ingredients: the timing information of the whole beat and its constituents, and the dynamics of the individual components. In fact, there is far more to it than that. There are many factors that influence the flow of a drum beat and I will cover the most important ones.

I am Armenian, born in Iran, and have lived in other equally wondrous and safe havens like Lebanon and Kuwait. As a child, I had an obsession with sound, not exclusively music, but sound in its entirety. The diverse cultures to which I was exposed have afforded me the benefit of experiencing some exotic time signatures, dynamics, and timing elements. I always believed that the East held the title for advanced timing variations in music and obscure pattern structures, and for a while this was true. Today, we are blessed with a fusion of cultures and artistic practices. None are more infused with cross-cultural influences as the drum beats we incorporate in modern music.

Let’s break down the different areas that, collectively, form ‘The Nod’.

The Sounds

In dance-based music the choice of drum sounds is critical, and we have come a long way from processing live, acoustic kits into workable sounds that can live alongside a fast and driving BPM (beats per minute). Instead, we use drum samples and, in many cases, layer these samples with other samples and acoustic sounds. In the case of urban music, and the more defined and extreme sub-genre Hip Hop, we tend to go with samples from famous drum modules and drum samplers like the Emu SP1200, Roland TR808/CR78, and the MPC range—most notably the earlier versions such as the MPC60/3000.

The drum samples that we layer and process within a beat must meet very specific requirements. These include topping and tailing, mono/stereo, acoustic/noise/ tonal, and pitch/duration specifications. Let me briefly explain, ahead of the longer discussions later in this book:

  • Topping and Tailing: This process entails truncating a sample (removing dead space before and after the sample) and then normalising it (using Peak Normalisation to bring the sample’s amplitude/level up to 0dB). We do this for a number of reasons. Crucial considerations include sample triggering, aligning samples on a timeline, and referencing gains within a kit or beat.
  • Mono/Stereo: A drum sample that displays the same information on both channels is a redundant requirement unless the dual-channel identical information is required when layering using the ‘flip and cancel’ method. (Watch my video Art of Drum Layering Advanced, or read the article I wrote for Sound On Sound magazine entitled ‘Layers of Complexity’ for more information.) The only other instance where a stereo drum sample would be used is if the left and right channel information varies, as would be the case if a stereo effect or dynamic process were applied, or if the sample were recorded live using multi microphones, or if we were encoding/decoding mid/side recordings with figure-8 setups. We try to keep kick samples, in particular, in mono. This is because they remain in the center channel of the beat and, ultimately, the mix. For other samples like snares, claps, and so on, stereo can be very useful because we can then widen and creatively process the sample to taste.
  • Acoustic/noise/tonal: Acoustic drum sounds will invariably have been tuned at the playing and recording stages but will need to be re-tuned to the key of the track in which the beat lies. Tonal drum samples, like the legendary 808 kick drum, will also have to be tuned. More importantly, the frequency content of the sample will determine what type of dynamic processing can be applied. A sine-wave based tonal kick will have no harmonics within the waveform and will, therefore, be reliant on innovative dynamic processing techniques. Noise-based samples contain little or no tonal information, so require a different form of processing because the frequency content will be mainly atonal.
  • Pitch and Duration: Ascertaining and tuning atonal drum sounds is a nightmare for many, and this area is covered extensively in later chapters using specific tools and processes. Extending duration with pitch changes, altering pitch without altering duration, using time-stretching, and modulating pitch and/or duration using controllers and automation: all these are excellent forms of pitch manipulation.

Timing

  • Producers spend more time using the nudge feature and timeline of their DAW, refining timing information for beats, than on other time-variant processes. We have access to so many time-variant tools today that there really is no excuse to be unable to create either a tight and strict beat, or a loose and wandering beat, exactly as required. In fact, we have some nice workarounds and ‘cheats’ for those that have problems with timing issues, and I will cover these in more detail later.
  • Great timing in beat construction requires understanding several phenomena and techniques that I will explain in this book—BPM and how it relates to realistic timings for ‘played’ rhythms; Quantize, both in terms of divisions and how to alter these divisions; Ghost Notes and how they relate to perception; and Shadowing Beats, including the use of existing loops and beats to underlie, accent, and support the main beat. For example, if your drum beat is too syncopated and has little movement, you can reach for a Groove Quantize template in your DAW, or use other funky tools such as matching slice and hit points to existing commercial breaks.
  • The perception of a timing variance can be achieved in more than one way. Strangely enough, this leeway has been exhausted to death by Akai with the original Linn-designed pads and contacts. After the MPC 60 and 3000, Akai had no more timing variances in their hardware that could be attributed to ‘the MPC swing and sound’. Far from it. The timing of their DSP is rock solid. The timing of the pad’s initial strike, processed as channel pressure, note on/off and velocity curves, is what adds to the timing ‘delay’. This can be emulated on any pad controller that is sample-based because it is not hardware-specific. To further understand the perceptual formula, we need to look at the sample playback engine of all the top players. Bottom of the list lies Akai with their minimum sample count requirement, which demands so many cycles that if you truncate to a zero point sample start, the unit simply cannot cope with it. Add this ‘dead space’ requirement before a sample can be truthfully triggered to a pad that has inherent latency (deliberately designed by the gifted Roger Linn), and you end up with the ‘late’ and ‘loose’ feel of the MPCs. The sample count issue has now been resolved, and in fact, was corrected from the Akai 2500 onwards. I bring this up so that you are aware that there are very few magic boxes out there that pull out a tight yet loose beat. Nope. They all rely on physics to work. Yet, because of that requirement, we can work around the limitations and actually use them to our advantage. The MPCs have explored and exhausted these limitations quite successfully.
  • I love using pads to trigger drum sounds as it makes me feel more in touch with the samples than a mouse click or keyboard hit. The idea that drums must be ‘hit’ is not new, and the interaction that exists in the physical aspect of ‘hitting’ drum pads is one that makes the creative writing process far more enjoyable and ‘true’ to its origins. After all, the Maya didn’t have keyboard controllers. For this book, I will be using the QuNeo to trigger samples, but occasionally I will also trigger via the keyboard (Novation SLMK2), because spanning templates can be a little confusing for those that do not understand the manufacturers’ default GM templates.
  • Early and late processes in aligning beat elements are also a creative and clever workaround for improving static syncopated beats. Simple movements of individual hits using grid subdivisions can add motion to strict 4/4, 3/4 and 6/4 beats, which are the common signatures used in modern music.

Dynamics

  • Although we think of our brains as really smart organs, they are not actually that smart when it comes to deciphering and processing sight and sound. If you were to snap your fingers in front of your face, the sound would reach your brain via the ears before the visual information reaches your brain via the eyes. That may sound strange because light travels faster than sound, but it isn’t that strange when you take into account the time it takes the brain to decipher the different sensory input. In addition, the brain does not recognise frequency or volume without a reference. This is what memory is for: referencing. The brain has an instinctual response to already referenced frequencies and can turn off like a tap in a hurry when confronted with the same frequencies at the same amplitudes. However, when presented with the same frequencies at varying amplitudes the brain has to work to decipher and reference each new amplitude. This keeps the brain active and therefore interest is maintained. Next time you decide to compress your mix into a square wave because you think it will better ‘carry your mix’ across to listeners by rattling their organs, think twice. A narrow banded dynamic mix simply shuts the brain down, which then goes into ‘irritation mode’ because it has already referenced the constant amplitude for the frequency content in your track. The same processes take place when dealing with drum beats. The most interesting drum beats have acres of dynamic movement and do not rely on a single static amplitude for all the frequencies in the beat. Simple tasks, like altering the individual note velocities or amplitudes, will add huge interest to your beats. I would be surprised if Clyde Stubblefield maintained the same 127 velocity across all his hits whilst playing the drums.

Layering

  • Individual drum sounds can be layered to give both depth and width, resulting in a texture that can be both dynamic and interesting. If you need to delve into this area in more detail please refer to my book Art of Drum Layering, or the Advanced Drum Layering video which explores very specific layering techniques using phase cancellation, mid/side, and so on. But don’t confine yourself to drum sounds for layering. I have sampled and used kitchen utensil attacks, edited from individual amplitude envelope components, for the attack part of my snares and hi-hats, cardboard boxes close-miked with a large-diaphragm capacitor to capture the boom for kick bodies, and tapping on the head of a directional mic for some deep, breathy samples with which to layer my entire beats, and so on. If you can sample it, hell, use it!
  • Whole drum loops, treated as layers, can add vibrancy and motion to a static drum beat. Layering loops under a beat not only helps in acting as a guide for those that are not very good at drumming or creating grooves but also allows for some interesting new rhythms that will make the listener think you have incredible insight into beat making.
  • Layering tones beneath drum beats is an old and trusted method of adding low end. However, simply throwing a sine-wave under a beat doesn’t make it ‘have low-end’. You need to edit the waveform both in terms of frequency (pitch) and dynamics (in this instance: duration and velocity) and take into account the interaction between the low-frequency content of the beat and sine-wave along with the bass line. Many a Prozac has been consumed during the mix-down of this type of scenario.

Modulation

  • Using modulators to create both motion and texture in a drum beat is not as hard as it may seem at first. The trick, as with all these processes, is to understand the tools and their limitations and advantages. For example, a low-frequency oscillator (LFO) triggering the filter cut-off using a fast ramp waveform shape can add a lovely squelchy effect to a clap sample. Another technique that I have often used is assigning a sine-shaped LFO at a low rate with filter resonance as its destination to run through the entire beat. I then layer this ‘effected’ version with the original dry beat. This gives the perception of tonal changes throughout the beat, even though it is not random.

Drum Replacement/Ripping Beats

  • Creative beat construction techniques using drum replacement and ripping beats include: substituting your own drum samples for drum sounds within a beat; using the timing information from an existing drum beat as a Quantize or groove template for your own beats; ripping both MIDI and dynamic data from an existing drum beat; and using two beats at different tempos, matching their data to create a new beat that combined drum elements from both beats.

Let’s now look at some of the techniques used to shape and hone drum beats into working ‘Nods’. I will try to incorporate as much of the above as possible into real-life exercises using examples of common chart hits. In terms of tools, I have found that a decent DAW, a capable pad controller, and a good all-round keyboard controller will cover the areas that we require. A pad controller is not crucial, but it does allow for more interaction and dynamic ‘feel’ (we all love to hit pads).

Extract taken from the eBook Beat Construction

I am often asked why I teach my students how to mix to a pink noise profile, be it at the channel stage, or pre-master prepping. The answer is simple: ‘the most important aspect of production is the understanding and management of relative levels.’

When I first began this wonderful and insane journey into audio production I was blessed to have had producer friends that were also my peers. In those ancient days, the industry was very different. The community aspect was both strong and selfless. We were not competing with each other. Instead, we chose to share our knowledge and techniques. It was then that I was introduced to noise as a mixing tool, and coupled with my sound design experience I took to it like a pigeon on a roof.

I was taught the old school method of tuning your ears and mindset to work from a barely audible mix level. If I did not hear everything in the mix then I had to go back and work the quieter channels. If something stood out in the mix then I knew the culprit had to be re-balanced, and all of this relied heavily on relative levels.

Relative levels in a mix context deals with the relationships between all sounds, and that includes effects and dynamics. You may think that relative levels refers only to volume but that is not entirely accurate. Relative levels deals with all level management, from sounds to effects and dynamics. Eq is an excellent example of frequency/gain management, but so are reverb levels, balancing parallel channels or wet/dry mix ratios, and so on……..

An example of how this technique helps the student to understand all areas of relative gains is by throwing in the classic reverb conundrum. We’ve all been there. If there is too much reverb level then the sound will either lose energy through reverb saturation, sound too distant if the wet and dry mix is imbalanced, or sound out of phase. By continual use of this technique, the student learns how well the sound and its effect sit together, whether the dry/wet ratio is right and whether the right reverb algorithms were used. This level of familiarity can only help the student and is the only simple working way of attuning the ears to not only hear level variances but also if something somewhere sounds ‘wrong ‘.

In some ways, this is very much like ear training but for producers as opposed to musicians/singers.

When I feel my students have reached an acceptable level conducting level and pan mixes (another old school apprentice technique), I move them onto pink noise referencing. By the time they have finished countless exercises using all manner of noise responses, they develop an instinctive understanding of gain structuring every aspect of the signal path, from channel to master bus, and with that comes an understanding and familiarity of what sounds natural and ‘right’.

Supplemented with listening to well-produced music this technique has saved my students both time and money and it is such a simple technique that even Trump could do it………well…..with help of course.

Eddie Bazil
Samplecraze

If you prefer the visual approach then try this video tutorial:

Mixing to Pink Noise

Relevant content:

DIY Mastering using Pink Noise

The Different Colours of Noise