The previous instalment ended with a ‘Mixing With Headphones’ session template, set up and ready for mixing. In this instalment we’ll start putting that template into practice using the EQ tools it contains; in the fifth instalment we’ll look at dynamic processing (compression and limiting), and in the sixth and final instalment we’ll look at spatial processing (reverberation, delays, etc.). But first a word about ‘visceral impact’ as defined in the second instalment of this series, and some basic mixing rules to get your mix started…

Visceral Elusion

We know that with headphone monitoring/mixing there is no room acoustic, no interaural crosstalk, and no visceral impact to add an enhanced (and perhaps exaggerated) sense of excitement. In the second and third instalments of this series we discussed ways of working around the frequency response and spatial issues, but the lack of visceral impact is a trap we need to be constantly aware of – especially when first transitioning from speaker mixing to headphone mixing.

A good pair of headphones can effortlessly reproduce the accurate and extended low frequency response that acousticians and studio owners dream of achieving with big monitors installed in expertly designed rooms and costing vast sums of money. However, when mixing with headphones we have to remember that those low frequencies are being reproduced directly into our ears via acoustic pressure coupling – which means we do not experience them viscerally (i.e. we do not feel them with our internal organs aka our viscera) as we do when listening through big monitors. There is no visceral impact, which means we must be very careful about how much low frequency energy we put into our mixes. Increasing the low frequencies until we can feel them is not a good idea when mixing with headphones…

This is where the musical reference track and the spectrum analyser in our ‘Mixing With Headphones’ template are particularly valuable. If the low frequencies heard in our mix are pumping 10 times harder than the low frequencies heard in our reference track and seen on the spectrum analyser, they’re probably not right. It might be tempting to succumb to premature congratulation and declare that our mix is better than the reference because it pumps harder, but it is almost certainly wrong. That’s the point of using a carefully chosen reference track that represents the sonic aesthetic we’re aiming for: if our mix strays too far from the reference in terms of balance, tonality and spatiality then it is probably wrong and we need to rein it in before it costs us more time and/or money in re-mixing and mastering.

How do we avoid such problems when mixing with headphones? Read on…

BASIC RULES FOR HEADPHONE MIXING

The following is a methodical approach to mixing with headphones based on prioritising each sound’s role within the mix, introducing the individual sounds to the mix in order of priority, and routinely checking the affect that each newly introduced sound is having on the evolving mix by using the tools described earlier: mono switch, goniometer, spectrum analyser with 6dB guide, and a small pair of desktop monitors. This methodical approach allows us to catch problems as they occur, before they’re built into our mix and are harder to undo. The intention is to create a mix that, tonally at least, should only require five minutes of mastering to be considered sonically acceptable.

Note that the methodical approach described here is not suitable for a mix that needs to be pulled together in a hurry; for example, mixing a live gig that has to start without a soundcheck or rehearsal, or dealing with advertising agency clients who don’t understand why a 30 second jingle takes more than 30 seconds to record and mix. In those situations ‘massage mixing’ is more appropriate, i.e. pushing all the faders up to about -3dB, getting the mix together roughly with faders and panning, focus on keeping the most important sounds in the mix clearly audible, and continue refining the mix with each pass until the gig is finished or the session time runs out. In these situations, Michael Stavrou’s sculpting analogy [as explained in his book ‘Mixing With Your Mind’] is very applicable when he advises us to “start rough and work smooth”. Get the basic shape of the mix in place before smoothing out the little details, because nobody cares about the perfectly polished snare sound if they can’t hear the vocal.

Establishing The Foundation

For the strategic and methodical approach described here, start by establishing the foundation sounds that the mix must be built around. For most forms of popular music those foundation sounds are the kick, the snare, the bass and the vocal. Each of the foundation sounds should have what Sherman Keene [author of ‘Practical Techniques for the Recording Engineer’] refers to as ‘equal authority’ in the mix – meaning each foundation sound should have the appropriate ‘impact’ on the listener when we switch between them one at a time, and they should work together as a cohesive musical whole rather than one sound dominating the others. A solid stomp on the kick pedal should hit us with the same impact as a solid hit on the snare, a firm pluck of the bass guitar, and a full-chested line from the vocalist. Those moments should feel like they hit us with the same impact, and they should feel like they belong together in the same performance. That feeling is harder to sense without the visceral impact of speakers, but with a little practice and cross-referencing against our reference track we can get there.

Start with the most important sound in the foundation, and get that right to begin with. For most forms of popular music that will be the vocal, so start the mix with the vocal only and get it sounding as good as possible on its own, where it is not competing with any other sound sources. You may need to add one or two other tracks to the monitoring – one for timing and one for tuning – to provide a musical context for editing and autotuning but don’t spend any time on those tracks yet. Focus on getting the vocal’s EQ and compression appropriate for the performance and the genre. Aim to create a vocal track that can carry the song on its own without any musical backing. Create different effects and processing for different parts of the vocal performance to suit different moods or moments within the music – for example, changing reverberation and delay times between verses and choruses, using delays or echoes to repeat catch lines or hooks, and similar. Use basic automation to orchestrate those effects, bringing them in and out of the mix when required as shown below. Note that placing the mutes before the effects simplifies timing the mute automation moves and also allows each effect (delay, reverb, etc.) to play itself out appropriately instead ending abruptly halfway through – the classic rookie error.

Once we have the vocal (or whatever is the most important sound in the mix) ready, check it in mono, check it on the goniometer and check it on the desktop monitors to make sure that the stereo effects and processors are behaving themselves. Cross reference it with the reference track in the Mixing With Headphones template to make sure it is sounding appropriate for the genre.

Introduce the other foundation sounds one at a time; in this example they will be the kick, the snare and the bass. Use EQ, compression and spatial effects (reverberation, delay, etc.) to get each of these sounds working together in the same tonal perspective and dynamic perspective as the vocal, and in the desired spatial perspective against each other and against the vocal. Toggle each plug-in and effect off and on repeatedly to make sure it is making a positive difference. If not, fix it or remove it because processors that are not making a positive difference are like vloggers at a car crash: they’re ultimately part of the problem. Check each foundation sound and its processing in mono, check it on the goniometer and check it on the desktop monitors to make sure that its stereo effects and processors are behaving themselves.

With all of the foundation sounds in place we may need to tweak the levels of any spatial effects on the vocal that have become perceived differently after introducing the other foundation sounds.

Orchestrate the effects for the foundation sounds (as described above for the vocal) to help each sound stand out when it’s supposed to stand out and stand back when it’s supposed to stand back, thereby enhancing its ability to serve the music.

Always consider the impact each newly-introduced sound is having on the clarity and intelligibility of the existing sounds in the mix and, particularly, its impact on the most important sound in the mix – which in this example is the vocal. We should not modify the vocal to compete with the other sounds, rather, we should modify the other sounds to fit around or alongside the vocal. After all, in this example the vocal is the most important sound in the mix and we had it sounding right on its own to begin with. If adding another sound to the mix affects the sound of the vocal (or whatever the most important sound is), we need to make changes to the level, tonality and spatiality of the added sound. That is why we prioritised the sounds to begin with: to make sure the most important sounds have the least tonal, dynamic and spatial compromises, and therefore have the most room to move and feature in the mix.

Note that the goal here is to fit the other sounds around and alongside the vocal, not simply under the vocal. Putting foundation sounds under the vocal is the first step towards creating a karaoke mix or a layer cake mix; more about those in the last instalment of this series…

After introducing each new foundation sound to the mix be sure to check it in mono, check it on the goniometer (as described in the previous instalment) and check it on the desktop monitors to make sure it is not misbehaving in ways we cannot identify in headphones but will become apparent if heard through speakers.

With the foundation mix done, save a copy that can be returned to in case things spiral out of control. Thank me later when/if that happens…

Beyond The Foundation

Introduce the other sounds one at a time, weaving each of them among and around the foundation sounds while ensuring all sounds remain in the desired tonal perspective, dynamic perspective and spatial perspective with each other. The following text describes strategies for achieving tonal perspective; strategies for achieving dynamic perspective and spatial perspective are discussed in the forthcoming instalments.

Every new sound introduced to the mix has the potential to change our perception of the existing sounds in the mix, so check for this and process accordingly without messing with the foundation sounds. Pay careful attention to how each new sound impacts the audibility of spatial effects (reverbs, delays, etc.) that have been applied to existing sounds, and adjust as necessary.

Loud Enough vs Clear Enough

When balancing sounds together in the mix, always be aware of the difference between “not loud enough” and “not clear enough”. Novice engineers assume that if they cannot hear something properly it is not loud enough and will therefore reach for the fader. More experienced sound engineers know that often the sound described as “not loud enough” is in fact loud enough but is not clear enough due to some other issue with how it fits into the mix (e.g. its tonality, its dynamics or its spatial properties). And in some cases we realise that the sound deemed not loud enough is actually being buried or masked by another sound that is too loud in the mix and needs to be fixed.

Bring each sound up to the level it feels like it is supposed to be at from a performance point of view, regardless of how clear it is. We can determine if its level is right by soloing it against other sounds that are meant to have similar authority in the mix. If the sound is at the right performance level when solo’d against sounds of similar authority but is hard to hear properly in the mix, the problem is not the sound’s fader level but rather its clarity and/or separation within the mix.

In most cases this means the sound’s overall level is correct but some parts of its frequency spectrum are either not loud enough or are too loud, and we need to use EQ to boost or cut just those parts of the sound’s frequency spectrum to make it clear enough and bring it into the correct tonal perspective for the mix. We’ll discuss this process later in this instalment.

If it is hard to find the right level for a sound that gets too loud at some times and too soft at other times, it suggests that sound is probably not in the same dynamic perspective as the other sounds and will require careful compression to rein it in. (See ‘Dynamic Processing’ in the next instalment.) Sometimes a sound is in the correct tonal perspective and dynamic perspective for the mix but gets easily lost behind the other sounds, or continually dominates them, due to having an incorrect spatial perspective (e.g. too much reverb). We use spatial processing to create, increase or decrease the sound’s spatial properties and thereby assist with separation (See ‘Spatial Processing’ in the sixth instalment of this series.)

It’s also possible that the problem is unsolvable at the mixing level due to ridiculous compositional ideas that have since become audio engineering problems. For the remainder of this series let’s remove that variable by assuming we’re working with professional composers who know how to build musical clarity and separation into their compositions.

To solve these ‘loud enough but not clear enough’ problems we use tonal processing to adjust the balance of individual frequencies within a sound, dynamic processing to solve problems with sounds that alternate between too loud and too soft, and spatial effects to provide separation from competing sounds. Let’s start with tonal processing, or, as it is generally referred to, ‘EQ’ and ‘filtering’…

EQUALISATION & FILTERING

The use of equalisation and filtering serves three purposes in a mix: correcting sounds, enhancing sounds, and integrating sounds. In our ‘Mixing With Headphones’ template described in the previous instalment we added three EQ plug-ins to each channel strip. Here’s what they’re for…

Corrective EQ

This is used to fix fundamental problems in individual sounds and clean them up before putting them in the mix, which means we should choose a clean EQ plug-in that is not designed to impart any tonality or character of its own into the sound. The emphasis here is to use something capable rather than euphonic. A six-band fully parametric EQ with at least ±12dB of boost/cut, along with high and low pass filtering and the option to switch the lowest and highest bands to shelving, is a good choice.

Corrective EQ is used to make an excessively dull sound brighter or to make an excessively bright sound duller, to fix sounds that have too much or too little midrange, and to fix sounds that have too much or too little low frequency energy. It is also used to remove or reduce the audibility of any unwanted elements within the sound such as low frequency rumble (high pass filter aka low cut filter, low frequency shelving cuts), hiss and noise (low pass filter aka high cut filter, high frequency shelving cuts), and unwanted ringing and resonances (notch filters, dips).

The goal of corrective EQ is to create objectively good sounds. What is an objectively good sound? It’s a sound that does not contain any objectively bad sounds, of course. It is hard to define what sounds are objectively ‘good’, but it’s easy to define what sounds are objectively ‘bad’.

Objectively ‘bad’ sounds are resonances and rings, low frequency booms and rumbles, unwanted performance noises and sounds, hiss and noise, and similar unmusical and/or distracting elements that don’t belong in the sound as we intend to use it.

One of the most common applications of corrective EQ is removing unwanted low frequency energy. Most sounds contain unwanted low frequency energy below the fundamental frequency of the lowest musical note in the performance. It may not seem like much on any individual track but the unwanted low frequency content on each track accumulates throughout the mix, with two results. Firstly, it reduces the impact and clarity of kick drums, bass lines, low frequency drones and other sounds that are legitimately occupying that part of the frequency spectrum. Secondly, most monitoring systems are not capable of reproducing this unwanted low frequency information reliably (particularly below 70Hz), and forcing them to reproduce it affects their ability to reproduce other frequencies that are within their range – which thereby affects their ability to reproduce the mix. It’s like forcing one horse to pull a cart that requires two horses.

Strategically removing unwanted low frequency information from individual sounds brings clarity and definition to our mixes while also allowing a broader range of monitoring systems to reproduce our mixes properly. With these benefits in mind, it is always worthwhile starting any EQ process by viewing the sound on the spectrum analyser (built into the Mixing With Headphones template) and looking for activity in the very low frequencies that has no musical value. This will be low frequency activity that remains visible, whether audible or not, and can be seen bobbing up and down at the far left side of the spectrum analyser regardless of what musical parts are being played. Removing or reducing this unwanted low frequency information with a carefully-tuned high pass filter or low frequency shelving EQ (in either case pay attention to the cut-off frequency and the slope) will clean up the individual sounds and the mix considerably.

We use corrective EQ to remove or significantly reduce the audibility of the objectively ‘bad’ parts of the sound, thereby leaving us with only the objectively ‘good’ parts of the sound for enhancing and integrating into our mix. As always, after applying corrective EQ we should check the results against the original sound to make sure we have made an improvement and not just a difference.

Enhancing EQ

This is used to create subjectively ‘good’ sounds from the objectively ‘good’ sounds we made with corrective EQ as described above. What are subjectively ‘good’ sounds? They are sounds that contain no objectively bad sounds (which we removed with corrective EQ), and are good to listen to while also bringing musical value or feeling to the mix. We can do whatever we like with the objectively ‘good’ sounds to turn them into subjectively ‘good’ sounds, as long as we don’t inadvertently re-introduce the objectively ‘bad’ sounds we removed with the corrective EQ.

For this enhancing purpose we can use an EQ plug-in with character to introduce some euphonics into the sound. This could be a software model of a vintage tube EQ that imparts a warm or musical tonality, and/or something with unique tone shaping curves like the early Pultecs, and/or gentle Baxandall curves for high and low frequency shelving. Unlike the corrective EQ, the enhancing EQ doesn’t need corrective capabilities (a lot of vintage EQs did not have comprehensive features), and we can make up for any shortcomings here by using the corrective EQ and the integrating EQ.

The enhancing EQ is where the creative aspect of mixing begins: crafting a collection of sounds that might be individually desirable but, more importantly, collectively help to serve the meaning, message or feeling of the music. One of the goals here is to bring out the musical character of each individual sound while giving it the desired amount of clarity so we can hear ‘into’ the sound and appreciate all of its harmonics and overtones, along with the expression and performance noises that help to bring meaning to the mood of the music. In other words, to enhance its musicality.

When applying enhancing EQ try to use frequencies that are musically and/or harmonically related to the music itself. Most Western music is based around the A440 tuning reference of 440Hz, so that forms a good point of reference. The table below shows the frequencies of the notes used for Western music based on the tuning reference of A440, from C₀ to B₈. The decimal fraction part of each frequency has been greyed out for clarity and also because we don’t need that much precision when tuning an enhancing EQ. Integer values are accurate enough…

If we jot down the frequencies of the notes that exist within the scale(s) of the piece of music we’re mixing, we can lean into those frequencies when fine-tuning our enhancing EQ. For example, let’s say we had an enhancing EQ that provided a small boost at 850Hz. It’s using a frequency that does not exist in any Western musical scale that is based on the A440 tuning reference; 850Hz sits in between G^#₅ (830.61Hz) and A₅ (880Hz), and is therefore not a particularly musical choice. Nudging that enhancing boost down towards 830Hz (G^#₅₎ or up towards 880Hz (A₅) will probably sound more musical and is, therefore, definitely worth trying.

We should always nudge our enhancing EQ boosts towards frequencies that do exist within the scale(s) of the music we’re mixing – we wouldn’t let a musician play out of tune, so why let an enhancing EQ boost be out of tune? Likewise, we should always nudge our enhancing EQ dips towards frequencies that don’t exist within the scale(s) of the music we’re mixing – if we’re going to dip some frequencies out of a sound, try to focus on frequencies that aren’t contributing any musical value. Less non-musicality means more musicality, right?

It’s also worth noting that when a sound responds particularly well to a boost or a cut at a certain frequency (let’s call that frequency f), it will probably also respond well to a boost or a cut an octave higher (f x 2) and/or an octave lower (f / 2). More about that shortly…

Integrating EQ

While corrective EQ and enhancing EQ are used for cleaning up and creating sounds, integrating EQ is used for combining sounds together, i.e. integrating them into a mix.

Creating good musical sounds with enhancing EQ is fun and satisfying, and might even be inspiring, but we must be constantly aware of how those individual sounds will interact when combined together in the mix. It’s common to have, for example, a piano and a strummed acoustic guitar that sound great individually but create sonic mud when mixed together because there is too much overlapping harmonic similarity between them. They are both using vibrating strings to create their sounds and therefore both have the same harmonic series, which makes it harder for the ear/brain system to differentiate between them if they’re playing similar notes and chords.

Another form of sonic mud occurs when composers create music using sounds from different sample libraries and ‘fader mix’ them together. Because each individual sample sounds great in isolation, the assumption is that simply fader mixing them together will sound even greater. That is like pouring a dozen of our favourite colour paints into a bucket and giving it a stir on the assumption it will create our ‘ultimate’ favourite colour. What do we get? A swirling grey mess, every single time, and it’s the same when mixing a collection of individually enhanced sounds.

That’s what integrating EQ is for: helping us to integrate – or ‘fit’ – the individually enhanced EQ sounds together into a mix or soundscape, ensuring they all work together while remaining clear and audible. As with our choice of corrective EQ, the integrating EQ should be a clean plug-in that does not impart any tonality or character of its own. A six-band fully parametric EQ with high and low pass filtering and the option to switch the lowest and highest bands to shelving is a good choice here.

We use integrating EQ to maintain clarity and tonal separation within a mix. We listen to how our enhanced EQ sounds affect each other when introduced to the mix, and we make appropriate tweaks with integrating EQ to fix any conflicts and restore the preferred elements of each sound. How?

INTEGRATING EQ EXAMPLE

Let’s go back to the earlier example of the piano and the strummed acoustic guitar, where each instrument sounded good on its own but both instruments lost clarity and tonal separation when mixed together. Imagine the piano and the acoustic guitar have been loaded into our Mixing With Headphones template. Using the individual channel solo buttons along with the spectrum analyser on the mix bus allows us to examine the frequency spectrums of the piano and the acoustic guitar individually. Conflicts between their frequency spectrums can be identified by temporarily adjusting both sounds to the same perceived loudness, then alternating between soloing each sound individually and soloing both simultaneously.

For this example let’s say that, due to our clever use of corrective EQ and enhancing EQ, both sounds are full-bodied and rich but therein lies the first problem: they’re both competing for our attention in the midrange. That means we have to apply integrating EQ with the goal of making them work together in the midrange.

We start by prioritising each competing sound based on its musical and/or textural role in the mix. We want to use minimal integrating EQ on foundation sounds and featured sounds that play musically significant parts, preserving the musicality and tonality that we’ve already highlighted in those sounds with the enhancing EQ. Textural sounds and background sounds are more forgiving of tonal changes so it is smarter to apply any significant integrating EQ changes to those sounds. Let’s examine the roles of the acoustic guitar and the piano in this particular piece of music to prioritise them accordingly.

Although the rhythm and playing of the acoustic guitar is helping the drums and bass guitar to propel the music forward, it is never actually featured in the mix of this piece of music. Therefore its primary purpose is textural; it provides a gap-filling layer of musical texture in the background. We can use a lot of integrating EQ here if we have to, as long as it doesn’t interfere with the acoustic guitar’s textural role.

What about the piano? In this piece of music, the left hand is playing a textural role with gentle low chords that complement the bass guitar and thicken the acoustic guitar. The right hand, however, is playing a musically significant role by adding sharply punctuating chords along with short melodies that fill the spaces between vocal lines, and those melodies often conflict with the acoustic guitar. These observations tell us that we can manipulate the piano’s lower frequencies (left hand, textural) as required to make it work in ensemble with the bass guitar and the guitar, but we need to be very conservative with any EQ applied to the midrange (right hand, musically significant) to avoid altering the tonality of the punctuating chords and short melodies.

Having established that the acoustic guitar’s tonality has a lower priority than the piano’s tonality in this piece of music, it is an appropriate starting place for applying integrating EQ.

Let’s make a clarifying dip in the acoustic guitar’s spectrum, right where the two sounds share overlapping peaks in their spectrums – which is almost certainly the cause of the problem.

This upper midrange dip will change the tonality of the acoustic guitar, of course. Fortunately, in this case it will make it more subdued and appropriate for the background textural role it plays in the mix. More importantly, however, it will contribute to the overall clarity of the mix by creating room for the piano without altering the piano sound itself.

To fine-tune the depth of the dip (i.e. how many dB to cut) and the width of its bandwidth (Q), we should switch the applied EQ in and out while soloing the guitar and piano separately and together and also checking the results on the spectrum analyser. We want to dip just enough out of the acoustic guitar to leave room for the right hand parts of  the piano to be heard clearly, but no more.

We might find that the required depth and bandwidth of the dip has improved the clarity of the piano within the mix, but the acoustic guitar has become less interesting. We can musically compensate for this change in the acoustic guitar’s tonality by adding small boosts an octave above and below the dipped frequency in the acoustic guitar’s spectrum as shown below.

Again, switching the EQ on and off while soloing the instruments individually and together while watching the spectrum analyser will help us get the settings just right. [The compensating EQ technique shown above can be applied whenever a sound has been given a necessary integrating EQ peak or dip that subtracts some of that sound’s musicality: add small boosts an octave either side of any significant dips, and small cuts an octave either side of any significant boosts.]

The process of applying integrating EQ and compensating EQ to the acoustic guitar might reveal other areas worth working on. For example, let’s say this ‘soloing with spectrum analysis’ process revealed some upper harmonics in the piano sound that were worth bringing out. Applying a small integrating EQ dip in the acoustic guitar’s spectrum will create room for those upper harmonics of the piano to shine through, and applying a small compensating EQ boost in the acoustic guitar’s spectrum an octave higher will do the same for the acoustic guitar’s upper harmonics.

All of the integrating and compensating EQ changes detailed above have improved the clarity of, and separation between, the acoustic guitar and the right hand parts of the piano by focusing on their respective textural and musical roles. We’ve already established that the left hand parts of the piano were playing a textural role, as is the acoustic guitar, so let’s see how they sit alongside one of the mix’s foundation sounds that shares some of the same spaces within the frequency spectrum: the bass guitar.

Adding the bass to the piano and acoustic guitar, and switching/solo’ing between them, shows that there is some worthwhile upper harmonic detail in the bass sound that fits nicely into a natural dip in the piano’s spectrum but is being masked by one of the complementary EQ boosts we added previously to the acoustic guitar. Because the bass is a foundation sound that we’ve already got sounding right within the foundation mix, we want to avoid altering it if possible because it is one of the internal references for our mix. Rather than boosting the upper harmonics of the bass, we’ll reduce the complementary boost added previously to the acoustic guitar just enough to allow those upper harmonics of the bass to be audible again – as shown below.

This process has also revealed that between the bass, the left hand of the piano and the acoustic guitar there is more low frequency bloom in the mix than we’d like. It’s not necessarily boomy or wrong, but it is bordering on sounding bloated and muddy in the low frequencies – especially when compared to the low frequencies in the reference track we added to the Mixing With Headphones template before starting this mix. We don’t want to change the low frequencies in the bass because we got them right when establishing the foundation mix, and we know that any alterations to the foundation mix are likely to result in a ripple of changes throughout the mix. In this example, reducing the low frequencies of the bass to minimise the risk of the mix sounding ‘bloated’ will make the kick sound as if it has too much low frequency energy or is generally too loud. This will lead us to make changes to the kick, and the domino effect will topple through the mix from there.

Because we’ve been working with the acoustic guitar so far, we’ll start there by adding a subtle low frequency shelf or a gentle high pass filter to pull down its low frequencies just enough to clarify what is happening between the bass guitar and the left hand of the piano parts.

From here we can see the left hand parts of the piano sound, which are playing a textural role, are the remaining cause of the excessive bloom. We can wind them back with some low frequency shelving or a gentle high pass filter on the piano’s integrating EQ.

While we’re working on the piano, let’s bring out those upper harmonics we revealed earlier by adding a small boost in the piano in the same area we previously made a small dip in the acoustic guitar.

All of the integrating EQ and compensating EQ changes detailed above have resulted in the bass, piano and acoustic guitar sitting together clearly and musically in the mix, and yet most of that improvement was achieved by making changes to the lowest priority sound of the three: the acoustic guitar. Its entirely textural role in the mix made it the most sensible choice to make integrating EQ changes to. Two very subtle changes were made to the piano to improve its placement in the mix, and no changes were made to the bass guitar – which is in keeping with our goal of using the foundation sounds as a point of reference to build the mix around.

We should not get too hung up on soloing the acoustic guitar and worrying about how its sound has been changed by the EQ when heard in isolation. The reality in this situation is that it doesn’t matter what the acoustic guitar sounds like in isolation (i.e. when solo’d) because the listener is never going to hear it in isolation, and that’s because it is never featured in the music. It remains a background textural sound. Therefore the only thing that really matters beyond its musicality is how it affects other sounds in the mix. In this example, the applied integrating EQ has allowed the guitar to sit nicely behind the piano rather than under it. As we will see in the following illustrations, the acoustic guitar’s spectrum (and therefore its tonality) has been altered to allow it to fulfil its spectral role in the mix: filling in the spaces between the other instruments.

The illustration below adds the kick drum’s spectrum (shown in orange) to the image so we can see how it works with the bass and the piano.

As shown above, we have used the strategic application of integrating EQ to alter the perceived volume of each sound in the mix by a combination of boosting important parts of a given sound’s spectrum and/or cutting parts out of competing sounds’ spectrums, rather than making global ‘brute force’ fader changes. Each of these sounds was already loud enough in the mix, it just wasn’t clear enough, and we’ve used integrating EQ to clarify it.

The illustration below shows the spectrums before any integrating EQ was applied. There is too much overlap in significant parts of each sound’s spectrum, resulting in a poor mix that is lacking in clarity and tonal separation.

Complementary EQ, Shared EQ & Opposite EQ

In the integrating EQ example given above we introduced the concept of complementary EQ, where an EQ cut was accompanied by complementary boosts applied to the same sound, typically an octave (or other harmonically valid interval) above and/or below the centre frequency of the cut. If the cut was, say, -2dB at 880Hz, the complementary boosts would be placed at 440Hz (an octave below 800Hz) and 1.76kHz (an octave above 880Hz).

Each boost would have the same bandwidth (Q) as the cut, and each boost would start at +1dB with the intention of collectively returning the 2dB that was lost to the cut (conceptually aiming to maintain the same overall energy in the signal but redistributing it within the spectrum). However, the amount of boost and the choice of frequencies will ultimately be decided by ear, because nobody cares about the theory if the end results don’t sound good.

Sometimes an integrating EQ dip has an adverse affect on the sound it is applied to, and in those situations we can resort to shared EQ. The integrating EQ example shown above started by placing a dip in the acoustic guitar’s frequency spectrum to clarify the piano sound. Let’s say the dip needed to be -3dB at 880Hz with a bandwidth (Q) of 1.5 in order to do its job, but the change to the acoustic guitar’s tonality was more than we were willing to accept. In this situation we can copy that same EQ on to the piano, and share the 3dB difference between the two instruments. For example, perhaps a dip of -2dB is acceptable on the acoustic guitar, and we can make up the difference with a +1dB boost in the same part of the spectrum on the piano without adversely affecting its tonality. Now we have created the same 3dB difference at 880Hz required between the piano and the acoustic guitar, but have changed it from one large EQ change on one instrument to two smaller EQ changes shared between two instruments.

When using integrating EQ with sounds that are competing with each other but not in any clearly obvious or significant manner, it’s worth taking advantage of differences between each sound’s frequency spectrum by using opposite EQ. For the integrating EQ example used earlier we saw that the piano sound had a peak in the upper range of its spectrum where the acoustic guitar did not, and we put a subtle dip of matching bandwidth (Q) at that place in the acoustic guitar’s spectrum to increase the separation between the two sounds. This use of integrating EQ might have no significant effect on the acoustic guitar’s sound (perhaps the acoustic guitar doesn’t contain much musical value in that area) but creating the dip will further separate the two sounds while bringing the piano forward in a way that sounds better than boosting the peak on the piano’s spectrum – which might sound unnatural or perhaps even make certain notes ‘ping’ out (for which the piano tuner ultimately takes the blame).

TONAL PERSPECTIVE

After applying corrective EQ, enhancing EQ and integrating EQ, it is always important to check the tonal perspective of each sound. Does the tonality of each individual instrument sound as if it belongs in the same mix as the other instruments?

It is easy to lose track of tonal perspective and end up with one or two sounds that are very good when heard in isolation while also maintaining clarity and separation in the mix, but they don’t sound as if they belong in the same mix, e.g. they’re considerably brighter or duller than the other sounds. They are not in the mix’s tonal perspective.

This is the same problem that happens when we start combining sounds from different sample libraries, as mentioned earlier. Each sample library brand has their own sound engineers, producers and mastering engineers, therefore each sample library brand evolves its own ‘sound’ in the same way that some sound engineers, producers and boutique record labels evolve their own ‘sound’. The samples might all sound good individually, but there’s no guarantee (or likelihood) that samples from different sample library brands will work together without some kind of integrating EQ. It’s like sending all of the drum tracks but nothing else to one engineer to mix and master, all of the guitar tracks but nothing else to another engineer to mix and master, and all of the vocal tracks but nothing else to another engineer to mix and master – each engineer might do a great job on their parts, but there is no guarantee or likelihood that the individually mixed and/or mastered stems will automagically work together when combined in a mix. The individual sounds need to be tailored to fit together using integrating EQ, not simply layered on top of each other using fader levels.

When EQing individual sounds within a mix we must ensure they all sound as if they belong in the same mix, i.e. they have the same tonal perspective. If one sound proves to be overly bright or overly dull within the mix it should be fixed in the mix, because fixing it later is going to take more time in re-mixing and/or more cost in mastering.

After introducing any significant EQ changes to a sound – whether they’re corrective, enhancing or integrating – always solo the sound and switch the EQ in and out while checking on the spectrum analyser and the 6dB guide to make sure the sound’s tonality is behaving itself and not steering the mix towards being too bright or too dull.

By following the strategic step-by-step process demonstrated in this instalment, i.e. introducing one instrument at a time to our mix and checking it on the tools built into the Mixing With Headphones template, we can make high quality mixes in headphones that, tonally at least, should land within five minutes of mastering to sound acceptable.

THE ROOM

We were recording Gretta’s vocals in my production suite. It’s a space that Gretta is familiar with and comfortable in. It was originally designed as a drum room, so there’s acoustic treatment in there but not to the degree that you’d treat a vocal booth.

We moved the microphone around and tried different spots to find a position that didn’t accentuate the bad stuff. Once we determined where that was, we taped an ‘X’ on the floor and didn’t move from that position. Just like most people recording from home, you could hear a little bit of flutter and a little bit of background noise. In our case, we also had people walking past the studio door occasionally.

THE GEAR

For consistency’s sake we wanted to settle on one microphone for all the vocal recordings. We didn’t have a bunch of expensive mics to choose from. We settled on a Chandler TG mic. It’s certainly not a budget mic but it’s not out-of-reach either. Gretta has an amazing voice and really good control. But the biggest challenge we found was the 2 to 5kHz range, which can be harsh. We were trying to find a mic that gave a tape-like roll-off in that frequency area.

We didn’t have racks of preamps to choose from. We recorded the vocals through a pretty standard UA Apollo chain — nothing out of the ordinary. We didn’t track any plug-ins on the way in. We thought about recording with some gentle compression but I think we decided recording raw would give us more options later.

THE PSYCHOLOGY

As a vocal producer it’s your job to coax the very best performance from the artist, and a large part of that is knowing when to push and when to back off. It’s the interpersonal stuff.

These days studio engineers are expected to fix performances after the fact. Artists tend to have short attention spans now. They expect things to happen quickly and, often, budgets are tight and time is short. Your job is to make the artist feel comfortable and make sure they can perform at their best.

Gretta needed a safe space to be emotionally vulnerable, but she’s also very hard on herself and needed support at points (even when she delivers a near-perfect performance!).

MIX ON THE RUN

Gretta is definitely someone who needs to hear a high-quality rough mix as she goes. I would create a good-sounding mix with some effects, processing, and some panning of backing vocals, so Gretta could get involved with the song as we listen back during the recording process. None of what I was doing would be used in the final mix, it was just for the purposes of listening. That’s another piece of advice I’d offer: artists these days won’t understand why they’re not hearing something ‘produced’ during the recording process, in fact, if they don’t, it could badly demoralise them.

MELODYNE FORENSICS

After about a day of recording vocals for a song I would spend about a day, or a day and a half, on Melodyning and editing everything. For this album, Gretta was after a processed pop sound. So everything is very, very closely tuned and edited. Most of the harmonies have at least four stacks – maybe a couple of BVs in the left and right, with the lead in the middle. There’s an amazing level of detail, so I’m doing my best to be just as detailed with breath editing and de-clicking – making sure the fades were all exactly right. I’ll also be forensic with the time editing, making sure all of the BVs and the lead move really closely together. If every BV double was slightly out of time relative to each other then we wouldn’t achieve that super-tight pop sound.  You can read more about the pitch and timing edit process later in the article.

VOCAL CHAIN

The S’s on Gretta’s vocal were the biggest challenge. We were making quite a bright-sounding record and a very loud pop record, so we would manually de-ess in addition to using de-esser plugins.

As far as plugins go, at the start of the chain was a more forensic EQ (like a FabFilter Pro-Q3) then a couple of stages of gentle compression – we generally used a UA LA2A and a 1176. Gretta’s vocal doesn’t respond to one-stage of more severe compression, so we’re better off having a couple of stages.

I’m a big fan of the UA Pultec EQ-P1A, for a gentle presence boost or to cut some low/mid. It’s the sort of EQ plugin that forces you to use your ears and not your eyes, which I like.

At all times I’m aiming to finish with a vocal quality that sounds as close to the other songs as possible.

Consistency is everything for this album. The goal was not to have a really organic vocal performance, but a consistent vocal that would consistently would cut through at all times.

Finally, we  would often use the McDSP ML4 compressor/limiter over the vocal group. When you hear a low/mid build up or the highs build up, the ML4 lets you control that sort of thing very effectively. It’s a really useful tool at the end of vocal chain tool to just control stuff when you have a lot of vocal elements coming together.

For backing vocals, the other producer on the album, Gab Strum, is a big fan of the Eventide H3000. He’s got a hardware unit and loves it. I tend to use the SoundToys Microshift plugin or the plugin version of the H3000. Those widener plugins are great on BV stacks to make them more expansive in the mix. For the ‘oohs’ and ‘aahhs’ those type of plugins are unstoppable. I’m a fan of EchoBoy for delays as well.

MELODYNE MAESTRO

Here’s a final thought: you don’t need crazy-expensive tools to do a good job! Close editing and a basic copy of Melodyne Essential (A$129) will get you 90 percent of the results the pro engineers are achieving. The other tools I’ve mentioned make life just a little bit easier but you’ll get a similar result with your DAW’s bundled plugins.

Remember: great-sounding vocals aren’t something you can rush. You need to take your time to make the artist feel comfortable; take your time listening and doing the best job that you can! More often than not, the vocals are the focus, it’s where the humanity and emotion of the song will mostly come through. A good lead vocal comp will show the artist in the best light and make them feel great, and this just takes time and care. Once you’ve got a great lead vocal, making sure the BVs support the lead and help tell that story is the goal. Time, care and Melodyne will get you there!

FAMILIARISATION TOUR

Using the DM3 for the first time requires less familiarisation than most and there are plenty of helpful features. This is when you notice the mixer’s TF series lineage: the one-knob EQ and compression functions, the presets and scenes. For those not familiar with the TF series, the consoles have a library with preset scenes for typical music/speech applications that give new users (or operators in a hurry) a helpful pre-patched starting point. Naming the channels on the DM3 with the virtual keyboard is simple and includes the choice of different languages. Channel colours are limited to eight and there’s not many instrument icons, there’s no tuba icon, for example (yes, I mixed a band with a tuba, and felt mildly dudded).

The 8+1 physical channels have the usual Sel/Cue/Mute buttons, while everything else you need is on the screen. Touching a box or an edit icon takes you there and activates the one-touch knob ready to tweak the level or main parameter. Touching again opens up more functions and options. Channel processing starts with a four-band parametric EQ, with separate HPF. The RTA integrated into the EQ section is not overbearing like some, and can be switched off, or switched to display a piano keyboard for reference. There are two dynamics processors, nominally set as a gate and as a compressor. GainFinder is a visual gain-setting helper. Each channel can be delayed. A multi-band compressor and de-esser are noticeable by their absence, perhaps being saved for an update.

ONE TOUCH, ONE ENCODER

When I first met Yamaha’s QuickPro Presets and One-Touch EQ/Compressor functions on the TF Series I thought they were gimmicky but after using them I came to appreciate how they are another control option rather than simply a shortcut for newbs. A library of QuickPro Presets gives typical EQ/comp settings for a long list of commonly used mics and DIs. The One-Knob function continuously varies how much of the suggested EQ or Comp settings get applied to the channel. The One-Touch EQ preset for an SM58, for instance, gives you increasing amounts of a simultaneous boost to the high-mid vocal range and reduction in the low-mids. The One-Touch compressor simultaneously adjusts the threshold; increases gain reduction; and raises the channel output level as more is applied. The settings make sense and it’s fun to audition the presets but in use I found I was often only wanting to change one parameter at a time and reverting to manual drive.

EFFECTS & SENDS

Notably, there are only two effects. That’s okay by me – a reverb and a delay is enough for a 16-channel show. The effects include three versions each of Yamaha’s REV HD and REV R3 reverbs from the TF series, among a total of 18 different types of effects to choose from.

If the on-board effects aren’t enough or you’d like to use your fave recording plugins instead you can download VST Rack Elements that hosts Send/Return VST plugins.

The boxes in the Send section flip the faders to the six mix buses or matrix send levels. The mix busses (generally foldback) can also be accessed via Send From boxes on the input channels and the send level adjusted by the knob. The Mix Out channels get the same processing as the input channels less the separate HPF, but with the addition of a 31-band GEQ (limited to 15 bands at one time) that can be flipped onto the faders or controlled from the screen.

There are six fader layers on the DM3. The first two are the 16 input channels, the third is FX/MON channels, the fourth is the outputs. Layers 5 & 6 are custom layers you can populate on the fly and I used these a lot. Throwing the channels you need to ride onto one layer saves a lot of thinking. Below the Layer buttons reside six user-defined buttons that allow control of your choice of functions including Mute Groups, FX mutes, tap for delay and meter peak-hold. If the DM3 is used as a recording interface the user-defined buttons can control the transport functions of the host DAW.

GOTH BANJO MUSIC HALL

I’ve been using the DM3 for around three months by the time I put pen to paper for this review. It’s been a handy. It can be anything from a nifty problem-solver to a fully-featured FOH or foldback mixer. The first time I took it out, I was using a 32-channel analogue system for the 2023 Maldon Gothic Festival and over the weekend the DM3 got used for recording the live acts, playing backing tracks, BGM between acts, and as an on-stage sub-mixer for a magic act.

The Guildford Banjo Jamboree is one my favourite annual events and I wanted to use the DM3 as FOH mixer in the Public Hall for the Friday bush dance and the main Saturday concerts. Starting from a Preset scene and tweaking it to suit the event was an easy way to get going and I was soon running all 16 channels in the house plus five sends of foldback. Running the PA subs from a mix out gave me good control of EQ and dynamics, plus delay, and saved packing a separate crossover.

The DM3 looked a little out of place at the traditional bush dance. People kept asking if I could watch TV on it, but I was happy with the sound in the room and found mixing quite enjoyable. The channel and output EQ options are precise and you get what you expect from them. The comps are normal – they do what they’re asked – perhaps without the character of favourite analogue devices (the high-end Rivage range has simulations of those) but they work. The FX are limited in number but they’re excellent quality. The physical faders are real but not very smooth, not like analogue faders. It’s more of a digital console issue rather than Yamaha’s, and could be a cost thing, but they’re noticeably notchy.

I also supply a system for the old gold-rush era Guildford Music Hall across the road. The guys mixing in there over the weekend noticed that I’d ‘gone digital’ in the main room. They had no complaints about the sound but were mighty relieved that I’d supplied a 16-channel analogue desk for them to use in the Music Hall. It might be different for the new kids but there are a lot of amateur/casual/senior mixers out there who freeze when they see digital consoles… even a nice friendly one like the DM3.

I make a simple recording of the concerts every year, usually onto a Zoom recorder from the mixing desk. Recording onto the DM3’s USB drive was easier, the recording/naming routine is foolproof, and the recorded sound quality is very good, although I’d prefer the recordings to be pre master bus EQ, or switchable, to avoid printing the house EQ. Something for a firmware update perhaps.

GOING PLACES

The Whole Loddon Love Tour saw the DM3 on a run of shows in flood-affected towns in the Loddon region of Victoria. Some were outdoor shows but others were small pubs with no dedicated FOH mix position. The 24-channel analogue desk and effects/drive racks I otherwise would have taken would have been hard to squeeze in. These small rooms made a good case for an iPad mixer but the DM3 was a better solution for me – it’s not wireless but it’s more precise to operate and it will fit in any space.

An outdoor school concert I do every year is a good test of how a digital console performs in daylight. One year I recall having children holding cardboard sheets above the console to try and create some shade because I couldn’t see the screen at all. That was under a gazebo too, it’s glary out here. The DM3 passed the test, with the brightness up full, I could see the screen quite well, and it’s got dark/light background screen modes with separate intensities for screen and panel, plus two custom banks to store your preferences.

A seven-piece cover band at the local sports club was probably about as big as you could fit into the 16 inputs but did demonstrate how capable the DM3 is with heavily processed pop bands, where I think digital consoles do their best work. The channels were all loaded up with no signs of protest and it remained easy enough to operate with a band that needed lots of FX and fader-riding. Clear displays of everything going on helps, the Custom Fader layers minimise fader flying.

GROUP THERAPY

There are a few things the DM3 doesn’t have. No DCAs – although you can make groups from the omni mix outs and route them back to the main outs, and there’s also a Channel Link page to create groups of linked channels that can have parameters changed simultaneously. There’s no talkback mic so you use up a channel to talk to stage. No auto mic mixer for corporate events. No multiband comps or de-esser. None of that really matters, it’s not competing on quantity of functionality, it’s the functionality of the package that’s important here and by that measure it punches well beyond its weight.

Apps include the offline DM3 Editor, a good way to learn the desk before you use it or to pre-plan shows, that can be saved and loaded when needed. DM3 StageMix is the app for iPad-only remote control. MonitorMix is available for Android as well.

All Yamaha gear seems inherently reliable, they make a lot of machines you don’t want to fail in the middle of a high speed corner (or equivalent) so they make them to work and last. Yamaha pioneered the development of digital consoles and while I still occasionally see tech riders specifying ‘no LS9s’ its current range from the DM3 up, are all leading-edge premium consoles.

SLIDING INTO THE DM

Given the quality, both physical and fidelity, the Yamaha DM3 seems like a bargain to me. There are cheaper options with similar channel counts, especially iPad-control systems, but the combination of Yamaha quality in a modern compact console, with an excellent touchscreen and real faders is a level above.

Potential customers for the DM3 include small-medium sized bands/duos/DJs, streamers and home/project recording studios. Portable PAs, where size and weight are important. Venues that use less than 16 channels looking for a plug ’n’ play replacement for their ageing analogue mixer, or any sort of corporate/educational setting. I want one.