Guide to Mastering

YOU'VE JUST COMPLETED the mixdown of the final track to go on your groundbreaking demo EP. You listen through your album several more times, but the more you compare it to your favourite CDs by the artists you admire, it always seems to fall short. It sounds dull and lacks impact, it doesn't flow right and, quite frankly, you're getting tired of hearing the same track over...

These feelings are shared by pretty much all of us involved with making music, including some of the best mixing engineers and producers in the world. This is why, prior to sending an album or single off for commercial CD or vinyl duplication, a lot of people pass their recording by a mastering engineer's keen ear to ensure it's technically suitable for release. Engineers possess both the tools and the complete understanding to make any necessary changes to present the music in its best possible light. This could involve any number of things, from dynamics processing for a heightened sensation of impact or spectral balancing using EQ for maximum high fidelity, through to bass and stereo enhancement or authoring for CD duplication. The mastering engineer is the absolute final check of quality before the music hits its target, whether that be record label A&R offices or even the retailers' shelves.

DIY or pro

With the explosion of the home studio generation in recent years, manufacturers have seen to the needs of budding wannabe engineers with a series of fairly affordable products to carry your musical vision from inception to completion, without ever having to set step inside a professional studio. So why is there such a strong emphasis placed upon having your music treated to a professional mastering service?
The immediate benefit is the mastering engineer's fresh set of ears. Unlike you, they haven't heard the tracks an umpteenth number of times, including its various stages of arrangement, remixes, or the parts that didn't make the album, so they can remain objective and hear it for its face value.

Mastering studios are equipped to pass your music through some of the finest equipment in the world. The odd cheap effects unit is fine on a single instrument, but hardly the sort of thing you'd trust your main mix through. Engineers know what things should sound like, and how to get it there. They'll also hopefully recognise when something should be left untouched, and not process for the sake of processing. But it is perfectly possible to do it yourself...

There are plenty of people producing favourable results with their own smaller studio set-ups. Being involved at every stage of the process, including mastering, is undeniably part of the attraction of owning our own gear. There's also the worry of placing your dream into the hands of a stranger. Your chosen mastering engineer must have an affinity and complete understanding of your vision, otherwise they're hardly going to be in tune with your expectations. However, by learning the craft of mastering, you'll come to understand what's required to make a tune sound great, and also the limitations.

Mastering should definitely not be considered the magical cure for a bad mix. "Don't worry about that mate, we can fix it in mastering," just isn't going to cut it in most cases. A great mix makes an even greater master. Practice your mastering skills on your demos. If you eventually move on to releasing your music commercially, still consider having your music professionally mastered, if only on an educational level to discover how you might further refine your skills. So allow me to show you how to begin doing pro-level mastering for yourself...

Listening in colour

Whether you listen in colour or in black and white can make all the difference to your recording, and though that may sound intriguing, let me explain. Take Spielberg's epic war drama Schindler's List, or any black and white film. The further your'e drawn to the characters and the storyline, the less your mind realises you're not watching in colour. The human brain compensates for any deficiencies you may see (a bit like when you can imagine you can see a full landscape, even though something like a pillar or a fence may be in the way). I've dubbed this perfectly natural phenomenon as the 'black & white effect', and it applies to music as well.

It's so easy to listen to a cruddy recording yet be totally touched by the musical genius of the performance. But if that performance is three minutes of a radio show, it's essential that it shines in all its colourful glory, or it will fall behind the competition. During the many months (sometimes years) of writing, recording and mixing your album, you tend to become so intimate with the musical storyline that it soon becomes impossible to make objective decisions in what could improve the recording, hence the mastering process. If you plan to be your own mastering engineer, you must always remain grounded to recognise this potential problem from its onset.

When mastering, carefully note any areas requiring immediate attention, and be wary of other issues creeping in right under your nose during the mastering process that may result in a substandard 'black and white' master. You have to work quickly. Listen, make an alteration if necessary, compare results and move on. If you get caught up trying to fix one problem for too long, take a break and come back to it when you're feeling fresh and able to make a fair judgement.

Continuity

The mastering engineer's role is to create a sense of continuity. This may involve ensuring an album flows smoothly with regards to level matching and the transition between songs, or so each track will sit nicely alongside other independently produced and mastered material on radio, for example.

I was recently working on a mastering project for a new music label wishing to showcase their artists on CD in a continuous mix. So rather than just one artist's vision to work on, there were several, each with differing musical styles, and produced in different studios (some entirely hardware, another using only software-based synths and samplers), and varying levels of production.

Before I could even consider the DJ-like mix aspect of the job, I had to work through the tracks first and subtly touch up each one using my mastering tools. The end result was a product that enhanced the work of each artist, while still retaining the essence of their individualism, which would finally gel together as an album with a quality comparable to other compilations on the market.

Hard or soft

There are two trains of thought when it comes to the mastering format of choice. Traditionally, mastering was very much a hardware-orientated affair, calling on finely engineered analogue compressors and limiters, exciters, bass enhancement circuity and equalisation. Unfortunately, such gear is prohibitively expensive for most of us.

With the arrival of CD, the focus turned towards digital processing algorithms. For the PC or Mac-based home studio, this has lead to huge popularity of reasonably priced, standalone software applications and plug-ins that run within your audio software's environment. They offer a distinct advantage over their analogue hardware counterparts through their ability to instantly save and recall user set-ups. They also allow multiple occurrences of the same processor, meaning you can simultaneously run a new compressor on every track on an album, for example, with each one configured for the track in mind. Using hardware, you'd have to possess several units for this sort of flexibility, or be prepared to write down your settings as you move from track to track, and possibly feel less inclined to go back and make changes out of sheer inconvenience.

Still, things aren't necessarily rosy in the plug-in world either. Once you start to price some of the upper-end hardware, you'll no doubt see an immediate attraction to software, but there's more to it than mere economics. Most entry-level packages are bundled with an assortment of audio processing tools. In fact, with the VST platform being open to development by anyone, you'll stumble across plenty of freeware effects as well. These are generally fine for creative applications on individual sounds, but probably won't stand up to the stringent standards required for mastering your finished track. Instead they may result in a loss of depth and definition, and inflict an artificial digital character.

Third-party developers, whose core business relies on quality plug-in products, are a much better investment. The Waves collections (www.waves.com) in VST, DirectX, MAS, TDM and RTAS are particularly good for their highly musical yet transparent parametric EQs and multiband compression. Possibly their most dramatic mastering product is the L1 Ultramaximizer, which has often been referred to by users as "worth the admission price alone".

The L1 incorporates a look-ahead peak limiter (to capture ultra-fast volume spikes as they occur) with an automatic make-up gain for maximum volume potential, plus the appropriate dithering algorithms to reliably convert your 24-bit audio file down to a CD-friendly 16-bit. This plug-in will easily raise the apparent level of your material to a commercial standard, but can absolutely wreck the dynamic emotion of a performance if abused.

Do-it-all

An all-in-one package aimed at the home mastering market is IK Multimedia's T-Racks 24 (www.t-racks.com). Rather than relying on a host program to function, T-Racks 24 is completely free standing; just import your audio files in to the program and you're away. The emulation of valve-based analogue hardware is very much designed to colour your sound, albeit in a positive manner, therefore opinions on its worth in preserving the original recording in a mastering role are a little mixed. However, T-Racks users absolutely rave about it, and it might be of particular interest to those of you who use a lot of software-generated instruments and are looking to add a pinch of analogue character to spice up your recordings.

Another couple of quality mastering packages are Steinberg's WaveLab (www.steinberg.net) and Emagic's WaveBurner Pro (www.emagic.de). Both incorporate wave editors, VST mastering effects and CD burning facilities.And don't feel you necessarily need dedicated mastering software. The audio environment in your present audio sequencer should be adequate if the effects are up to scratch and may be auditioned and tweaked in real-time, rather than waiting for the entire file to be processed first.

In some cases, a mix of hardware and software is used, but bear in mind that every time music passes back and forth through the digital-to-analogue converters (DAC), degradation to the waveform is inevitable. This is a strong reason why professional studios employ precision converters from the likes of Apogee (www.apogeedigital.com) and Prism Sound (www.prismsound.com) to interact their digital audio with the outside world. Perform some critical listening tests with your own equipment. You may find it best to remain in the digital realm, if only to avoid the digital-analogue headache, but also due to the average analogue rack gear being far from noiseless in itself.

The hard way

Lastly, there are a number of hardware-based digital processors available that are packed with essential mastering tools. They're like the best of both worlds: custom DSP (digital signal processing) chips mounted within a sturdy rack chassis. Examples include the dbx Quantum and Quantum II (www.dbxpro.com), Drawmer Masterflow DC247(www.drawmer.co.uk), and the ever-popular TC Electronic Finalizer Express and Finalizer 96k(www.tcelectronic.com). These devices are an excellent solution for situations where a computer isn't appropriate (such as live performances) or for those looking for a product that offers superb quality, straight out the box, with a minimum of fuss.

If you're using the processor in conjunction with your computer, it's essential to take advantage of the digital connections to avoid an unnecessary analogue conversion. Be wary of the seemingly preset nature of these boxes, as mastering is never a preset process. Like all aspects of music production, it's a matter of getting in beyond the presets and working the settings to discover your own magic formula that works in harmony with the material you're mastering.

Mastering monitors

Your audio monitors act as a magnifying glass and finetooth comb for picking out any errors and inconsistencies in your audio material, allowing you to correct it prior to duplication and commercial distribution. In a typical mastering house, the popular nearfield monitors we often find ourselves mixing on are nowhere to be seen.

In favour are the larger (and considerably more expensive) full-range main monitors, such as the Genelec 1035B (www.genelec.com) or Quested HM212 (www.quested.com), that offer unsurpassed levels of clarity,imaging, dynamics, and a frequency response that just isn't available to the smaller nearfield market. Often found soffit-mounted (flush to the wall), these monitors fire into a control room that's been acoustically optimised to minimise reflections, distortions, and resonances.

Of course, most of us can only dream of ever working in such a pristine environment, so you have to make the most out of your current monitoring set-up. Ideally you shouldn't be mastering on the same set of monitors that were used for mixing, otherwise rouge nasties may slip through the net without being noticed, so, at the very least, a secondary speaker system should be available, alongside your primary monitors.

Get to know how music sounds on a range of speakers, such as multimedia speakers or the popular Auratone Cubes (to emulate the typical television and small radio) and bigger hi-fi systems. If you're preparing music likely to be played in nightclub, be wary of over-hyping sub bass frequencies due to the inadequacies of your present monitors.

This is where a mate's megaWatt, stereo-on-wheels, car audio system could come in handy. The more references you have access to, the better. This helps you predict how your masters in progress will eventually translate on your intended audience's wide range of playback systems.

Next comes the room itself. Forget the speaker manufacturer's advertising blurb and technical specs for a moment; stop and listen to how the room sounds. Walk around and observe how sounds change throughout the room. These acoustic variations, caused by soundwaves collecting and reflecting, and rising and falling throughout the room, are perfectly natural phenomenons, yet can heavily influence the decisions you make during your track mastering. In my own studio, I discovered a mid-bass resonant peak after critically listening to numerous recordings I was familiar with. If I ignored it, I might have risked removing these frequencies during mixing or mastering through equalisation, so I rearranged my set-up by repositioning my workstation and monitors a little towards the centre of the room until I felt everything sounded more natural.

Are there any improvements you can make to your own monitor placement? Be aware of the room walls, floors, corners and ceilings that may be hyping bass frequencies, and note any hard surfaces that can reflect and smear mid and upper range frequencies. Sometimes it may be simply a matter of installing a set of heavy curtains over a window, or shifting a couple of racks of equipment. Most monitors are magnetically shielded to protect computer and TV screens from picture distortion, but also have a think about how your soundwaves may be bouncing off the sides of your computer monitor screen.

Mastering your meters

Mastering requires highly accurate level metering. The LED meters found on most mixing desks, as well as the displays in audio software, are generally set to track the fast travelling peak of the audio waveform with precision.

While this is essential for keeping a watchful eye on clipping when recording digitally by exceeding 0dBFS (FS = Full Scale, the maximum level a waveform may be encoded digitally), program peak metering does not relate too well to the real world. Our ears tend to respond to the slower, average volume level, while the ultra fast transients go unnoticed. Two songs with identical signal peaks may appear different in volume simply because one of the songs average level was lower. This is why analogue VU meters are still popular among studios because they give a good representation of how the human ear perceives volume.
Look for an RMS setting on your digital level meter, such as the PAZ Psychoacoustic Analyzer (www.waves.com) plug-in for Mac to get a real world idea of the average volume. Most people strive for a loud mix, but be wary of setting your average level too high. Apart from becoming fatiguing to listen to, you may run into problems in radio broadcast situations where the station's limiters confuse the main energy of your track with the instantaneous peaks and attempt to squash it further. As a general rule, the difference between your peak levels and your average levels (known as the peak to average ratio) should never be less than 6dB.

One final point with metering: although it makes a terrific reference, a level meter is still not perfect. It registers evenly across the entire audio spectrum, whereas our hearing is more sensitive to mid-range frequencies, so always let your ears make the final verdict.

Another couple of worthy features from the PAZ plug-in are the Spectrum Analyzer, to visualise how energy levels are distributed among the audible frequencies (handy for comparing bass levels between different songs, for example); and the Stereo Position Display which also deciphers out-of-phase material. For a quick phase demonstration, reverse the polarity of the wires connected to just one of your stereo speakers (connect positive to negative, and negative to positive) and play back some music through them. Note the sensation of extreme width, minimal bass, and an obvious hole in the centre of the music. This is the due to the left and right channels working 180° out of phase from each other. The cones are literally pushing in different directions.

Sometimes, a sound can be thrown out of phase while recording or mixing if a channel is delayed for a stereo effect, or a studio cable or patchbay is wired incorrectly. Should the out-of-phase material be played through a single speaker set-up (like a television, for example), the true mono component of the audio would completely cancel out and become inaudible. So, be wary of out-of-phase sounds in the mix, and if in doubt, check it on a monaural source for compatibility.

Setting up

In order to get set up, the first step is to prepare your tracks for the session (see the Step-by-step #1 box on p113 for more on this). Close down all unnecessary computer applications so the CPU's clock cycles are available solely for music reproduction and accurate on-screen level metering. If you're accepting mastering jobs, ideally the tracks should be supplied on DAT or CD-ROM. Even portable hard disks are common these days. Audio on CD-R may be used, but is best avoided due to lost bits of data being unaccounted for during the ripping process when returning the CD back to a computer file.

Encourage your customers to supply their material without any main bus effects processing like compression or normalisation, and have them retain the same bit depth and sample rate that their mix was created in.

Every time a file is converted from a higher bit/sample rate down to 16-bit/44.1kHz (CD format), there is some degradation, either through truncating the extra bits from each word, or from the preferred approach of dithering. I personally mix in 24bit/44.1kHz and stay that way until the final stage of mastering when I'm ready to commit to dithering down to 16-bit.

Hearing is believing!

On a final note before delving into the actual process of mastering, you must consider the most important piece of test equipment in the studio: your ears. Every decision you make will be run past your ears for analysis, so there's a lot of responsibility riding on your ability to hear things as they really are.

Unfortunately, our quality of hearing fluctuates throughout the day, particularly after loud or extended listening periods, and when we're feeling tired and stressed, so always devote a fresh day to mastering. You may also wish to consider having your ears tested professionally, which would be a very good idea if you're setting out on a career in mastering.
Before starting a session, pop on a familiar (and professionally mastered) CD that compliments the style of music you plan to be working on. Don't monitor too loudly or you'll spoil your ears before you've even started! While listening, note the bass presence and definition; the prominence of the vocals, synths or guitars; and the detail of the cymbals and other treble in the mix. Observe your level meters: watch where the peaks of each beat sit, as well as the quieter periods of the piece. These practices serve to calibrate your hearing by placing a fresh reference in your mind of what things should sound like, in living colour!

Final advice

Maintain moderate listening levels. Once you think you've nailed a setting, turn it up to full room level and see if everything still sits right, before bringing your monitoring level right back down again to preserve your hearing's sensitivity. Always refer back to your reference material throughout the session to ensure you haven't steered off track into 'black & white' territory.

Listen for the message the composers are putting forth. They're ultimately the creative forces behind the music, and it's your role as the mastering engineer to enhance their original vision by working in harmony with it.

Although mastering is very much a technical process, the fundamental settlement of each mastering decision falls squarely back on your understanding and love for music as an artform. This appreciation continues to grow throughout our lifetime, allowing each of us us to broaden our proficiency at mastering through experience.

DIY--> STEP-BY-STEP #1: Setting up a mastering session on computer

A guide to setting up a mastering session for an album in Pro Tools. Apply the same principle to any multitrack audio sequencer with real-time effects...

1.Open a fresh session with sufficient stereo tracks to cover the songs
you plan to master, plus another for your reference track. This point of reference should be a piece of professionally produced and mastered music that you aspire to, and in the same genre as the tracks you're mastering.

2.Add the insert effects you might use for mastering. Although you might not end up using every one as it really depends upon what you feel each track requires, I like to set up the most commonly used tools. A few examples are a parametric EQ, stereo enhancement, a multiband compressor, and limiter.

3.Listen through all the tracks first, and make careful notes to ensure nothing is overlooked. Go back to you reference track to calibrate your senses, then work on the track that you feel is already close to perfection. Use this track has an example of where the remaining tracks should be steered towards.

4.Work your way through the tracks making subtle changes if required. Compare them to each other to ensure they're of uniform consistency, and of course how they stack up against your reference track. The PAZ Analyzer comes in handy for comparing volume levels and frequency content, but your ears should always be the final judge.

5.Trim the starts and finishes of each track. It's a very good idea to insert at least 100ms before the start of each track to ensure the first beat or note is not missed by some CD players that mute their output for several milliseconds upon playback. Add any fade ins or fade outs. It's crucial the fades take place after the compression or limiting, otherwise the change in level beforehand will upset the processor's response.

6.Take the opportunity now to listen to how each track will flow into the next as they will eventually appear on the finished album. Adjust the relative volume of each track to suit, using your ears, not necessarily your meters. Bounce each track (including the effects) as a separate file to disk, not forgetting the blank portion before each track starts. Load up your CD burning software, add your tracks, set the gap between each song to taste (commonly two seconds), and burn CD using the Disk At Once setting (and not Track At Once).

STEP-BY-STEP #2: Sorting out the EQ

For accurate equalisation with pinpoint precision, you'll need a multiband parametric EQ that offers full control over the frequencies...

1.Place the EQ first in the chain of mastering effects. This ensures subsequent processing (eg, compressors) won't respond to frequencies that have since been removed. Start by setting a high-pass shelving filter to around 30-35Hz. This eliminates the subharmonic content of the music, and can actually improve the bass definition by returning lost amplifier power and minimising speaker instability caused by frequencies below their limits.

2.A neat trick to bring out certain instruments and improve clarity is to slightly notch out the frequency immediately below the instrument you want to bring out. Watch out that you don't take out the body of the mix though. It is very much a juggling act; a sharp dip somewhere between 150-300Hz may help reduce a muddy mix, while gradually peaking up above 8kHz may provide some high fidelity air. Don't concentrate on specific instruments though, just the overall tone of the music. Use a high Q setting to focus in on a narrow band of frequencies, or a low Q to cover a broader range of frequencies.

3.Go back to your reference CD and make comparisons. There's no preset formula for the right EQ, and if the track's been mixed well, you should be able to leave it alone. You are more looking to gently nudge the EQ into place by a few dB, rather than a surgical operation. Should any problem frequencies ring out too much, locate the exact frequency by sweeping a filter peak up and down the audio spectrum until the offending frequency stands out more, before reducing the EQ back down at that point. Compare your EQ with the effect bypassed to ensure your changes are beneficial.

STEP-BY-STEP #3: Multiband compression

For transparent compression resulting in a louder mix with minimal pumping effects, you might find it's better to use a compressor that treats different frequency bands of the audio spectrum separately (on these pictures, we've shown a TL Audio Ivory compressor)...

1.Before considering a multi-band compressor, you should be well accustomed to using a regular single-band device. Refer to The ultimate guide to compression feature from FM114. When placed across an entire mix, a compressor's job is generally to lift the overall perceived volume of the music and can also help create a tight and more defined mix. It works by restricting the fast volume peaks in the music. This in turn allows you to turn up the output resulting in a louder sounding mix whilst the peaks are kept safely under control. NB: The dynamic range (ie. difference between loud and soft musical passages) will be reduced by compression.

2.A multiband compressor is two or more compressor circuits working alongside each other, but set to concentrate on their own section of frequencies. Start with a compression ratio (comp) of 2 or 3:1 on all bands, although you may choose to increase the lowest band up to 4:1 for dancefloor-based music. The ratio determines the degree of gain reduction. For example: 1:1 means that the volume remains unchanged; 4:1 means that for every four units of volume that enters above the threshold, there'll be just one unit of volume at the output.

3.Start with the lowest band (the bass) and lower the threshold (circled) until about 3dB of compressed signal is registered on the gain reduction meter. Then raise the output gain for that band to compensate for the loss of volume incurred by the action of the compressor.

4.Now lower the threshold of the next band until about 3dB of gain reduction is registered, and raise the output (circled) to compensate. The upper band(s) covering the treble frequencies will probably require no compression, so just lower the threshold to catch any peaks, although slight high frequency compression can be desirable to emulate tape saturation.

5.You'll find that the EQ balance across the mix will now be all over the place as the volume of each band has been altered. To readjust, use the output gain controls on each band, bypassing the compressor at times to check for the original tonal balance.

6.Go back and listen carefully for transients in the music like the thump of a kick drum, or the crash of a cymbal. How do they sound before compression? Is the change for the better? They should still retain their dynamics without restraint. Increase the attack portion(s) of the compressor to allow more time for the initial peaks of these instruments to pass through, or lessen the threshold or compression ratio.

STEP-BY-STEP #4: Using the Waves L1 plug-in

The Waves L1 Ultramaximizer plug-in is an excellent yet simple way to dramatically lift the perceived volume of your music. It should be the last thing you do to your track in the mastering process...
1.L1 is capable of pushing the waveform peaks extremely close to the maximum digital limit of 0dBFS (Full Scale), meaning further alteration to the volume following the L1 (eg, compression or EQ) may result in clipping/distortion of the waveform. Also, the L1 can implement a dithering routine to smoothly convert from say, a 24-bit stream to16-bit, ready for CD duplication. For these reasons, always place the plug-in last in your effects processing chain.

2.Set the Out Ceiling no higher than -0.5dB. This controls the automatic make-up gain that compensates for the reduction in output due to the action of the limiter. Basically, this sets the level your music peaks will leave the limiter at. It's a good idea to not set this to the maximum (0dBFS) in case some peaks still get through, or is read by some CD players as a digital overload. Now pull down the Threshold and observe your music becoming louder. Compare your average volume increase to that of your reference material. While this plug-in is nothing short of amazing, listen carefullly for over-compression taking the life out of the performance making for a very fatiguing listening session, and heavy abuse of the threshold will generate audio distortion.

3.Choose the Quantise bit rate for your destination. Whenever you move from a higher wordlength to a smaller wordlength, the file should be dithered for the cleanest conversion. For example, if you recorded and mixed in 24-bit, stay this way until you've completed your mastering and you're ready to dither your 16-bit file for CD. If you were originally working in 16-bit, you can convert to 24-bit for mastering, and then dither your file back to 16-bit. Due to the degree of noise introduced, the dithering process to 16-bit should only be performed to the file once.

Adding herbs and spice

Following on from the standard array of EQs and compressors, studios often have a few more tricks up their sleeves in treating their tracks to seize your attention. Much of the technology behind these intriguing products is buried amongst patented secrets, but those that use them rave undeniably over their effectiveness. A few examples are: the SPL Vitalizer (www.soundperformancelab.com) for its interesting yet highly musical slant on EQ; the Aphex products like the Aural Exciter to add sparkle though the harmonic restoration and Big Bottom for dramatic bass density (www.aphex.com); plus the S1 Stereo Imager (www.waves.com) plug-in to expand the instrument panning out and beyond the speakers.

So whether you think you're after a dash of one of these special effects or just a hint of Lexicon reverb, the results are often quite stunning, but as with anything in the mastering process, less is often more. Use these types of effects subtly when mastering, and always make comparisons to your original version to ensure you haven't actually made things sound worse in the process. Save the heavy-duty effects for your music composition.

To normalise or not to normalise

Popular with many DAW (digital audio workstation) operators is the Normalise function. This is a failsafe method for raising the volume of an audio clip towards the maximum level it may be encoded digitally (0dBFS), without fear of overload or clipping. The computer process begins by first scanning the file to automatically locate the loudest waveform peak, then the volume of the entire section of audio is increased until this peak reaches the digital limit (or a user-determined percentage thereof). Sounds great in theory, right? Perhaps not...

First, a high-resolution digital recording will only reap the benefits if the waveform was captured as a loud signal in the first place. Increasing the gain digitally (like normalising) won't help. Instead, it only reveals the limitations of the digital realm in the same manner as zooming in on a computer image will show up the block-like pixels that form the picture. Also, a pop, click or one-off transient can easily trigger a false reading so there may not be a significant change after normalising, plus any process applied to a waveform will incur a degree of noise or distortion.

Mastering involves additional dynamics processing, either through balancing the perceived song volume levels or through EQ or compression. This makes prior normalisation totally unnecessary. It's the average or perceived volume of the music that matters in the real world, not an ultra-fast peak that can't be heard. For example, when mastering a collection of songs for an album, you'll find yourself adjusting each track's volume differently to ensure a smooth transition throughout the course of the CD.

If you then normalise each track, the perceived volumes will be thrown all over the place in favour of the program peak. This might mean the subtle opening notes of an Ibiza inspired chillout song you chose to close your album with may now end up blaring at maximum volume, destroying the harmonious effect you were trying to create.

Credits to Intermusic.

Jquake