Modern Mastering CM

> make music now / modern mastering

MODERN MASTERING EQUALISING DYNAMICS STEREO IMAGE MULTIBAND LOUDNESS

42 / COMPUTER MUSIC / January 2015

modern mastering / make music now <

DOWNLOAD Grab all the videos and get the files to do it yourself at vault.computermusic.co.uk

Finish off your tracks with the sound of now! We’ll take you inside a contemporary mastering studio and show you how to apply 2015’s sonic polish using only software With modern music-makers now taking on the multiple roles once distributed among several engineers, it’s no secret that the once-mysterious field of audio mastering – the final stage of the music production process – is now accessible to all. While many post-processing studios still use rooms full of expensive gear, the availability of mastering software (such as Ozone 6, the latest version of iZotope’s flagship mastering package, reviewed on p94) means that everyone can have access to pro-quality mastering tools. However, with the mix engineer attempting more and more to take on the duties of the mastering engineer – a technical and often delicate craft that dedicated professionals spend many years perfecting – there’s now a new generation of producers that simply slap plugins on the master bus and squeeze the life out of their music in the hunt for professional loudness and sheen. It doesn’t have to be that way, though – and Computer Music is here to ease the pain with our huge Modern Mastering guide!

We’re going to give you the expert insight that you need to tackle this critical final stage of the music-making timeline like a pro. You’ll sit in on a virtual ‘attended session’ and shadow professional mastering engineer and lecturer John Paul Braddock as he masters a track in an exclusive Mastering Masterclass video session. Then follow along with our hands-on tutorials to learn the workflow and techniques that’ll sharpen your own mastering skills. Discover how to use modern software tools to master music at home, bring out the best in a well-mixed track, and help it compete against other commercial records. Even if you’d rather leave your tracks to be mastered by a professional, you’ll gain a truer understanding of how a mastering engineer will approach your track, and thus be able to provide him or her with the best mix possible. These techniques can also quite easily be applied during sound design, production or mixing sessions. And with video for every tutorial, plus tutorial files to help you along, your greatest masters ever are about to become reality…

January 2015 / COMPUTER MUSIC / 43


Mastering fundamentals songs and albums must compete with other professional records. A skilled mastering engineer has the equipment, experience and ears to help a mix sit beside other commercial releases, adding that final 5-10% of polish. Collections of individual tracks – whether destined for an EP, album or compilation – can each sound tonally and dynamically separate from each other. The mastering engineer will ensure the final collection of songs all sit together as a single cohesive product.

Before you start piling the plugins onto your latest project’s master channel, we should define exactly what ‘mastering’ is. Broadly put, it’s both the final stage in the creative production of a track or album, and the initial stage of its manufacture (where a physical product is concerned, such as a CD or vinyl) or output (ie, WAV/MP3 delivery to a download portal). In the early days of mass-produced music, it was the job of a skilled individual to physically transfer a final mix from tape to the ‘master’ vinyl record, which would then be sent for duplication. As record labels realised that skilfully applied processing could make a big difference in the perceived quality of music, mastering engineers were given more creative freedom to enhance mixes with equalisation and compression techniques. In today’s digital age, the mastering engineer is still essentially the middle-man between the finished mix and the consumer, correcting any ‘errors’ in the mix, making enhancements where needed, and editing/ submitting the final ‘master’ file for release – all without compromising the artistic vision of the original artist or creatively altering the mix.

Mastering masterclass

Why master? So why doesn’t the mix engineer simply take on the task of mastering too? A mastered track should sound as good as possible on as many playback systems as possible, achieving a professional and consistent sound whether listened to on a car stereo, club soundsystem, cheap earphones, television, mobile phone, and so on. A producer or mix engineer has likely spent countless hours creatively blending multiple elements together to craft the final mix, and in doing so has ‘overlistened’ to the track in the same studio, which may not be the ideal listening environment. The mastering engineer is a final, experienced pair of ears that can objectively listen to the track, correct errors

JP’s mastering philosophy is that a track should be balanced in all parts of ‘the cube’

introduced by an imperfect mixing studio, and transparently sweeten a piece of music further. Fans and listeners are used to hearing modern music coated with a professionally-mastered ‘sheen’ – sparkling highs, deep bass, consistent frequency spread and dynamic balance. Today’s

“A skilled mastering engineer has the equipment, experience and ears to help a mix sit beside other commercial releases”

Monitoring and acoustics Assuming the mix itself is ‘good’, a dense stereo mix requires quite delicate processing, as extreme settings could lead to unnatural side effects, interfering with its overall integrity or vision. With this in mind, we’ve used very subtle settings in our tutorials, so to appreciate these fine details, watch the videos and listen through the Tutorial Files on the highest quality pair of monitors or headphones you can. For a pro mastering engineer to subtly improve a mix, and help it sound consistent across a broad range of speakers and output formats, they require highly accurate monitor speakers, situated in a neutral acoustic 44 / COMPUTER MUSIC / January 2015

environment. Great monitoring allows you to hear exactly what’s going on across the full 20Hz-20kHz frequency range, helping you spot mix errors and inconsistencies. A suitable studio space and acoustic treatment stops environmental factors from interfering with the sound emerging from the speakers. If you’re looking to master music yourself, but your own studio or production environment features less-than-ideal monitoring or acoustics, we’d suggest addressing the issue as soon as possible. Coincidentally, sprucing up your acoustic knowledge is as simple as turning a few pages to our Studio Acoustics feature on page 67.

John Paul “JP” Braddock of Formation Audio is a mastering engineer with over 25 years working experience in the music industry. Not only does he make his living from mastering, but he’s also an experienced lecturer and teacher, making him the ideal man to give readers first-hand insight into the mastering professional’s craft. When introducing audio students to the concept of mastering, John Paul begins by outlining his take on the ‘cube’ theory. “Imagine a three-dimensional cube that represents the ideal ‘space’ a stereo mix should fill,” he explains. “The height of the box equates to frequency, front to back relates to dynamic range (-inf to 0dBFS in the digital realm), and the left-to-right plane is the stereo field. In the first diagram, a track’s overly dynamic low-mid range pokes out of the bottom-front of the cube, low-frequency stereo elements reach too far out of the sides, and the track’s overall interrelationships are inconsistent. Mastering the track correctly can help it fill all three dimensions of the cube, as in the lower diagram, giving us a tonally and dynamically even mix that will translate well across all playback systems – without excessive brickwall limiting or heavy-handed processing” For our seven exclusive Mastering Masterclass videos, we take you on a trip to John Paul’s mastering facility for an in-depth session. After providing an overview of the mastering engineer’s role and a brief tour of the mastering hardware he uses on a daily basis, you’ll see him get hands-on with the final mix from 209’s House Track-Builder feature, helping it fill the aforementioned virtual ‘cube’ while improving the mix’s overall tone, weight and loudness characteristics. Let’s head into the studio…

WARNING! HEADPHONES MUST BE WORN The difference that mastering techniques make can at times be subtle. To fully appreciate them, we strongly advise you to listen critically and compare the before/after WAVs (in the Tutorial Files folder) in as good a listening environment as you have access to. That means using monitor speakers and/or decent quality headphones!

MONITORS MANDATORY

modern m mastering / make music now <

1. Introduction and overview

“Hopefully, by the end of this, you’ll have a good overview of how you can make your tracks better”

In this first video, John Paul outlines the technical role of the mastering engineer in relation to the other creative stages of the production and mixing timeline: to make a track sound balanced across a wide range of playback systems, and ‘fit for purpose’ without creatively altering the artist’s original vision.


> make music now / modern mastering 2. Load in, levelling and equal loudness Before attempting to master a track, John Paul is adamant that you should be aware of the human perception of ‘loudness’, and how level imbalance between tracks and processes can influence our critical decision-making at the mastering stage. (head to page 51 for more info, or Google ‘Fletcher-Munson’ to find out more about human loudness perception). “As ‘louder’ sounds perceptively ‘better’, then any time we apply a process, if that process is making the audio louder, we’re hearing the process plus the difference in perceived loudness. Equally the opposite way around: if we compress something, but don’t use make-up gain, the compressed signal sounds weaker when A/B-ed with the uncompressed signal, even though in reality it could sound better.” This is especially important when applying EQ boosts or cuts. “If we apply one (EQ) aspect, we also apply another aspect to the negative somewhere else. Every action has an effect. If we apply treble, we lose bass; if we apply midrange, we lose bass and treble; if we cut the mids out, then in effect we’re creating bass and treble, once we’ve rebalanced at equal loudness. It’s really important that, at the mastering stage, we’re aware of the fact that if we apply a boost or cut, we’re also applying a gain change. We’ll constantly be A/B-ing and trying to re-level every time we apply a process, to make sure we’re actually changing the mix in a way we want to, and aren’t being fooled by our ears. Our ears are our best tool, but equally our worst enemy when it comes to the loudness issue.”

See JP practically demonstrate how loudness imbalance can fool the ear, and learn how to perform a ‘sum difference’ test to discover what your plugins may be doing to your signals.

EQ will lead to gain change, which must be compensated for with re-levelling

3. Mastering hardware and transfer path overview Take a tour of the Formation Audio mastering studio, and have a good look at the choice selection of hardware and software processors that can be inserted at different points in the signal chain – otherwise known as the ‘transfer path’. 46 / COMPUTER MUSIC / January 2015

modern m mastering / make music now <

JP tackles our house track head-on in this exclusive Mastering Masterclass

Selected kit list Apple Mac Pro Windows 7 Professional Magix Samplitude Pro X RME AES-32 TC Electronic PowerCore Universal Audio UAD Quad UAD-2 Manley Massive Passive EQ, Pultec Pro EQ, Cambridge EQ Crookwood M3 Mastering console Lynx Hilo A/D D/A Transfer converter

Chandler Curve Bender EQ (EMI TG12345) mastering version Manley Mini Massive EQ Manley Vari Mu (mastering version with HPF option) TC Electronic System 6000 MKII Reverb & Mastering Pack Crookwood Monitor Controller Benchmark DAC1 ATC SCM25Pro monitors



4. Corrective processing with multiband compression To show you exactly how a professional mastering engineer can improve a track you might be familiar with, we asked John Paul Braddock to master the final stereo mix from 209’s ‘House Track-Builder’ feature – after we removed our own master bus processing, of course! Braddock begins the mastering session by loading the mix into Magix Samplitude Pro X and auditioning small sections of the mix. “I’m scanning across the track, to get a sample of how it feels.” As we previously mentioned, John Paul imagines a mix within a virtual ‘cube’, which relates to the ideal 3D space a track should occupy. “We can hear the mix is too ‘lumpy’: the bass end is too weighty and is pushing out of the box. Equally, the top end lacks detail – it’s not necessarily too dull, but it doesn’t contain enough transient detail and isn’t ‘spiky’ enough. We want to make the mix fit nicely within the cube, then lift the top end so the track occupies more of the space. If it occupies more of the space, it will sound more balanced.” “Because this bass end is a bit ‘floppy’, we need to apply some multiband compression, to tighten it up and pull it together. However, we’re not going to apply [multiple bands] across the whole track, as that would EQ the entire track and turn [frequency ranges] up and down in varying levels.” Instead, John Paul focuses in on the track’s bass end and compresses that area in isolation.

After compressing specific frequency bands, John Paul turns his attention to the track’s stereo components

Here, John Paul begins to master the stereo mix from 209’s House TrackBuilder feature.

5. Mid/side, tonal and dynamic enhancement

See our pro mastering engineer sweeten the individual mono and stereo components of the mix.

Once the problematic low end is controlled, Braddock shifts the track’s overall tone by equalising the mono (mid) and stereo (side) information of the track separately. “The top end needs a lift, but more to the mid aspects than the side. Currently we’ve got a ‘scoop’ to the treble: rather than a linear line of

The Chandler EMI TG12345 Curve Bender is a powerful hardware equaliser that Braddock uses to full effect


treble across the top of the ‘cube’, we’ve got a dipping effect”. An EQ unit is used to raise the mono information’s treble and dip the side signal’s treble. Subtle master compression then “pulls” the track together. Throughout the session, John Paul always keeps a copy of the original unprocessed mix, labelled the ‘reference’, side-by-side with the version he’s processing, named the ‘source’, or the beginning of his transfer path. “At all points, we can A/B between these two versions, to see if we’re changing the track in the exact way that we want to,” points out JP.

Voxengo’s free MSED plugin is used to separate our track into its mono and stereo components

modern mastering / make music now < 6. Soft clipping and limiting With the track corrected and enhanced, Braddock now uses a brickwall limiter to reduce the peaks and overall dynamic range, and help the track reach a similar level to other commercial releases without inducing distortion. However, he’s keen to emphasise the importance of all the other previous processing stages. “We’ve tonally and dynamically balanced the track, and filled our ‘cube’. Because of this, we can now apply limiting to it in a positive sense, so we’re just skimming off the peaks – achieving perceived loudness because we’ve tonally and dynamically corrected the track, as opposed to

“The transients have been reduced, but there’s still detail in the mastered waveform” applying limiting just for loudness. We can now take the peak information off in an even way, without it sounding too ‘affected’.”

Using the BL2 limiter from his TC Electronic System 6000 MkII, John Paul uses an initial stage of soft-clipping to remove the track’s highest peaks, before applying brickwall limiting to raise the track’s perceived level, being careful to A/B with the unprocessed track at equal volume. Check out the waveforms of the unprocessed and limited track below/left. Although peak reduction has been applied during mastering, the waveform hasn’t completely been chopped off – the transients have been reduced, but there’s still detail in the mastered waveform.

See John Paul use soft-clipping to remove the track’s highest peaks before applying brickwall limiting to raise the perceived level – all the while A/B-ing it with the unprocessed track at equal volume. The original (top) and mastered (bottom) waveforms – spot the difference!

7. Summary To finish, John Paul compares his finished master with the original unprocessed track. “Once we’ve mastered a track, we need to listen back to what we’ve done. When we’re using DAWs, and technology in general, it’s very easy to assume that our technology has done what it said it would do. “Because we’re the final stage of the process, we need to listen through the track and check that it sounds exactly as we want it to sound. Have we enhanced all the aspects we want to enhance? Have we made it better? If we can’t answer ‘yes’ to these questions, then we need to go all the way back to the beginning and start again.” To check out Braddock’s finished master, head to the Tutorial Files » Modern Mastering » Mastering Masterclass with Formation Audio folder. The first WAV file is the original unprocessed mix, and the second is the

Formation Audio master, turned down to match the unprocessed file. By A/B-ing between the two versions, you’ll hear how the track’s bass end has been tightened, and the overall track is more balanced in the high-mid and treble areas. The third file is the same master as the second but has been turned up to a ‘commercial’ level – a loudness that has been achieved through several stages of careful, reasoned processing. “It’s really important that we’re aware that limiting is a ‘post-process’ – until we’ve got the dynamic and tonal balance correct, the limiter won’t work effectively, if we’re trying to get some form of volume out of the track, to make it a ‘commercial’ level.” You can check out more of JP’s mastering resources and articles online at www.formationaudio.co.uk January 2015 / COMPUTER MUSIC / 49


Mixing vs mastering The majority of computer-based musicians are stuck in the composition, production and mixing mindsets. Our projects contain multiple tracks of building blocks that combine to form large multitrack mixes, so we’re used to piling on plugins, applying heavy-handed settings, and having control over each element of a mix. This approach is counter-intuitive when approaching mastering. The mastering engineer is really just a final, impartial pair of ears. To master music yourself, it’s important to realise that the creative mixing role has already been completed. Your task is now a technical one – to correct, sweeten and enhance a mix. That’s it. Disengage from the individual aspects of a track, and view it as a finished body of work that can be slightly enhanced, not proactively altered. Objectivity is one reason why it’s extremely difficult to master music that you’ve spent hours

> Step by step

producing and mixing yourself. Most artists find it tough to ‘emotionally detach’ themselves from their own song and make critical processing decisions without slipping back into ‘mix engineer mode’. We’ve all tried to create a quick ‘test’ master – to hear how a track sounds under the strain of master processing, or to DJ the track alongside commercially-mastered tracks – but generally, a separate mastering individual will perform a better job. They’ll also have the advantage of hearing your music in a different (and ideally better) environment, exposing any mix errors that your studio may disguise. If you want to master with true objectivity, source unmastered mixes from other artists and friends, or, if you’re desperate to master your own tracks, sit your projects aside and forget about them for weeks – or even months – before approaching them with your ‘mastering hat’ on.

Making a test master while mixing is fine, but the mastering mindset is more than just throwing on plugins

1. Preparing a multitrack mix for mastering

TUTORIAL

FILES

1

Here’s a typical ‘mixed’ project comprising many audio and MIDI channels. You could apply processing directly on the master channel, but you might be tempted to adjust individual mix elements, too, switching you back into the mindset of the mix engineer. Instead, render your track as a single stereo file, so that it can be tackled from the mastering engineer’s perspective.

2

Set your DAW’s cycle markers around the entire track – including a few seconds of silence at the beginning and end – ensuring you capture any extra track elements such as reverb tails or leadin effects. This is especially important for analogue mixes or restoration projects, as isolated noise at the beginning or end can be used as a ‘noise profile’ for noise removal plugins or restorative processes.

3

If the track contains any global volume fade-ins or fade-outs, remove them before rendering the final mix, as they’ll sound unnatural when dynamics processing is applied to them. If in doubt, export a second ‘reference’ version with fades applied, so that you or the mastering engineer can replicate them at the end of the mastering process.

POWER TIP

>How many bits?

4

Remove any master buss processing before exporting the final mix, as it could impair the mastering engineer’s ability to apply their own processing to the track. If a particular master effect is adding a quality that you like, simply make a note of it so that it can be replicated during mastering, if desired. Get the highest peaks hitting around -3dB to -6dB, to prevent clipping and leave headroom.


5

Render the mix at the same sample rate as the project, and at 24- or 32-bits to ensure that you don’t lose any quality or resolution in the final stereo file. It’s important to maintain audio fidelity at every stage of the production process, so be absolutely sure that you aren’t inadvertently lowering the quality of the track at this final stage.

We asked John Paul Braddock of Formation Audio if there’s any benefit to rendering out your 24-bit mix project as a 32-bit stereo file. “If you’re aware of your project’s output level, and not clipping the master bus, you don’t gain anything from rendering your project at 32-bit float. In principle, you’re monitoring through a 24-bit converter, so what you’re hearing is the 32-bit signal truncated and dithered to 24-bit anyway! If your master bus is clipping, then 32-bit will save you, as the extra headroom will keep the overs as rendered. It’s best to keep it within the correct dynamic range for a mix bounce. For an actual mix session, 32-bit can be useful – if you freeze a track, you know it won’t clip.

modern mastering / make music now < > Step by step

1

2. Referencing, levelling and equal loudness

Here we’ve loaded Session Mix.wav, a short section of an unmastered stereo file, into a fresh project. By comparing this mix to a professionally mastered reference track that’s proven to sound good on a wide range of playback systems, we’ll be able to assess exactly how ours can be improved. We load Reference.wav onto a second audio channel and mute it.

2

Our professionally mastered reference track has been limited to 0dBFS, making it far louder than our unprocessed session mix. As we’ve seen in our Mastering Masterclass tutorials, a louder track sounds subjectively ‘better’; to counteract this phenomenon, we balance our tracks’ levels to a calibrated reference point, so that there’s no gain difference to influence our objectivity.

3

RMS metering gives an accurate indication of average level over time – far more closely related to our perception of loudness than peak metering. We load an instance of iZotope Ozone 6 on each track and watch their RMS readouts as the tracks play through, with both muted so as to keep things entirely ‘visual’. The session mix peaks at -16 to -18dB RMS, while the reference reaches around -6dB RMS.

POWER TIP

>On the meter

4

To match the reference track’s RMS value to our session mix, we turn the region’s gain down directly on the clip itself, leaving the channel faders free for later adjustments. Turning the reference down to -11dB brings its RMS value more in line with the session mix.

5

RMS average metering – as opposed to normal peak metering – can help you balance the levels of tracks within a mastering project, whether it’s a processed and unprocessed version of the same track, a session mix against a commercial reference, or a selection of album tracks across a project. Once levelled with metering, turn off the meters and use your ears to judge exactly which track is louder. Our hearing is better at comparing levels than any meter, so trust your instincts and fine-tune the last dB or two by ear.

It’s important to use your ears and fine tune the gain difference between the two tracks, too. Our reference still sounds subjectively a little louder than the unmastered track, so we turn it down to -12dB. Now we can A/B compare our mix and the reference track at equal volume, without the level difference influencing our perception of tone and dynamics.

The Fletcher-Munson ‘equal loudness curves’ are the results of a series of audio experiments. Subjects were played test tones at a particular frequency, and told to note when these tones sounded as loud as a set reference frequency. The results prove that not all frequencies are heard equally – our ears are far more sensitive to midrange than to low and high frequencies, and what’s more, the ‘response curve’ of our ears changes at higher listening levels, with bass becoming easier to hear, for example. This is why some hi-fi’s have a ‘loudness’ or ‘bass boost’ button, to bring up the bass level when listening at low volumes, giving the bass-heavy effect of high-volume listening, thus increasing the impression of ‘loudness’. The dips in the graph at 2-5kHz show that those frequencies must be at a far lower level than bass and treble

to be perceived at the same loudness. To illustrate this, load a well-mixed, vocal-led commercial track in your DAW, then turn your headphones or monitors right down. It might be difficult to perceive the deepest bass or highest treble, but the midrange frequencies and highmid of the vocal will likely still cut through. Now, turn the volume up gradually, and listen to how the track’s bass weight increases and the treble cuts through more. It’s the exact same track, but it sounds ‘better’ the louder it becomes. Mastering engineers will often use dedicated average level metering systems to keep levels in check, such as RMS, K-System (invented by mastering guru Bob Katz), Dorrough, and broadcast formats outlined by the International Telecommunication Union (ITU) and European Broadcast Union (EBU).

Sound pressure level (dB)

Equal loudness across the spectrum 130 120 110 100 90 80 70 60 50 40 30 20 10 0 -10 10

100

1000

10k

100k

These curves show how loud various frequencies must be played for us to perceive them at equal loudness



Fixing mix problems at the mastering stage Before you begin applying any overall enhancement to a mix, you must correct any errors or mistakes present in the source material. If you have the opportunity to return to the mix, then this is always best, but sometimes this is not possible – when mastering a track you have been supplied, for instance. There may be glitches, pops, or clicks due to poor editing or caused by software – these small details can go unnoticed, but they’ll be magnified with later broad enhancement, so consider checking the mix with headphones to pinpoint these small details. Noise removal plugins can help remove these sort of errors, but it’s always advisable to correct these within the original mix first. ‘Honking’ resonances or frequency clashes can also be problematic. Listen out for areas where instruments overlap and particular frequencies get too loud, or grungy-sounding bunches of adjacent frequencies are caused by harmonics stacking near to each other. Tools such as multiband compressors or dynamic EQ allow you to focus on these very specific areas and correct them in isolation. A poor mixing environment or inexperienced engineering can also lead to dynamic issues within a track. Overly sharp transients can poke above a mix, or certain areas may appear too dynamic, or ‘flappy’. Again, one or two gentle bands of careful multiband compression can smooth out problem areas. Be careful with regular static EQ, as you’ll likely affect other areas of the mix too, so leave this for the later ‘sweetening’ stage. Listen to the track’s separate stereo components in isolation – the left channel, the right channel, mid signal and side signal. Understand that you can monitor and process these parts individually if needs be – for example, if a hi-hat on the left side of the mix is too piercing, you can apply processing to the left side of the mix, leaving the balance of elements on the right side unaffected. We’ll be addressing the topic of stereo processing in a few pages’ time. To reiterate, while pinpoint master processing and correction can go a long way in tidying up issues, remember that it’s always easier to correct errors at the source within the original mix project.


> Step by step

3. Cleaning up a mix at the mastering stage

1

Before you add enhancement or ‘sweetening’ to a track, you need to fix any problems introduced at the mix stage. A mix engineer’s monitoring environment might conceal frequency clashes and recording errors, and it’s the mastering engineer’s job to clean these up first, so that further processing doesn’t exacerbate them. Import Clean Up Mix.wav onto a new audio track in your DAW and loop it.

3

A narrow band of multiband compression could solve this problem, but instead we’ll turn to dynamic EQ, which combines the frequencyshaping flexibility of EQ with the leveldependent triggering of compression. Load iZotope’s Ozone 6 Dynamic EQ plugin onto the track. We can see the highend resonance in the analyser display.

2

We briefly play through different sections of the track to help us decide what our mix needs. It’s important not to ‘overlisten’ and let your ears get used to how the track sounds. The mix is relatively dull and weighty, but an overly-harsh hi-hat occurs every so often. We need to control this resonant high frequency, so that broad tonal adjustments later don’t make the harshness even worse.

4

Set Band 4’s Frequency to 15.5kHz – the exact point of the resonant peak – then select the Peak Bell curve. Set a Gain of -25dB and narrow the Q width to 4.00. Now, pull down the Threshold – a setting of -28dB causes the resonant spike to trigger the EQ’s notch. Shorten the band’s Release to 100ms, returning the EQ notch to unity more quickly.

POWER TIP

>Bands-u-need

5

Make sure no other part of the track exceeds the threshold at this frequency, as that would trigger the EQ band’s gain reduction, notching the area unnecessarily. Compare the track before and after EQ to hear how we’ve efficiently tamed that high frequency peak without affecting any other mix elements.

When you load up a multiband plugin, its bands are usually all active, showing off just how many features, bells and whistles are available. Remember that, just because a multiband plugin has lots of bands, that doesn’t mean you have to use them. Overzealous multiband processing will unnaturally and uncontrollably shift a signal’s tone and dynamics around. Instead, begin by switching off each band’s processing, then carefully isolate a problem area with a single band or two – ensuring the processing is being triggered only by the desired mix issue.

> make music now / modern mastering > Step by step

1

Import CM200 Track.wav onto an audio track in your DAW, then load CM209 Reference.wav onto a second track and mute it. Briefly audition each one in isolation, and note how the first track’s overall tone appears to be slightly weighted towards the low-mid and treble areas, while lacking high-mid presence, in comparison to the reference.

3

To reduce the excess of treble, select the EQ band 7, which is set to a Baxandall High Shelf. -3.0dB Gain at 9kHz gently shelves down the track’s high frequencies. To lift the bass end up slightly, set band 1’s Baxandall Low Shelf to 0.5dB at 150Hz. With a well mixed track, broad shelves like these are often all you need to gently shift the mix into the desired frequency area.

> Step by step

1

Linear-phase vs. ‘regular’ EQ

4. Broad tonal tilting with mastering EQ

2

Let’s gently shift the overall tone of our track (CM200) towards that of the reference track, which has a better tonal balance. We load iZotope’s Ozone 6 Equaliser plugin onto the first track, click band 3 on the display, and change the filter type to Analog Low Shelf. Set the band’s Gain to -1dB at a Frequency of 500Hz with a gentle Q of 3.0.

4

Bypass the EQ and you’ll hear that the unprocessed signal is now louder than the newly-EQed one, so increase the Output Gain to +2.0dB. Not only have we reduced the track’s low mid and treble, but we’ve also added emphasis to the highmid area, bringing its tone more in line with the reference track.

When we boost or cut a certain frequency using a ‘regular’ (or ‘minimum phase’) EQ, the phase of that frequency area is shifted ever so slightly, delaying the timing of that frequency and possibly resulting in transient ‘smearing’ and colouration – an effect which a mastering engineer may wish to avoid over a full mix. In this instance, a linear-phase EQ can be used, which will keep the timing/phase relationships consistent across all frequencies. There are benefits and drawbacks to each. A linear phase EQ offers a cleaner processing alternative in terms of phase but can cause audible ‘pre-ringing’, especially in the low frequencies – plus it can introduce a higher CPU hit and added latency. In addition, the phase distortion caused by a traditional minimum-phase EQ can also contribute to the character, adding ‘mojo’ to a sterile mix. Overall, we’d advise you to reach for a linear phase EQ when absolute cleanliness and transparency is required when applying broad tonal adjustments, but consider a more ‘characterful’ minimum phase EQ when a mix already seems a little cold and clinical, or you need to make tight cuts (which can exacerbate pre-ringing in linear phase EQs). Linear phase EQ is also desirable for left/right or mid/side stereo processing, to minimise phasing artifacts between the two channels.

5. Master ‘feathering’ EQ

By distributing EQ boosts and cuts over several stages, you can minimise resonance and phase shifting when mastering. Here, we’ve loaded Feather Mix.wav into our DAW. The track could do with a mid-range EQ boost to even out its tone. FabFilter’s Pro-Q 2 is used to apply a +5dB bell boost at 450Hz, with a Q width of 1.00.


2

After copying the settings over to the ‘B’ state, we halve the boost to +2.5dB. Now we add two smaller boosts either side of the first, and at half the gain level of the centre boost – so +1.25dB bell boosts at 400Hz and 500Hz (both at 1.00 Q). We then use the A/B button to switch between the single larger boost and the three smaller ones.

3

The three smaller bell boosts are affecting the same frequency range as the first single boost, but the EQ processing is divided over three gentler stages. This softens the effect, resulting in less harshness and focus than you’d get with one large boost, while also minimising the phase shifts that can be introduced by extreme EQ boosting.

modern mastering / make music now <

Stereo processing Four elements of a stereo mix can be treated separately – the left channel, the right channel, the mid signal, and the side signal. To the uninitiated, mid/side (or M/S) processing can seem complex, but it’s actually rather straightforward. A stereo signal’s ‘mid’ information is simply the information that can be found identically in both the left and right channels – or everything that’s mono. Its ‘side’ information is all the information that isn’t equal in the left and right channels, so anything that’s stereo. Panned instruments, stereo reverb, wide effects and stereo sources are present in the side signal, whereas core mono mix elements – drums, bass, lead vocals and instruments – will generally occupy the mid signal. Sounds that are entirely out of phase in a mix will always feature in the side signal, but not in the mid, and panned instruments will usually feature in both the mid

> Step by step

1

You can visualise your mid and side signals using Flux’s free StereoTool plugin from fluxhome.com

and side signals to some degree. These aren’t hard and fast rules, as every mix is different, so we’d advise you to strap a simple M/S tool across

a track’s channel and isolate the mid and side signals for yourself. Voxengo’s MSED (free from bit.ly/MSEDplugin) or Brainworx’s bx_solo (bit.ly/bx_solo) are two great free options. By altering or processing the mid and side independently, a mix’s stereo balance and width can be corrected and/or enhanced. If a track needs more or less ‘width’, the side signal’s gain can simply be turned up or down. Perception of width can be applied with simple EQ shelves or boosts to the side signal, and a particular frequency range can be ‘evened out’ across the speakers, as we’ve seen in our earlier Mastering Masterclass tutorials. If a track’s stereo information is too dynamic, compression can increase its solidity – again, adding perceived stereo presence. Many modern plugins now feature an inbuilt left/right or mid/side mode, enabling simple M/S processing without any extra routing.

6. Mid/side master processing

Let’s look at how we can monitor and treat the mid elements (the mono information) and the side elements (just the stereo information) of a stereo mix in isolation. Load up MS Track.wav, found in the Tutorial Files folder, onto a new audio track in any DAW. Again, we’re using Cubase 7.5 here.

2

We’ll load up the free bx_solo by Brainworx (free from bit.ly/bx_solo) on this channel. The plugin allows us to monitor either the left, right, mid or side channels by toggling the corresponding button. When we isolate the side signal, it’s clear that the track’s stereo information is rather dull and quiet in comparison to the mid information. Now bypass bx_solo.

3

Now load up Voxengo’s free MSED (available from bit.ly/MSEDplugin). As well as letting us monitor the mid and side signals via the Mid Mute and Side Mute buttons, the plugin also features a gain control for each – meaning you can adjust the levels of the mono and stereo signals independently. Turn the Side Gain knob up by 3dB, and hear how the track’s stereo width is increased.

POWER TIP

>Peak freqs

4

Many EQ plugins allow you to affect the frequencies of the mid and side information separately. With MSED still active, load iZotope’s Ozone 6 Equalizer next in the chain. It’s set to regular Stereo mode by default, so hit the M-S button over on the left of the plugin. We now have access to two independent EQ controls, switchable by hitting either the Mid or Side buttons on the left side.

5

Select the Side signal’s EQ, then click on Band 7 and apply a high shelf boost of 5dB at 1kHz. This brightens the side signal, emphasising the high-mid and treble in the hi-hats and bass riff. Move bx_solo last in the plugin chain and solo the side signal to hear this more clearly. We’ve been heavy-handed with our settings in here – usually, you’ll want to gently sweeten rather than drastically alter the width in this way.

As we learned on p51, our ears are more sensitive to high-mid and midrange frequencies – especially the 2-4kHz area. If a track seems to lack stereo width, ensure its side signal has enough presence in this area in relation to the mono signal, and vice versa. Broad bell or high shelf boosts or cuts will help gently re-balance the upper-mid frequencies of the two signals. As with all master processing, don’t feel you need to apply mid/side processing just because you can!



1

Macrodynamics and microdynamics

7. Master ‘manual’ compression

When mastering a track, you might find a verse or intro section to be comparatively quiet in relation to a louder chorus or drop section. Rather than rely upon a dynamics plugin, simple volume automation or audio region levelling can transparently balance track sections. Load Manual Track.wav on a new track in your preferred host software.

2

A song’s dynamics can be defined as its variation in volume over time, and how loud or quiet it is in relation to other songs and itself. When mastering, consider the dynamic relationships of entire track sections working together across a song or album. Is the verse too quiet or loud in relation to the chorus sections? Do any particular areas stand out as being too loud or quiet? Is the loudness of the intro detracting from the impact of the chorus? Any macrodynamic inconsistencies like this are best addressed using basic ‘manual’ compression techniques. Simply chop out the offending audio region and turn that section up or down by a dB or two, or automate the track’s volume. A song’s finer dynamic interrelationships – or microdynamics – are the shorterterm dynamic variations between individual instruments. It’d take a lot of time and effort to draw in volume changes over such a short span of time, but thankfully compressors and dynamics processors can work automatically over a far shorter time period.

The small one-bar ‘fill’ effect before the second drop is extremely quiet in comparison to the rest of the track. We’ll carefully chop out this region of audio and turn the segment up by several dB on the region itself. To prevent clicks at non-zero crossings, we’ll carefully crossfade the regions at either edge.

POWER TIP

>Tiny shifts

3

Levelling whole sections manually is useful if an entire section of a track is too loud or quiet, but it won’t work if some of the mix elements stand out above each other in level – in this case, you’ll probably need to affect the mix’s interrelationships on a micro level with compression techniques, which we’ll get onto soon enough.

> Step by step

1

In this tutorial (and accompanying video), we turned our quiet audio region up by a huge +6dB. This is, of course, an exceptionally large gain increase, and is only necessary thanks to the custom track we’ve used for illustrative purposes. In a real mastering scenario, huge adjustments like these aren’t usually necessary. Only a decibel or two of level-shifting can still make a profound difference upon the perceived balance of a piece of music, adding extra life, contrast and emotion between track sections.

8. Adding average weight with master parallel compression

Parallel compression involves blending an unprocessed dry signal with a heavily-compressed duplicate copy. This technique offers a great way to add average weight to an entire track while maintaining transient detail. Load Parallel Track.wav onto a new audio track, loop it up, then open up an instance of DMG Audio’s Compassion on this channel.


2

Master parallel processing should be a lot less extreme compared to mixing. For parallel use, attack settings should be set as fast as possible (ideally 0ms) to ensure that no transients poke through. We’ll dial in a -30dB Threshold amount, 3:1 Ratio, minimum 0ms Attack, and a Release of 400ms. Bring the Makeup gain up to 12dB to re-level.

3

We can now turn the Dry/Wet knob counter-clockwise to mix between our dry and compressed signals. We’ll set this at 15%. Bypass the plugin to hear how we’ve added weight and sustain while keeping transients. If your compressor lacks a Dry/ Wet control, compress a duplicate copy of the track, or send the signal to a return track with the compressor inserted.

HOT VIDEOS ON THE CHANNEL

SUBSCRIBE FOR: Exclusive hands-on sessions with the latest software

u-he Bazille: Exclusive hands-on bit.ly/CMBazilleVid

Classic Producer Masterclass vaults videos from the 150+ videos, with more added every week!

Cableguys VolumeShaper 4: Hands-on bit.ly/VolShpr4vid

FabFilter Pro-Q 2: Exclusive hands-on bit.ly/ProQ2vid

www.youtube.com/ComputerMusicMag


1

9. ‘Gelling’ a mix with master bus compression

Master bus compression is an oftenmisunderstood technique, so let’s look at how to use it effectively. Create a new audio track in your DAW and drop CM205 Track.wav onto the channel. We’ll load Native Instruments’ Solid Bus Comp – a plugin emulation of the classic SSL G-Series Buss Compressor – as an insert on the track.

> Step by step

2

We set a low Ratio of 1.5:1, a 0.8s Release and a Threshold of 6.0dB, before re-levelling with a 1.5dB Make-up gain amount. A very slow Attack setting will ensure we preserve transients, so the compressor only gently pulls down the track’s sustain elements – try out fast attack times and note the destructive effect of the compression. We’ll settle upon the slowest 30ms Attack time.

3

Extreme compression effects are generally best used at the mix stage, when you have more control over individual tracks and groups. Our smooth 1-2dB gain reduction is extremely subtle, but serves to gently ‘gel’ the overall mix together. You might not notice this ‘gluing’ effect unless you listen on good quality headphones or monitors – we’ll bypass the plugin to carefully observe its effect.

Loudness 10. Master limiting techniques

1

At this stage, your track should be tonally and dynamically balanced, making it sound perceptively ‘loud’ in relation to other commercial examples at equal level. Now it’s just a case of reducing peaks with a brickwall limiter. Here’s CM205 To Limit.wav – lightly compressed from our previous tutorial. We’ve inserted AOM’s Invisible Limiter on its channel.

3

We place T-RackS’ Classic Clipper before the limiter, set to 10dB Gain and -10dB Output, to clip off 1-2dB of the track’s highest peaks. This reduces the limiter’s workload, meaning we can increase its Input Gain to +11.5dB. A/B the result at an equal level to the unprocessed signal. There should be a slight loss of transient detail, but otherwise no real sonic change.


2

Just increasing the Input Gain will disrupt our balance of equal loudness, so make sure you turn down the Output Gain by the same amount to re-level. Many limiters allow you to link the In and Out Gain parameters for this purpose. To do this here, we’ll toggle the Unity Gain Monitoring button, then apply 11dB of Input Gain for 3-4dB of gain reduction.

4

If you hear any distortion or negative side effects, back off the gain reduction. If you can’t achieve enough perceived loudness now, don’t apply more limiting – consider going back for better frequency adjustment, dynamic control or parallel compression. To finish, raise the Output Gain up to full scale – around -0.1 to -0.5dBFS – and render your master.

The dynamic range of a song refers to the difference between its loudest and quietest points. By reducing a track’s dynamic range (by bringing down the highest peaks), the overall track can be turned up, resulting in an increase in average level and perceived ‘loudness’ – at the expense of dynamics and transient punch. Music has been steadily increasing in average level over the past few decades, and this trend rocketed with the invention of the digital brickwall limiter. Labels and artists have always wanted their music to be as loud as the competition. As we know, ‘louder sounds better’, but this race for level led to a public outcry a few years ago, dubbed ‘The Loudness War’. Whether you’re a fan of loudness trends in popular music or not, it’s unavoidable that a modern record must reach similar loudness levels to other commercial tracks within its intended marketplace. However, inappropriate application of limiting can easily lead to unpleasant ‘hypercompression’ and distortion, so it’s important that your track is well mixed (ie, tonally and dynamically balanced) before master limiting is applied. Simply cranking a limiter’s input gain over your master will make your track ‘loud’, but at the expense of clarity and fidelity.

modern mastering / make music now < > Step by step

11. Mastering the

1

In our Mastering Masterclass videos, you’ve seen professional mastering engineer JP Braddock process and enhance 209’s house track with mastering hardware. Let’s now replicate his process and master the same track in-the-box using native third-party plugins. Load CM209 Session Mix.wav onto a new audio track in any DAW – we’re using Cubase 7.5 here.

4

In the Compressor section, we’ll dial in a 2.5:1 Ratio, 39ms Attack, 350ms Release and a Knee of 2.8. Setting the Threshold to -30dB achieves 3-6dB of gain reduction. In the second band, set a 2:1 Ratio, 50ms Attack, 300ms Release and a Knee of 7.0. A Threshold setting of -33dB then triggers approximately 3dB of gain reduction. We’ve levelled out the bass’s dynamics, and we’ll make up this region’s drop in level later with EQ.

7

Let’s raise the mid signal’s treble with a high shelf, emphasise it with a highmid bell boost, then shelve down the treble in the side signal. Raise the mid’s High band to 2.50 at 5k6 (5.6kHz) and turn the shelf’s Q right down for a gentler slope. On the mid’s High Mid band, raise a bell boost of 7.00 at 3k3 (3.3kHz) with the Q knob at 7 o’clock. On the side signal’s High band, make a shelf cut of 13.00 at 8k2.

209 house track with plugins

2

To begin, multiband compression tightens up the track’s overly dynamic bass end. Load iZotope’s Ozone 6 Dynamics plugin on the channel, right-click on the rightmost band and select ‘Remove Band’ – we’ll only need three. For each band in turn, set the Compressor Thresholds to 0dB, initialising them but not yet applying gain reduction. Set the third band’s Compressor and Limiter Ratio values to 1:1.

5

The track has too much treble in the stereo field, but not enough in the middle of the mix. To correct this uneven distribution, we can equalise the track’s mono information and stereo information independently. Load Native Instruments’ Passive EQ next in the chain, and toggle its Mid/Side mode via the switch at the plugin’s top-centre. The left-hand controls now affect the mono signal, and the righthand controls will affect the stereo signal.

8

We raise the side signal’s Gain to 1.5dB, increasing the track’s width and compensating for the loss of level from the treble shelf cut. To raise the track’s bass end (that we compressed earlier), we boost the mid’s Low band shelf to 10.5 at 330Hz, with the shelf’s Q set fully right. Load a second Passive EQ instance, and apply a bell boost of 1.00 to the mid signal at 27kHz, further enhancing its treble detail.

3

In the plugin’s Options menu, head to the Dynamics tab and change the Crossover Type to Digital, giving us a cleaner, more transparent sound at the crossover points. Drag the leftmost band’s crossover all the way left to 74Hz, then drag the other crossover point to 151Hz. Solo the leftmost band to hear the track’s sub frequencies in isolation.

6

Unlink all the bands by turning off the purple LEDs under each, then deactivate the mid’s Low Mid band and the side’s Low, Low Mid and High Mid bands (by hitting the title switch at the top of each). Pull the left HP knob up to 22 to high-pass the mid signal at 22Hz, cutting unwanted subsonic frequencies. Next we’ll make some cuts and boosts. And by the way, note that this EQ’s gain knobs do not represent dB values!

9

Full-band compression will gel things together – we load NI’s Vari Comp, setting the Threshold to 7.00, Output to -10.5 and Recovery to 4000ms for only 1dB of gentle gain reduction. Now T-RackS’ Classic Clipper’s Gain is set to 6dB, with the Output rebalanced to -6dB. AOM’s Invisible Limiter is last – its Input Gain is set to 7.35dB for 1-2dB of limiting.



12 mastering tips HANDS-OFF MIXING We’ve gotta say it again… Don’t expect mix problems to be solved at the mastering stage! EQ clashes, dynamic issues and other errors are all best addressed from within the mix project. If you’re applying drastic amounts of processing, revisit your mix – or if you’re mastering for someone else, explain the issues and see if they can remix.

KEEP IT FRESH Don’t over-listen to a track! John Paul Braddock explains: “What I don’t want to do is to listen all the way through the track for six minutes, because as soon as I’ve done that, I’ve got used to how it sounds, rather than being objective. It’s crucial that we don’t spend too much time listening to the music. This might sound counter-intuitive, but we’re not mixing it any more. We’re not trying to listen to the detail; we’re trying to get an overview – to sample the overall tone of the song.”

Try distributing a plugin’s workload over several stages for a potentially more transparent effect

CONSIDER IT DOES IT CANCEL OUT?

STAGED LIMITING Several gentle stages of limiting or compression can help take the load off one single plugin. For example, three limiters with a gain reduction of 1dB might sound more natural than a single 3dB limiting stage. It depends on the plugins used, so give it a try, and listen objectively.

MAKE IT UP

Use ‘sum difference’ testing (also known as a ‘null test’) to hear if a plugin is ‘passive’. Get to know which plugins add gain boosts or frequency changes in their default state. John Paul Braddock discusses: “Many plugins will actually apply a tonal or level change even before any settings are dialled in. I‘ve noticed that, after analysing several types of plugins, you’ll load up a plugin with ‘no processing’, but the actual output might be louder. Perhaps those plugin manufacturers know what we now know – that ‘louder sounds better’ – or maybe that’s just a side effect of the plugin’s design. The important thing is that you analyse the tools you’re using, and don’t make assumptions. Be critically aware of your own tools.”

For transparency, try to use as few EQ or excitement stages as possible. So, if a track has too much bass and not enough treble, try using a single broad shelf to cut bass, then re-level by increasing the EQ’s makeup gain. This will shift the track’s weight towards the treble more naturally than two EQ bands.

Remember to compare your final processed master with the unprocessed session mix – at equal level – to see if you’ve actually achieved the outcome you intended. If not, don’t be afraid to start again from scratch.

Once you’re ready to master a track, don’t just dive in and start processing. The aim is to gently improve, not ‘mix’. Take a more considered approach. Briefly compare the mix to a reference track at equal level, plan exactly what correction or enhancement the mix needs, try it out, re-level, then evaluate.

ON THE KNOBS Type in parameter values and use stepped plugins (with fixed 0.5dB-1dB ‘notched’ controls) where possible. It’s easy to just crank up a knob, but typing in values makes you think about what you’re entering. Stick to 0.5/1dB steps at a time, as half a dB will make a significant difference when mastering.

REFERENCE WITH EQUALITY

MID/SIDE DIY A plugin with an unlinked left/right mode can also be used to process in mid/side. Simply load Voxengo’s free MSED on the channel and set it to ‘Encode’. Now load your plugin after MSED and unlink the left and right channels. Place a second MSED last in the chain, and set it to ‘Decode’. The left side of your plugin now processes the mid (mono) part of your signal, and the right affects the side (stereo).

MORE THAN AVERAGE Regular downwards, full-band compression can clamp down on peaks and transient detail, ruining dynamics if not applied carefully. If your track needs extra average weight, consider blending it in through the use of parallel compression – you can bring up the average level of your track while keeping the detail intact.

COMPRESS GENTLY A touch of downwards compression can pull (or ‘gel’) the overall mix together, but keep attack times slow so you gently clamp down on the mix’s sustain and not the transients. A low ratio and around 1-2dB of gain reduction should be all that’s necessary.

LIMIT LAST

Evaluate your mastering success by comparing your processed master with the unprocessed version – remember to set both tracks to equal loudness for a fair comparison


Many think of limiting and loudness as the main staples of mastering, but this attitude often leads to amateur results, flattened mixes and distortion. Final peak limiting should only be tackled when a track’s overall tonal, dynamic and stereo balance are in order. So leave limiting till last!

Modern Mastering CM

Recommend Documents