Snoman R. Dance Music Manual: Tools, Toys, and Techniques

Подождите немного. Документ загружается.

PART 3

Mixing & Promotion

404

This same process applies when a mix that sounds too ‘nasal’. In that case,

place a cut at 500 Hz with a Q that will create a slope from 250 Hz back up to

1000 Hz and slowly reduce the width of the Q .

Above all, bear in mind that wide bands will sound more natural to our ears.

Try to keep the Q quite wide unless you’re making surgical corrections to a spe-

ciﬁ c frequency. If you ﬁ nd that you have to use a very narrow Q or too much of

a cut to remove the problem, then the trouble is a result of the mixing and you

should return to the mixing stage to repair it.

High Range

Finally, carefully listen to the high-end frequencies in the mix. In the improb-

able event that these are too harsh or bright, wide cuts placed at a centre fre-

quency of 14 KHz can reduce the problem, but it is more likely that they will

not be bright enough when compared with commercial CDs.

Although brightness could be added to the signal by placing a wide Q with a

boost of a few decibels at 12 –16 KHz, as touched upon in the previous chap-

ter, gain boosts are usually avoided, with many engineers preferring to use

harmonic excitement for brightenness. Not all engineers will use harmonic

excitement, preferring to leave the signal unaffected. In this instance they’ll

employ the Baxandall curve.

This is created by using a wide Q setting on a parametric EQ and boosting by

6 or 8 dB at 20 KHz. The width of the Q introduces a gradual transition curve

from 10 KHz upwards which is not only similar to the ear’s natural response

but also introduces small amounts of harmonic distortion similar to valves.

Both these factors contribute to a pleasing sound that is often preferrable to

the sound acheived using harmonic enhancers.

While the majority of these adjustments should be made in stereo to maintain

the stereo balance between the two channels, occasionally it may be more beneﬁ -

cial to make adjustments on only one channel. This is useful if the two channels

have different frequency content from one another or there is a poorly recorded/

mixed instrument in just one of the channels. In these instances, subtlety is the

real key. The best approach is to start with extremely gentle settings, slowly pro-

gressing to more aggressive corrections until the problem is repaired.

Also, when it comes to this form of surgical EQ, the most important rule of

thumb is to avoid ‘itchy ﬁ ngers ’ and to listen to the material carefully before

making any adjustments. You need to identify exactly what the problem is before

applying any adjustments and even then adjusting only the sections that require

it. If the ‘problem’ frequency sounds ﬁ ne while it plays along with the rest of

the track but appears wrong in the breakdown, then it’s more sensible to correct

it during the breakdown than to attempt to change the frequencies throughout

the mix. Ideally, there should be more things right with the mix than there are

wrong, and if you concentrate too heavily on a fault you could end up mess-

ing up the rest of the frequencies while trying to repair just one small problem.

Mastering

CHAPTER 20

405

Like mixing, this type of EQ should also be applied decisively and rapidly.

Although it is difﬁ cult to perceive on a home monitoring system, changes of

less than 1/2 dB are audible and our ears can quickly become accustomed to

the changes, so it’s vital that you compare the recent EQ movements with the

original mix.

As a general rule of thumb you should play the same piece of the track being

mastered over and over for up to 10 min before switching to the un-EQ’d

comparison and listening to that for another 10 min to hear the changes that

are being made. On this same note, this should also be referenced against

no-expense-spared professionally mastered music of the same genre, both

aurally and visually, using a spectral analyser.

BOOSTS AND CUTS

Although frequency boosts are generally best avoided altogether to retain a

natural sound, any EQ adjustments will affect the perception of the overall

tonal balance of the whole mix and so must be applied carefully. For example,

even applying a 1/2 dB boost at, say, 7 KHz could have the same result as dip-

ping 300 –400 Hz by 2 dB. Always bear in mind that all the frequency ranges

on a full mix will interact with each other and even small changes can have an

adverse affect on every other frequency, so whenever you make a change you

will have to reconsider all the others too.

More importantly, the combination of all these EQ settings will determine the

overall dynamics of the song. This is one of the most critical aspects of any

mix – it must sound correct for the style of music. When anyone listens to a

mix, the ﬁ rst thing they’ll hear is the overall EQ; if this is wrong for the genre

of music then the track will fall ﬂ at. This is where a mastering engineer’s exper-

tise and subjective viewpoint plays yet another role. Every genre of music is

approached and balanced differently, and there is no deﬁ nitive guide to how

any one genre should be balanced.

As an example, consider two entirely different genres of music: techno and chill

out. A techno record will generally need to sound sharp and aggressive to suit

the style. This can be accomplished by cutting the lower- and higher-range fre-

quencies, so the mid-range becomes more apparent. Conversely, chill out will

be quite slow and nowhere near as aggressive, so a cut in the mid-range will

accentuate the low and high frequencies, making it appear gentler to the ear.

Similarly, euphoric trance is encapsulated by the ﬁ lter swept lead motif, so this

would need to be pulled to the forefront of the mix by dipping the low and

high frequencies, whereas in hip-hop the bottom end is bigger and the top end

is more aggressive, and thus, the mid-range is usually attenuated.

As with mixing, the key behind setting a good tonal balance with EQ is devel-

oping an ear for the different frequencies and the effects they have, but it’s

worth employing a spectral analyser so you can also see how the movements

PART 3

Mixing & Promotion

406

are affecting the overall balance of the music. Additionally, makes an analyser

it possible to measure the frequencies of a professionally mastered track and

then compare these with the results you have on your own mix.

A general guide to mastering EQ frequencies:

0 –40 Hz

■ All these frequencies should be removed from the mix using a high-

shelving ﬁ lter.

40 –200 Hz

■ These frequencies contain the bottom end of the mix and are not usually

attenuated or boosted. If this range has no punch or depth, or appears

‘ boomy ’, then it is common practice to use compression to cure the

problem.

100 –400 Hz

■ These frequencies are often responsible for a ‘muddy’ sounding mix. The

effect can be reduced by cutting the gain.

400 –1.5 KHz

■ These frequencies can contribute to the ‘body’ of the sound, and both

compression and EQ gain can be used to introduce punch and impact to

the mix.

800 –5 KHz

■ These frequencies often deﬁ ne the clarity of the instruments. Perceived

increases applied around here will occasionally pull out the fundamen-

tals of many instruments, making them appear more deﬁ ned.

5 –7 KHz

■ These frequencies are usually responsible for sibilance in a recording and

can be removed with a parametric EQ set to a narrow bandwidth to cut

the problem. Alternatively, compression can be used.

3 –10 KHz

■ These frequencies account for the presence and deﬁ nition of instru-

ments. Introducing perceived 1 or 2 dB increases here can augment the

presence of some instruments, while cutting can reduce the effects of

sibilance or high-frequency distortion.

6 –15 KHz

■ These can contribute to the overall presence of the instruments and

the ‘air’ surrounding the track. Introducing small gain boosts here will

increase the brightness but may also boost noise levels.

Mastering

CHAPTER 20

407

8 –15 KHz

■ Creating a boost at these frequencies will add sparkle and presence to a

mix (although it will also boost hiss if there is any), while cutting them

will help to smooth out any harsh frequencies.

20 KHz and above

■ This is beyond the limit of human hearing, but setting a wide Q of

around 1 and boosting (not perceived boosting!) by a few decibels at this

frequency will create the Baxandall curve, creating a more ear-pleasing

effect on the mix ( Figure 20.1 ).

Naturally, these are just guidelines, and in the end you should rely on your ears

and judgement as to what sounds right for your mix – provided of course, that

you have a pair of monitors that give an accurate ﬂ at response, a room with a

perfectly controlled response, an excellent EQ unit, an independent viewpoint

on the mix and a pair of ‘golden’ ears.

Thankfully of developments in computers and associated processing power,

more manufacturers are taking advantage and are releasing units that will aid

with balancing a mix properly. One such product is HalBal, currently only

available for the PC platform. This software analyses the music and shows the

frequency response in a graphical form that you can then edit easily.

MASTERING REVERB

Mastering reverb (applying reverb across an entire mix) is rarely required in most

mastering situations because it should have been applied properly during the

mixing stages, but in certain cases it may still be required. Typically, this is when

FIGURE 20.1

Spectral analysis of the

results of mastering EQ

on dance music

PART 3

Mixing & Promotion

408

instruments don’t seem to gel with one another, creating a series of holes in the

mix, or when the mix seems to be lacking the fullness of other similar mixes.

Although from an ideal viewpoint you should return to the mixing stage and

repair these problems, it may not always be possible, so lightly (and carefully!)

applying reverb over the entire mix can help to blend the music together to

produce a better result.

Of course, the amount of mastering reverb that should be applied is totally

dependant on the mix, the spacing between the instruments, and the amount

of reverb that has been applied to each during the mixing stage, but there are

some general guidelines that can be followed.

First and foremost, the reverb must be of a very high quality as any tail noise

created by poor algorithms will only serve to destroy the mix further. Indeed, if

you don’t have access to a good reverb unit or plug-in, it is not worth applying

because it isn’t worth turning an average mix into a poor one just for the sake

introducing some cohesion between the instruments.

We’ve already looked at the uses and parameters of reverb in a previous

chapter. Essentially the same methods and ideologies also apply during the

mastering process, although a little more prudence should be exercised. Also,

unlike all other effects in the mastering chain, reverb rather than is an insert

as a send effect used because you don’t want to wash over the mix completely,

only add small amounts to the original signal. To start with a typical return is

around 15% wet signal against 85% dry, but as always this depends entirely

on the mix in question. Nevertheless, using this setting you can set the other

reverb parameters and then back off the wet signal until it suits the mix.

Following the mix fader, it’s worthwhile adjusting the decay time. Broadly speak-

ing, you should aim to set this so that the tail is little longer than the reverb that’s

already been applied during the mixing stage so that the instruments mix together

more coherently. Typically, for most styles of dance music this would equate to

around 2 –8 ms; any longer and the mix could appear too ‘washy ’ and undeﬁ ned.

This, however, is related to the room size, and as a general starting point a square

room (if the option is available) with a relatively small size to begin with will

inevitably produce the best results. It’s important to note that the size of the room

will also determine the diffuse settings you should use; for a small room this

should be set quite low so as not to create an inaccurate impression.

As you’re working with a stereo reverb unit, you need to emulate the diffuse

effect that results from the left and right reverberations reaching the ears at dif-

ferent times. If you set this high but use a small room, you would essentially

be creating the effect of a small conﬁ ned room with walls that are far apart.

Although on some mixes this inaccurate impression has been known to work

quite well, ideally it is best to keep small diffusion settings with small rooms

and vice versa. More importantly, any adjustments with the diffusion can have

a signiﬁ cant effect on the stereo aspect of the mix, so always check the mix in

mono too.

Mastering

CHAPTER 20

409

Applying reverb to a signal may also introduce or enhance any sibilance or

low-end artefacts. To remove these it’s preferable to use the reverbs low- and

high-pass ﬁ lters. On average, for most dance mixes a high cut-off applied at

2–4 KHz should prevent the introduction of sibilance, while a low cut-off set at

80–120 Hz will prevent the low body of the mix from becoming too muddy.

Possibly the most important aspect of applying reverb, though, is to contin-

ually bypass the effect to check that you’re not getting too carried away with

its application. Ideally, as with mixing, reverb should be applied so that when

it is present it isn’t particularly noticeable but when removed the difference

becomes immediately perceptible.

DYNAMICS

Following EQ (and perhaps mastering reverb), many engineers will compress

the entire mix in an effort to add volume, punch, clarity and possibly emo-

tion. This cannot be accomplished simply by strapping a standard compressor

across the entire mix, however, since the loudest parts of the track will drive the

compressor, forcing the entire mix to pump with each kick.

While all good dance mixes should pump with some energy, much better

results can be gleaned from using multi-band compression. Using these, it’s

possible to alter the dynamics at different frequency bands, which can result

in a clearer-sounding mix. For example, compressing the high end of a mix

heavily can ﬂ atten the sound of a mix, so it’s best applied lightly. On the other

hand, the lower ranges of a mix are usually compressed more heavily, not only

to even out the sound but also to help the mix pump along and add extra vol-

ume to some frequency bands of the mix.

This frequency-speciﬁ c volume is an important aspect of most dance music,

especially on the lower frequencies, and should not be underestimated. Almost

every record company will want their music to measure up to the competition

in terms of volume as it’s a long-proven fact that if you play the same musical

passage to an audience at different volume levels, most will prefer the louder

playback. However, while it would seem sensible to reduce the dynamics of a

mix as much as possible to increase the volume, there are limits to how far you

should push it.

A typical home hi-ﬁ system will reproduce about 80 or 90 dB of the dynamic

range, but this is going to be pointless if the music is squeezed into just a few

decibels. In fact we naturally expect to hear some dynamics when listening to

music, and if they are eliminated it isn’t going to sound real no matter how

good the hi-ﬁ system reproducing the music is.

Furthermore, while louder music will inevitably capture the attention of the

listener, it will only be for the ﬁ rst minute or two. While the track may appear

hard-hitting at ﬁ rst, if it’s too ‘hot’ then it can quickly become fatiguing to lis-

ten to and most listeners will turn off even if they like the track. In fact, any

PART 3

Mixing & Promotion

410

ruthless dynamic restriction is only really considered useful for playback on

radio stations.

All radio stations will limit the signal for two reasons:

(a) They need to keep the music’s level at a speciﬁ c maximum so as not to

overmodulate the transmission.

(b) They want to be as loud as possible to hold an audience’s attention.

This means that no matter what the input level is, the output level will always

be constant. The inevitable result is that even if your mix happens to be a cou-

ple of decibels louder than every other mix, when it’s played on the radio it

will be at the same volume as the played before or after.

Incidentally, some artists will argue that by compressing a mix heavily at the

mixdown/mastering stage the radio station cannot squash the signal any fur-

ther so there is no possibility of any broadcasters destroying the signal. While

there is some logic behind this, setting the average mix level too high through

heavy compression can actually create more problems because the station’s

processors might confuse the energy of the music with the signal peaks and

attempt to squash it even further, totally destroying what’s left of the mix. This

will seriously reduce the overall punch of the music, making it appear lacklus-

tre when compared to other dance records.

Consequently, you will often have to settle for a compromise between the out-

put volume and leaving some dynamic range in the mix. As a very general rule

of thumb, the difference between the peak of the kick and the average mix level

(known as the peak-to-average ratio) should never be less than 4 or 5 dB. This will

enable you to increase the overall gain while also retaining enough dynamic

variation to prevent the mix from sounding ‘ﬂ at ’ on a home hi-ﬁ system.

Alternatively, if you really want a loud mix with less punch for radio airplay, then

it may be worthwhile releasing two different versions of the record. The promo-

tional records can be compressed heavily for radio airplay, while the record store

version can be mastered with more punch for clubs or home equipment.

There’s more to multi-band compression than simply using it to add some

punch to a dance track, and a manipulation of dynamics around the vocals or

lead melody can introduce additional impact. This is often what gives music

that extra listenable edge, although acquiring this dynamic feel often requires

a fresh pair of experienced ears, too, as it involves listening out for the mood,

ﬂ ow and emotion it conveys.

As an example consider this rhythmic passage:

’ … … I love to party … … to dance the night away … … … … … I just love to party ’

Beats 1

234

1 2

3 4

Of course, this isn’t a particularly complex vocal line, but notice that the

accents occur on the downbeat of each bar. If you were to apply compression

Mastering

CHAPTER 20

411

heavily throughout the mid-range (where the vocals reside) then this vocal line

could easily become:

’ … … I love to party …to dance the night away … … … … I just love to party ’

Beats 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4

By compressing the vocal area hard throughout, you have lost the dynamics of

the track and the resulting feel has been totally removed. However, by carefully

compressing and only applying light amounts to speciﬁ c parts of the vocals

you could get this result:

’ … … I love to party … … to dance the night away … … … … … I just love to party ’

Beats 1 2

234

1 2

3 4

As the accents show, you’re not just working on the vocal but the mix that’s also

behind it, so you’re also affecting the music. This can help to add a dynamic

edge as the music will move dynamically with the vocals, enhancing the gen-

eral feel and emotion of the song.

When you are working with such precise dynamic manipulation, you must set

the compressors parameters accurately and carefully. If the release is set too

long the compressor may not recover fast enough for the next accent or sub-

accent, while if it’s set too short, the sound could distort. Similarly, if the attack

is too short, then the initial transients of the word and music will be softened,

which defeats the whole point behind applying it in the ﬁ rst place.

More importantly, any compression will result in your having to increase the

make-up gain on the passage being compressed, and as we’ve discussed, loud-

ness can have an adverse affect on our judgement. Consequently, even applying

accent compression in the wrong place will lead you to think it sounds better,

so it’s important to keep the gain the same as the uncompressed passages to

ensure that it isn’t just pure volume that’s making the passage sound better.

When it comes to setting up a multi-band compressor for a mix, there is no set

approach and there are certainly no generic settings. It depends entirely on the

mix in question. All the same, there are some very general guidelines you can

follow to initially set up a compressor across a mix.

■ On a three-band compressor, start by setting each band of the compres-

sor to 2:1. Then adjust the frequency of each band so that you’re com-

pressing the low range, mid-range and high range.

■ On a four-band compressor, adjust the crossover frequencies so that one

band is compressing the bass frequencies, two are compressing the mid-

range and one is compressing the high range.

■ On a ﬁ ve-band compressor, adjust the crossover so that two bands are

concentrated on the bass, two are compressing the mid-range and the

ﬁ nal is compressing the high range.

PART 3

Mixing & Promotion

412

The crossover frequencies will depend on the mix being squeezed, but try

to ensure that the vocals are not compressed with either the low- or high-

frequency ranges of the compressor. By doing so, you will be able to adjust a

single (or double on a ﬁ ve band) compressor concentrated on the vocals alone,

to add the aforementioned dynamic movement to a mix.

Applying this multi-band compression to parts of the mix will upset the overall

balance of mix. Rather than start EQ’ing again, adjust the output gain of each

compressor to try to gain a more level mix. It’s important at this stage to real-

ize that overall track volume shouldn’t be an immediate concern. Ideally, you

should be aiming to produce a mix that has plenty of bottom-end energy with

a well-balanced mid and high range.

If the bass end of the mix is short of any bottom-end weight, then it is worth-

while lowering the threshold further and raising the ratio until the bass end

comes to the forefront. It should be noted, though, that the mid-range may

simply need less compression to lift up the bass and it’s a careful mix of all the

bands that creates a fuller mix.

For instance, if the mid-range is compressed heavily it will be pushed to the

front of the mix, overtaking both the high and bass ranges, so rather than com-

press the bass even more it may be more prudent to lower the compression on

the mid-range. As a general guideline, for most dance mixes, it’s usual to com-

press the bass heavily and use lighter settings on the mid-range because this

can help the mix appear less muddied.

Both the attack and the release parameters can also affect the overall tonal bal-

ance of a mix, so these must be adjusted with caution. Usually, on the bass

frequencies you can get away with using a fast attack time, as it consists purely

of low-frequency waveforms with few or no transients. Also, the release setting

should be set as fast as possible so the compressor can recover quickly, but you

may ﬁ nd that it has to be set longer than both the mid and high range due to

the waveform lengths of the bass frequencies.

Generally, the faster the release is set the more the bass end will pump, so it

should be set so that it pumps slightly but not enough to destroy the feel of the

mix. For the mid-range, it is sensible to use longer attack times than with the

bass to help improve the transients that are contained here, while the release set-

ting should be set as short as possible without introducing any gain pumping.

Determining pumping at mid-range frequencies can be difﬁ cult, so it is prefer-

able to use a compressor that allows you to solo the current band to hear the

effect you’re imparting onto the mix.

Finally, the high range should use a relatively fast attack to prevent the tran-

sients becoming too apparent, and as the waveform of these is often consid-

erably shorter than lower frequencies you can also get away with using short

release times. As with the mid-range, however, do not set this release too short,

otherwise you may introduce pumping which will ruin the overall dynamics

contained here.

Mastering

CHAPTER 20

413

As a very general starting point, what follows is a list of typical settings for

three-, four- and ﬁ ve-band compressors.

Three-Band Compression

BAND 1: TO TIGHTEN UP THE BOTTOM END OF A MIX

■ Frequency: 0 –400 Hz

■ Ratio: 4:1

■ Threshold: 3 dB below the quietest notes so that the compressor is per-

manently triggered

■ Attack: 10 –20 ms

■ Release: 140 –180 ms

■ Gain make-up: Increase the gain so that it is 3 –5 dB above pre-compression

level

BAND 2: TO TIGHTEN UP AND ADD SOME ‘ WEIGHT ’ TO THE MIX

■ Frequency: 400 Hz–1.5 KHz

■ Ratio: 2:1

■ Threshold: Just above the quietest notes so that it is triggered often but

not continually

■ Attack: 10 –20 ms

■ Release: 100 –160 ms

■ Gain make-up: Increase the gain so that it is 1 dB above pre-compression

level

BAND 3: TO INCREASE CLARITY OF INSTRUMENTS AND REDUCE

‘ HARSH ’ OR ‘ ROUGH ’ ARTEFACTS

■ Frequency: 1.5 –15 KHz

■ Ratio: 2:1

■ Threshold: Just above the quietest notes so that it is triggered often but

not continually

■ Attack: 5 –20 ms

■ Release: 100 –130 ms

■ Gain make-up: Increase the gain so that it is 1 dB above pre-compression

level

Four-Band Compression

BAND 1: TO TIGHTEN UP THE BOTTOM END OF A MIX

■ Frequency: 0 –120 Hz

■ Ratio: 4:1

■ Threshold: 3 dB below the quietest notes so that the compressor is per-

manently triggered

■ Attack: 10 –20 ms

■ Release: 150 –180 ms