Audio engineering advice_

January 2023 V1.6

Conor Dalton's mixing advice:

Recurring mixdown issues i see regularly as a mastering engineer

Too much bass due to the mixing engineer using small speakers, and so having no clear way of
hearing low frequencies while they work as the speakers can't create the lowest octave(s).
Dominant Resonances: loud dominant single frequencies demanding your attention in a narrow
bandwidth. These can destabilize a mix, mask frequencies, and make for a frustrating sound. 
Phase problems due to too much wide panning and non time aligned instruments: use a
correlation meter to analyse the width of certain frequency ranges
Too much reverb usually due to the mixing engineer working in a room with speakers too close
together or with sub optimal monitoring conditions, or perhaps too many low and low mid range
frequencies in the reverb, creating a muddy washed out feeling that also contains resonances and
dominant frequencies that can start masking the dry direct instruments. 
One instrument dull in the frequency domain and lacking high frequencies: another instrument
harsh and bright, which leads to the mix having dull and harsh characteristics at the same time:
this can be hard to address at mastering as any attempt to x this problem will emphasise one of
these issues
Dips in the frequency spectrum overall or over emphasized frequency ranges leading to a
frustrated, weak unbalanced sound
Spikey hihats with multiple harsh dominant frequencies in the sensitive frequency range between
1kHz - 6 kHz
Non-balanced levels in terms of volume between the instruments: this is one of the most
important things to get right as if the bass is loud and the kick is quiet, the mastering processing
will have to compress the low end to rectify this which could be avoided simply but adjusting the
volumes of the instruments earlier in the processing.
Over-compressed material, if the mixing engineer tries to make the song "loud" at the expense of
punch.
“
TIP: You should have visual meters running on your
stereo output to familiarize yourself with how music
looks on a correlation meter and frequency analyser
and use volume matched reference songs to skip for
a few seconds to to reset your ears while you work.

Correlometer (free vst plugin) can visualize the width of the frequency spectrum
Span is a useful free tool to visualize the frequency balance of the entire mix, or individual elements. If you
hold ctrl and click it can solo frequencies, moving the mousewheel will increase or decrease the Q

What makes a mix translate well across many

environments? 
Clarity: the ability to identify the individual components of the mix, this is achieved by having the
instruments occupy their own space in terms of amplitude, frequency and phase. 
Technical issues are resolved: For example, cleaning up errors, unwanted noise, controlling
resonances, xing vocal pops and clicks, controlling sibilance on voice.
Time aligning the components of the mix as best as possible, so try and align the peaks and troths
of similar elements by moving the parts forwards or back in time by a few milliseconds, for
example the drums, you can also use a phase rotating plugin to tighten things further such as
Little Labs IBP or Voxengo Pha-979
Punch: Can you feel the energy of the transient of the kick and drums? Or have you sacri ced that
snap in favour of a warmer sound? The strike of the drums is important so that the dynamic energy
jumps out of the speakers, allow the transients to drive the groove. You can bring transient back to
over compressed material with a tool like Oxford Transmod.
 Depth: How close is the noise, is it forward or backward in the sound stage (determined by
volume of the sound in comparison to everything else). Use reverb/loudness/timing/width to
allow for a bigger overall image. The movement creates realism and a "stage", but don't forget
that your most important ingredients (vocals, lead synth, kick, bass) should feel center stage for a
con dent sounding performance.
Personality: a sonic identity that gives you a feel of the tone and personality of the noises. Texture,
tone and feeling that is unique to the noise. Is it clean, dirty? Raw, smooth? Is it saturated, dry, thin,
warm? Have some elements been doubled and detuned to ll a di erent octave? 
Balance: are the elements at a volume that allows them all to be heard without competing too
much and sticking out over everything else? Is something behind all the other instruments and
neglected so it becomes slightly inaudible? What about their frequency content, are elements
evenly spread among the spectrum? Is the frequency spectrum lled up with the ingredients of
the song or have you neglected certain ranges. 
Human emotional impact: Arguably the most important, and a combination of all of the above
categories. In human terms, how does it make you feel as a person? Can you connect with the
artists story? As Mandy Parnell once said: If you hear "Bob Marley - Jammin" while sitting at a
beach bar, your rst thought won't be "this song doesn't have a lot of high end frequencies", your
rst thought will most likely be "the song makes me feel good". How does a song make you feel
something? Well that's a whole other discussion! 

How to achieve powerful mixes:

It is a carefully tweaked combination of every decision made at every stage. It starts with the source
material: good quality source material is essential, whether that's samples, microphone recordings, analog
equipment, or VST instruments: the choice of sounds will determine how far you can take it.

Next is production and arrangement: are too many things playing at once? Are too many similar
instruments competing for space? Is there contrast in macrodynamics between the sections of the song? 

The mixdown: 
Techniques such as high pass ltering to remove non essential sub information on elements that
don't need it such as vocals, guitars, hats (only if necessary, since it could have a negative impact
as well due to phase shift with minimum phase EQ or pre-ringing on transients with linear phase
EQ), 
Stereo optimization (panning, widening, mid/side EQ), it's a good idea to not have too much wide
sub frequency information, or to have your main instruments sound like they are on the sides of
the stage
Fixing dominant resonances, with dynamic EQ for surgical precision centered around one
problematic frequency, or multiband compression if there are multiple resonances in a small
area. Dominant frequencies take away attention / space from the surrounding frequencies, so
they should be controlled, but it's not advisable to take a dominant frequency and make it a weak
frequency, but rather to make a dominant frequency step back to sound con dent but under
control.
Appropriate EQ, to balance the frequency spectrum. As a rough starting point Linear phase EQ is
potentially best for gentle wide Q boosts especially in the high end if you want to protect the
phase relationship while adding high frequency content, or minimum phase for surgical cuts on
transients or low mid range correctional work. Be aware of transient pre-ringing when using linear
phase EQ, be aware of phase shift when using minimum phase EQ.
Pay particular attention to the range between 1.5kHz and 6kHz, your ear is especially sensitive
here (see: etcher munson curve): if there are resonances here your song will feel harsh. If the
resonances are controlled here and this area is smooth and loud overall then the song will sound
big, clear and powerful
Contrast can be a useful mix trick, for example, a narrow verse and a wide chorus, or a quiet
breakdown which gives the rst kick drum back the impact it deserves. 
Automation is very important, we love (even subtle) changes. Good automation choices means
we can use less compression also. 
Too many dominant tones in the low mid range will make a song sound muddy, nd them with EQ
using your ears or use an analyser like SPAN and see what's going on there visually, are there big
boomy resonant notes? Are they in key or room resonances? What does the song feel like if you
reduce one of them with EQ (narrow Q), does it feel lighter overall? More open in the high end? 
Controlling the low mid resonances between 150Hz-500Hz can help make your track feel free
open and expressive, but if you lack too much low mids the song will not feel con dent or sincere
overall, we need this thickness. 
For vocal processing, controlling sibilance is essential, as harsh "SSSS" sounds can lead to a
stressful sensation, so using a de-esser or plugin like Soothe2 could be very helpful. Also
controlling 'plosives' that sound like bursts of air overloading the microphone. There are tools
from Izotope RX to address this. Vocal volume automation is essential too. 
Broadband dynamic processing if necessary (compression or expansion or a combination) to
control dynamics of individual elements, and also to control dynamics of groups (drum group,
synth group). Using a side chain to relax the threshold of compression on the lower frequencies
can be useful on mixbus processing or drum bus processing so the compression isn't always
triggered by the low frequencies only. 
Multiband processing can be used to x issues such as a build up of frequencies in certain areas,
for example a loud harsh violin in the high mid range in the second chorus, multiband processing
could tame that range and leave everything else alone (as with all processing only if bene cial).
Multiband processing is ideal to solve problems, rather than to shape your entire mix.
 Time based e ects like delay and reverb, to provide a sense of space: these should ideally be on
a send rather than on the individual channel. This way you protect the dry signal and compliment
it with a dedicated 100% wet channel (blended to taste) that provides you with a sense of depth
and space in a "room". 
Enhancing musical frequency areas or overtones, with saturation. This adds new richness into the
frequency spectrum. 

Below is a way to visualize

the e ects of saturation
Here is a 1kHz sine wave without saturation

and with (extreme) saturation (below):

If this is the results of saturation on a pure sine wave, then you can imagine how adding (subtle)
saturation into a complex signal will generate harmonics and richness in the frequency spectrum in a
potentially useful way but be aware of the risks too, such as rounding of the transients and losing
some of the accuracy of the source material. 

“
It's also worth mentioning here that it's very
important to have your corrective EQing BEFORE
saturation or colourful processing in the order of
your processing, as otherwise you are adding
harmonics to unwanted dominant frequencies or
frequency ranges, which can increase frequency
imbalance issues and make them even harder
(impossible) to remove. 
Setting levels
Setting levels takes some practice and experience, but it's one of the most important aspects of a
good mix. This determines how close the instrument feels to us on the sound stage. 
You could try the pink noise trick to see how that feels as a starting point for setting your levels: 
Use a [Pink Noise sample] and solo individual instruments while playing pink noise until you can
barely hear the instrument, then move onto the next instrument in solo, and do this with each
instrument until nally you mute the pink noise and you should be left with a roughly balanced
mixdown in terms of volumes. Please note this is a starting point, from here it's important to adjust
for taste to prioritize the important ingredients: it's welcome to be tastefully unbalanced,
prioritizing what is important. -18dB is a good starting point for the pink noise. 
4dB trick: turn an instrument up by +4dB or down by -4dB. What do you learn? If you turn it down
by -4dB and you can still clearly hear the instrument in the mix, then it was maybe originally too
loud. Same in vice versa

Headroom on your mixdown while you work:

It's important to allow for some headroom: say around -6dB for the loudest peak of the song
overall is a safe target. It doesn't have to be exactly this, that's just a rough gure.
Why? So the mastering engineer is dealing with dynamic audio that still has the transients intact
with no risk of clipping internally in plugins or during export while attempting to exceed 0dB. 
Turning down your master fader at the end will ensure that the signal is not clipping the output,
but some (older) plugins would still have the possibility of clipping internally while you work if you
are exceeding 0dB, so it's better to mix with headroom at the source than to turn down at the end.
This leaves the engineer with many more options when using compression and limiting, and
ultimately more control. 
Somewhere around -18dB Peak per individual channel in your software is a GREAT average level
for your individual channels while you mix. It assures clean and pristine sound and compatibility
with all plug-ins and gear and will allow for plenty of headroom on the overall mix.
Turn up your speakers, not your faders. Loudness comes later during mastering. 
Below is an example of a dynamic mixdown with headroom that would be suitable to send to the
mastering engineer
Gain Staging
With every change you make in the mix, you should volume match the output of the processor to
match the same level as the input so the source material leaves the processing at the same perceived
volume as it enters the processing, this is gain staging. This way you don't start eating into the
headroom of the mix, and when you turn the processing on and o you can hear the e ect of the
processing without the deceiving volume change. This is also the only way to really tell if you made
something sound better or just louder, don't forget loudness can come easily at any stage with a
limiter: it's much more important to protect the songs integrity while you make your adjustments rather
than chase loudness with every step.
Recognizing dominant resonances and
controlling them with EQ
A resonance is a buildup of high amplitude frequency energy in a narrow bandwidth, causing a
note to sing / ring / dominate a frequency range. Reducing / controlling them is best done early
in the signal path so the problems don't in uence the other processors later in the chain

To nd resonances, sweep with a high Q on an EQ, narrow bandwidth, and boost by a lot, move
through the frequency spectrum slowly, what note rings / sings / jumps out LOUD? Maybe you
could do with less of one or two of these loud frequencies, so now take a narrow Q small cut out
of the frequencies that really stood out. Act like a surgeon: do as much as needed, but no more.
Alternatively use a spectrum analyser plugin like Voxengo SPAN with a slow release to nd
resonances visually, ctrl click to solo frequency and while doing so use mouse wheel to adjust Q.
EQing in key can be useful, using an EQ from DMG Equilibrium or using Fab lter Pro-Q3 would
make this easy in the digital domain. You can use a musical keyboard app for playing piano notes
on your phone (for example "virtuoso") to nd out the notes in the song and keep that in mind
when EQing, snapping the nodes to the musical note using the musical keyboard at the bottom of
the EQ for extra musical accuracy and relevance
Room resonances are harder to nd this way however as they won't have predictable musical
key. 
To control resonances use static minimum phase or linear phase EQ with narrow Q and cut the
o ending frequency to taste if the resonance is always playing, -3dB is a starting point, but it
depends on the amplitude of the resonance.
Use minimum phase EQ if the resonance is on a transient to avoid pre-ringing, also best suited for
low mid resonance control usually. 
Or use dynamic EQ (linear or minimum) to reduce the resonance if it only occurs every so often,
this way you only target the frequency when it exceeds over the loudness threshold and don't
permanently cut into the frequency even when it's not problematic.  

“
Too many resonances prevent you from reaching the
overall perceived power of a balanced song, due to
build up of frequency concentrated in small areas and
not spread out evenly, these high amplitude areas hit
the limiter's threshold alone, and not as part of a team
in balance with all the other frequencies.
Rough example guideline for stereo placement

Above Image from Masteringthemix is a good example, but remember the choice of panning is a personal
one rather than a strict formula, experiment and trust your instincts. It’s ok to not follow this example. 

Use Voxengo MSED (free) to check your mixdown in mono. Does any instrument disappear? This
shows you that there is a mono compatibility issue due to excessive panning / wide source
material. 
Also it's wise to use this plugin to check the sides. Does something stand out as problematic?
Wide sub frequencies for example can give you an unsettling feeling and these can be usually
reduced to prioritize central sub frequencies instead. 
 Mixing Bass

If high pass ltering to remove the lowest sub information or remove DC o set, some cut-o
resonance (boost before the cut) is worth trying out to help it to sit a little nicer in the mix and cut
through. 
Usually a 12dB per octave roll o low cut works best in my experience but there are exceptions.
Experimentation is always advised. 
Too much sub and the whole track feels nervous and frustrated, sub under 20hz is pretty much
useless as sound systems can't recreate it, but EQ low cutting in minimum phase mode can
actually increase the peak level of the instrument due to the Gibbs e ect, or cause pre-ringing in
linear phase, so only cut if you need to get rid of these useless frequencies, otherwise the bass
can be left alone to avoid those destabilizing artefacts.
Look for resonances in the low midrange of the bass: does one note jump out louder than all the
others? If so it could do with some EQ control: Does it happen once every 20 seconds (then use
dynamic EQ), or does it happen constantly (then use static EQ)? 

Above is an example (not a guideline that works for everything, but just an idea)
 Sometimes if you want 'more bass' in the mix, you actually just want to hear it cut through,
meaning its the mid range of the bass that needs a boost, rather than the sub. This could be done
by EQ boosting, saturation, or even opening up the stereo eld of the bass for the higher octaves
on the bass over 200hz. This could also help it translate on smaller speakers that can't produce
the lowest octaves.
Setting the bass with a healthy relationship in terms of volume to the kick drum is essential. One
shouldn't overpower the other but both are very important.  If there is a clash, you can use side
chain ducking on the bass (triggered by the kick: fast attack fast release), or you could take an EQ
cut out of the bass where the kick drum fundamental frequency occurs (or vice versa).

“
Creative tip: take a bass, and duplicate it. Leave
'version A' alone, and on 'version B' smash it up with
distortion and overly aggressive limiting, blend B in
a tiny bit to compliment A, if you want rich
harmonics without losing the original feeling of the
source material. 

This video from Dan Worrall is excellent in discussing bass mixing challenges.

https://www.youtube.com/watch?v=1xPO2Q2QHXk
 Compression / Expansion: a
rough practical guideline

Downwards compression: Reduces amplitude of the loudest events in the source material
Upwards compression: Increases amplitude of the quiet events in the source material

Both compression types reduce dynamic range (e.g. distance between the shouts and the whispers)

Upwards expansion: Makes the loudest events louder (transient enhancing for example)
Downwards expansion: Makes the quietest events quieter (de-noise, de-hiss plugins for example)

Both expansion types increase the dynamic range.

“
Make sure you use the makeup gain to compensate
for the gain reduction when using downwards
compression
i.e. If downwards compression dynamically reduces the signal by -4dB gain reduction on the loudest
peaks, set the makeup gain to +4dB. 

This is the ONLY way you can hear the e ect of the compressor and not the deceiving volume change. 

Set a tiny loop of a few seconds and watch the peak level on the meters with the compressor on and
with the compressor o , they should be the same value. Some compressors have an equal loudness
bypass feature which does this automatically so you can compare before and after processing
accurately. 

For live recorded drums it's important to grab the initial transient which can often be way louder
than desirable for a balanced mixdown, using something like an LA2A or 1176 is common (linked
are free vst versions).
For sampled drums or loops, less compression is needed usually, a lot of processing could
already have been applied. 
For drum busses a slower attack and fast release for snappy energetic sound works well, or if
you want a warmer sound you can make the attack a little faster at the expense of punch.
Pads, Slow attack and slow release for a gentle reduction in the distance between the louder and
quieter sections can work great. 
Synths (quick snappy synth) Slow attack fast release 2:1 taking o 4dB or so and compensating
with makeup gain might be a regular place to begin with your compression settings, pushing the
sustain of the noise upwards and allowing the punch to exist untouched. Adjust to taste
Vocals, medium slow attack, medium release, about 6-8dB gain reduction is a common starting
point, making sure all words can be heard and understood clearly over the music and the
whispers don't get lost in the noise, and the shouts don't stick out too much. However with vocals,
it will also be important to automate the volume for individual phrases, words or syllables that get
lost in the noise, simply compressing won't x everything, vocals need extra attention as it's the
instrument we are the most familiar with and we know when a voice sounds "o " very quickly. It
could be useful to use two compressors for a vocal, one to catch quick peaks and one to control
the overall body. 
Bass: if certain notes are jumping out a lot louder than others use a compressor to reduce the
loudness di erence between the notes. This provides stability in the low frequencies which can be
very important. 
It's also sometimes necessary to side chain compress a bass, triggered by a kick for example,
sometimes this can rescue a relationship between kick and bass that has problems with the
groove. When doing so, a fast attack and fast release is essential.  
Multiband dynamic compression / expansion
Use to x amplitude issues in certain frequency bandwidths, for example a buildup of loud high
midrange frequencies in the chorus of the song, or a lack of transient snap in the high frequencies
Slow attack, fast release if you want to maintain drum punch, or fast attack fast release if you
want to atten and warm things up at the expense of the transient attack. 
Also suitable for smoothing out a high end volume imbalance issue, such as the hat feels harsh
and loud, yet there isn't much air overall on everything else. You could have the multiband
downwards compressor reduce the volume of the hat, then adjust makeup gain to turn up the
high end overall so nothing is jumping out alone but the high end is appropriately loud and
consistent in general. 
Don't use if you don't have a speci c reason to do so: this is a problem solver, not a default.
Overuse can make a track feel unnatural and over processed, especially if all the bands have
di erent attack / release and ratio settings

Parallel processing:
Taking one instrument, duplicating it and heavily over-processing the duplicated version with
creative compression and/or distortion and blending it in with a little bit of volume to tuck it in
behind the unprocessed version. Don't over-do it or use on every aspect of the mix.
Can be incredibly useful and versatile when dealing with drums, bass, or vocals especially. 
Can cause phase related artefacts so be careful of this, if in doubt check with a correlometer.

Image from Producelikeapro

 Limiting (perhaps most useful
at the mastering stage)

Limiters are compressors with an in nite ratio and they are experts in peak management. 
They go last in the chain (followed only by dither to move from one higher bit depth to a lower bit
depth)
Limiting is best left to the mastering stage, limiting should be used to protect the song from peak
distortion of transients trying to exceed 0dB, and also it's used to reach a speci ed level in terms
of loudness, but the limiter should never be used as the only way to achieve loudness (sonic
integrity shouldn't be sacri ced to achieve loudness) 
Nothing passes beyond their ceiling, except intersample peaks. 
Oversampling can help reduce intersample peaks
True peak limiting can also help reduce artefacts at the expense of a little less perceived loudness.
You can unlink the channels so it treats left and right di erently for a more open sound, or link
channels for more controlled sound. 
-2 to -3dB gain reduction is maximum before audible distortion/damage begins to occur, usually.
Therefore if you are hitting -3dB gain reduction and it still isn't as loud as you desire, this is an
issue to x in the mix balance rather than in the master limiting. Go back and squeeze a few more
dB from overly dynamic tracks in your song to let the nal limiter work less. The goal is to never
rely on the nal limiter to do too much.
Soft clipping could be useful before the limiter, to quickly deal with stray events that have no real
emotional bene t to having ultra high peaks that stick out way above everything else. Controlling
those before limiting can help the limiter to relax a bit.

“
Louder is not necessarily better if you have to
sacri ce sonic integrity to achieve it. 
There will come a point where you start to degrade the audio by over limiting: the ONLY way to
tell when you have gone too far is by volume matching while adjusting levels. Same goes for
compression. Don't be deceived by a change in volume. Listen for the change in dynamics and
distortion artefacts cause by the limiting. You can only hear the e ect when it's volume matched
to before and after processing.

Processing on the stereo

output
Anything that you have applied for creativity on the masterbus, such as subtle EQ, gentle
compression (with slow attack preferably), colour / vibe / tape, can remain on your pre-master if
you feel it adds something to the sound that you like, but also send a version to the mastering
engineer without these e ects if possible. However it's usually best to mix in a certain way so that
processing on the masterbus isn't necessary but feel free to "break the rules" if it makes you
happy.
Strong compression or limiting (or even more destructive, clipping) on the masterbus should be
switched o before mastering because if you over-do these e ects it's impossible to undo.
Best to x EQ balance issues on the individual channel rather than on the stereo output in general
If you made a self master or limited version, also send this so the engineer can get a feel for how
you heard it while working and your desired loudness target. If you made an amazing self master
then you will have the full attention of the mastering engineer to try and improve on what you did.

The importance of
referencing

Referencing is the practice of comparing your mix to other songs in your genre at a the same
perceived volume while you are working
Your references should be high quality productions that you aspire to sound like
If you don't have anything to compare your mix to, how do you know if you aren't going the
wrong direction with your mix? If you want to be a good person you need good role models.
Same for mixing. Compare and contrast: 
Are all of your reference songs brighter than yours? Maybe your mix needs a high end boost then
Check the mixdown reasonably loud when trying to evaluate the bass and check it at a quieter
volume when trying to judge clarity and making sure all the elements can be heard
 Ear training resources: 

Soundgym
Train your ear to hear the frequency you want to address, 
Pensado's ear training
Learn how to listen 
We hallucinate our expectations into “reality”. Proof: here
A great website to do volume matched comparison tests of exported les

Strategy

It’s recommended to get your music sounding as close to nished as possible before sending it to
mastering. (minus the aggressive limiting for loudness). Don't leave it too open ended expecting
that mastering will x a bunch of issues and drastically change the tone.
Remember: a great mix = a much better chance at a great master.
Your mixdown should sound 'like a record' as much as you can. Mastering should improve it
further while respecting your intentions, but the mixdown is key to good results, so don’t cut
corners: if you can x something during the mix, do so. All of these little 1% improvements add up
to a big result. 
Make your mix with no regard for the overall volume, which comes later at the mastering stage.
 General Advice

Be aware of your room modes in your studio: use a sine wave generator to sweep slowly
through the low frequencies 20hz-200hz one at a time in your listening position. Does one
frequency completely disappear? Here you have a room mode cancellation. Does one frequency
sound way louder than everything else? Here you have a room mode peak. If you have crucial
mix elements (such as a kick drum) with strong amplitude at one of these problem frequencies
your room resonances could mislead you into boosting when you should be cutting and cutting
when you should be boosting.
Having good headphones to check can be useful here, as headphones have a single diaphragm
that remains the same distance from the ear at all times, so room resonances don't in uence your
decision making. Maybe your headphones can bene t from a corrective EQ curve too.
REW is free software for room acoustic measurement, loudspeaker measurement and audio
device measurement
Want to record through expensive hardware from the familiar settings of your own studio (in real
time)? This is also an amazing way of satisfying your curiosity to the age old question of "do you
need hardware to sound good?" (spoiler: NO). Hear for yourself.

Exporting the mixdown to

send to mastering
Sample Rate

Sample rate should be 44.1kHz or above (Keep the sample rate the same as you have been using
while working / recording, don’t upsample or downsample on export).
Bit depth 

32bit without dither (The best due to the additional headroom and lack of dither required)
24bit with at dither (ok but now we have to dither twice in total, once for exporting mixdown, once
for exporting master)
16bit with at dither (acceptable but not ideal, better as a nal container format)

Dither

If you export your mixdown in 32bit oating point, you do not need to apply dither. This is optimal,
as then dither can be applied only once, at the very end of the mastering chain.
Dither adds low level noise to our audio signal. Yes, we add noise on purpose, and it is bene cial.
How can adding noise be a good thing?!! We add noise to make a trade. We trade a little low-level
hiss for a big reduction in distortion when moving from a oating point bit depth to a lower xed
point bit depth container format such as 16bit. It’s a good trade, and one that our ears like. This
noise keeps the gate open while the calculations are occurring, it's an entertainment system for
bored bits.
If you are exporting a mixdown for mastering and can only supply 24bit (also ok) or 16bit (not ideal
but possible) then it's best to use a at/TPDF dither instead of a noise shaped dither. Flat dither is
best suitable if further processing will be applied.
Noise shaped dither is best used only at the nal mastering export stage as no further EQ will be
applied. 

File types
.Wav or .Ai les are ideal
Stereo interleaved
Mp3 is not possible to work with at the mastering stage, mp3 is a nal container format but not
suitable for further processing (also, try to avoid mp3 samples or youtube rips for your source
material as the quality is compromised)

File management / labelling

Make sure you label your mixes properly, this is much more important than you might think, so you
can easily understand what mixes were used for the mastering even years later. This also really helps
with archiving the project so it can be searched for later with ease.
Example of good le name labeling structure for digital / CD:
1. Artist Name - Track name mixdown v13 - 44.1kHz 32bit 
2. Artist Name - Track name (remix1) mixdown v2 - 44.1kHz 32bit 
3. Artist Name - Track name (remix2) mixdown v7 - 44.1kHz 32bit 

Or for vinyl with the running order and side splits added:
A1. Artist Name - Track name mixdown v13 - 44.1kHz 32bit 
B1. Artist Name - Track name (remix1) mixdown v2 - 44.1kHz 32bit 
B2. Artist Name - Track name (remix2) mixdown v7 - 44.1kHz 32bit 

Fades and trimming audio

It’s sensible to make sure that the start and end of the pre-master mixdown are the way you want
them to be, and not with too much silence before the rst note or sloppy fade outs at the end (or
the audio cutting out with a click is also a common mistake, as seen in the image below.)

A subtle fade at the end would avoid a click / unnatural ending

Communication with the mastering engineer
It’s sensible to explain to the engineer if you have concerns with your mixdown (”I just can’t get that
bassline to sit nicely”) and also tell them your desires for your master (”Please make it dynamic
and clear”). This makes it more likely that you will end up with results that you’re happy with.
Another approach that’s useful is to send the engineer your pre-master mixdown, and ask them
for some feedback / consulting on the mixdown before mastering begins. They might tell you
some useful information like “Turn down bass guitar -2dB, boost kick high end with shelf +2dB at
12kHz, and reduce a harsh frequency at 5543Hz on your hats with narrow Q EQ cut -4dB”. This
means you can quickly go into your project, make those changes to the mix and export a new
version that’s better for the engineer to work with without these pressure points.
Not all engineers provide mix feedback for free, so it might be worth asking if they provide this
service before you begin. At Glowcast there is an additional charge for this service as it takes
roughly 25 minutes of time. 
It's wise to make a rough reference self master version, and send that over too. This can help the
engineer better understand your desires (and also even help you become aware of mix issues
during your own quick mastering process).

Revisions
It’s recommended to also learn about the mastering engineers revision policy, in case you would like to
make adjustments to the master that they make. Some engineers o er unlimited revision until satisfaction,
whereas others charge money for revisions.

At Glowcast, revisions are free of charge on the original mixdown until the client is satis ed. 

Sometimes mastering can highlight an issue that you might not have been aware of previously, so there
are times you might want to make a new mixdown after hearing the original mastering. At Glowcast
new mixdowns are charged at 50%, as it requires a re-mastering process that takes almost the same
amount of time as mastering from the beginning.

Ready?
Then Upload Files + Order Mastering
About the author: 

Conor Dalton is a Berlin based mastering engineer at Glowcast and Calyx and lecturer at Berklee
Valencia, his credits include Daft Punk, Je Mills, Rodhad, Amelie Lens, John Digweed, Kobosil, Carl Cox
+ Nile Rogers, Fink, Robert Hood, Moderat, and many more.

Experience:

Joao Carvalho Mastering, Toronto 2009

Glowcast Audio Mastering 2009-Present
Published: Music Tech Magazine worldwide
Scottish New Music Awards: "Sound Engineer of the year" 2012
Calyx Mastering Engineer: 2013-Present
Lecturer at Subsine Academy, Glasgow: 2015-Present
Lecturer at Berklee College of Music, Valencia: 2018-present
Lecturer at Catalyst Berlin: 2018-Present
Musician in the band Island People (Raster)
Musician in the band Headless Chaos

Education:

Sound Training Centre, Dublin, Music Technology: Distinction

UWS BscHons Degree: Court medal for academic achievement
 Conor's recommended
plugins for mixing and
mastering:

Best free plugins: 

Analog obsession plugins, (vibe, character, functional, everything, essential!) 
Airwindows plugins, (vibe, character, functional, everything, essential!) 
TDR plugins, (free and amazing with option to upgrade to more features for a small amount of $,
ESSENTIAL, super clean) 
Acustica: Co ee ThePun (free amazing Pultec style, great preamp) 
Voxengo free bundle (worth having these versatile tools)
Quick Haas (free quick and easy haas e ect)
Tan2 (free compressor)

EQ 
MAutoDynamicEQ (excellent capabilities and interface, check out the sonogram, wow! perfect for
resonance control) 
Fab lter Pro Q 3 (excellent all round, for surgical or smooth, also has automatic gain
compensation) 
TDR Nova GE (excellent dynamic EQ) 
UAD Massive Passive (silky smooth) 
Acustica Purple (Pultec style) 
 Acustica Cream2 (curve bender emulation, classic, warm, sparkly) 
EQuilibrium - DMG Audio (one of the best there is, incredible) 
MAAT - ThEQOrange (a classic, a remodel of algorithmix plugin, linear phase, SUPERBLY CLEAN) 
Eventide - SplitEQ (control transient / sustain separately)

Compressors: 
Kotelnikov TDR GE (clean, free version also available with less features, world class) 
Weiss DS1 MK3 (based on the expensive hardware, hard / snappy)
UAD Manley Vari-Mu (character, soft) 
Unisum (incredibly transparent, world class) 
Cream2 (warm) 
Coral2 from Acustica (clear and controlled) 
Magic Death Eye - DDMF (vibe, snap, power)
Fab Filter Pro MB (Multiband Compressor, transparent, clean)
Waves SSL comp (punchy, good for control), 

Colour 
Acustica Celestial (SSL fusion clone) 
Acustica taupe (tape emulation, vibe) 
Oxford In ator (a classic exciter, still a winner) 
Waves J37 (vibey tape emulation) 
Decapitator from soundtoys (An absolute classic) 
Airwindows Density (colour, character, free) 
SDRR (distortion, be careful, strong character, not subtle) 
Airwindows ToTape5 (realistic tape emulation, free) 
Acustica Cream2 (colour, air, vibe) 
Tone Projects Kelvin (nice, smooth, useful)
Black Box (Sounds great, useful in a lot of ways)
Corrective 
Oxford Supressr HD (good for taming a harsh frequency range or harsh vocal) 
Oxford Transmod (helps x over compressed material, enhances transients) 
spi oeksound (transient design) 
soothe2 oeksound (corrective resonance control) 
Izotope RX suite, (to get rid of pops, clicks and crackles and restoration, world class) 
Softube Weiss Deess (harsh vocals or high hats controller) 
TDR Nova GE (free or upgrade for cheap, great!) 
Trackspacer (Side chain ducking sculpted to your frequency balance) 

Limiters / Clippers 
Weiss DS1 MK3 (also a deesser and compressor) 
Fab lter Pro-L2 (all rounder, clear interface) 
Voxengo elephant (reliable, clean) 
Sonnox limiter (aggressive, loud, good for dealing with subby and bass heavy material) 
Waves L2 (bassy) 
Waves L3 (linear phase multiband limiter, clean, good to use before nal mastering limiter, don't
overuse!) 
Newfangled Audio - Elevate (multiband aggressive limiter, a little unpredictable but sometimes
incredible) 
AOM Invisible Limiter (punchy, clean, clear, but high risk of distortion on the wrong settings) 
Standard Clip (great clean tool
DMG Limitless (multiband limiter, great, confusing interface but worth learning)

Dither: 
Airwindows "Notjustanotherdither" (use when you want to export 24bit, there is no graphic user
interface so when you put it on a channel it won't open, but it is working) 
 Airwindows "NotjustanotherCD" (when you want to export 16bit, there is no graphic user interface
so when you put it on a channel it won't open, but it is working)

Analysers 
Voxengo Span (free. When you hold ctrl and use the mousewheel and click it will solo a
frequency) 
Flux Pure (shows you where your resonances are, looks great) 
MAAT DR Meter mkII (good for getting a strong understanding of Lufs levels) 
Ian Shepherd's "Dynameter" (powerful visual tool to show you your dynamics) 
Youlean loudness meter 2 (Lufs, Free, ESSENTIAL) 
MAAT 2bus control (free, check out your mono compatibility, and sides) 
Voxengo Correlometer (check how wide certain frequencies are) 

Other tools: 
Ian Shepherd's "Perception" (Gain matching plugin, very useful for before and after comparisons) 
Letimix "Gainmatchcontrol" (Gain matching plugin, same as above except cheaper) 
TBProAudio AB LM (Cheap Gain matching plugin, very useful for before and after comparisons) 
MCompare 30€ (Cheaper Gain matching plugin, very useful for before and after comparisons)
HOFA 4U+ BlindTest (volume matched hidden comparison) 40€ 
DR meter MKII (good for streaming requirements for metering) 

Phase correction
Little Labs IBP Phase Alignment Tool UAD (essential, phase aligning tool) 
MAAT RSPhaseShifter (phase aligning tool)
Voxengo Pha-979
Voxengo MSED
Creative plugins for colorful mix personality:
Mishby by Freakshow industries. This company is hilarious and these plugins are serious 
Guitar Rig 6: wow, an absolute power house to bring life into something that lacks personality /
charm, this is instant avour and strength (Don't be fooled that it has the word guitar in the name,
this is powerful for way more than just guitars, it's useful for synths, vocals, anything that needs a
new personality) 
Portal: Granular synthesis with beautiful interface, simple to use, complex sound 
Lovend by analog obsession (free) for FAT bottom end that will hit you in the stomach like a train
Blackhole Reverb: big, liquid, dreamy, smooth, beautiful 
Backmask by freakshow industries: Dystopian destroyer
Sound toys (essential)
Audio Damage (essential)

All in one useful tool:

Izotope Ozone (all in one toolbox for audio mastering / mixing. (Some modules are great: Exciter,
Low End Focus, Master Rebalance) 

Want to learn more?

Private tuition
Would you like to learn more about mixing or mastering? Would you like to take your
mixing/mastering skills to the next level? Conor is o ering online engineering workshops either for
single students or small classes of 2 or 3 individuals. It consists of webcam, screenshare, studio cam
and high quality audio feed in 32bit PCM. This guarantees you hands on dedicated and intensive training
from your own studio. Please get in touch for more information.

Pre-recorded 4/5 hour video mastering crash

course
Rather learn in your own time? I am also o ering an intensive beginners course in audio mastering
complete with written notes, audio materials and 5 hours of personal video tuition from Conor Dalton
about everything you need to know to get started in your mastering journey. This video footage has
never been online and is only available from this site, the footage was recorded in 2021. Get in touch for
more information.

Global guest lecturing/mastering workshops

Conor is available to give extensive training on audio mastering to groups in a studio or classroom
environment. With experience in teaching at Berklee College of Music (Valencia, Spain), Catalyst (Berlin,
Germany), MI Hollywood (LA), and Subsine Academy (Glasgow, Scotland). Courses could be intensive
weekend crash courses or shorter guest lecturing. Please get in touch to discuss further details.

Keep in Touch here, socials etc