Professional Mastering - Analog & Digital
Professional Mastering
251K+ subscribers
A lot of people think mastering will fix a mix. It can certainly help fix some issues, but notentirely. It makes a well-balanced mix sound great, but can only make a poorlybalanced mix sound moderately better. So today I'll show you how to get the tonalbalance and the level balance of your mix right first, so that your mix sounds as good aspossible before mastering.
We’ll also cover some specifications like how much headroom is needed, the correct fileformat, and so on.
So, let’s jump into...
When a mastering engineer gets your mix, they're working on the whole stereo file. Ifthe vocal is 3 dB too quiet, or the bass and kick are fighting around 100 Hz, any fix theymake affects everything in that range.
For example, if you boost the vocal presence on the master, it’s possible that thechange brightens the hi-hats or maybe the overtones of the guitars, or whatever else isoccupying that space. That's why the best masters start from a mix that's alreadybalanced.
So, I’d recommend focusing on 2 concepts.
Level balance: how loud each part is compared to other parts
Tonal balance: how the frequency response of each part fits together.
When affecting level and tonal balance, you first need to:
Start by bussing everything into four groups: lead vocals and lead vocal doubles,melodic instruments, drums and bass. The melodic instruments bus is everything that’smelodic but isn't the lead vocal or bass, so guitars, keys, synths and pads.
Four groups is a great starting point. It's few enough that you can hear how they relate,and it's roughly how most mastering engineers think about a mix anyway. Dependingon your genre, less may be needed; for example, if the track is just acoustic guitar and vocals, but with a more complex genre, try to avoid more than 6.
Before touching anything, pull all four groups down so the mix peaks around -6 dB, give or take a couple dB.
That gives you room to work. Let’s briefly listen to the stems before we start making changes - notice that the mix sounds muddy and unbalanced at the moment:
Watch the video to learn more >
I recommend setting levels before you begin to EQ. A lot of what sounds like a tonalinconsistency is really improper levels. A dull mix often just needs a higher level setfor a group with high frequency content. A muddy mix is often a bass part that's 2 dB tooloud.
I recommend you pick your anchor first - this is the group around which you’ll balancelevels and tone. In most pop, hip-hop and rock, that's the vocal. In dance music, it'susually the kick and bass together. Set the anchor, then bring the other groups uparound it.
Once you’ve established some levels you’re happy with, do the following:• Monitor at a low volume: At a very low volume, you should still hear everyinstrument group. Whatever disappears first is too quiet.
• Check in mono. If a part vanishes in mono, it should be increased slightly tocompensate.
• Compare to a reference at the same loudness.
You could use a reference you import into the session. I’m going to use this plugin to import a reference and compare my mix to the reference. It’ll split the reference into the relevant stems, and then adjust the levels. For now, I’ll bypass the tonal changes it makes so we can hear the effect of level changes.
Watch the video to learn more >
Now we’ll EQ, but think in terms of collective balance. Keep any changes within 1-3dB.Here are the areas that typically require the most attention.
• Kick vs. bass, roughly 50–120 Hz. Decide which one owns the very low range -typically the kick will have the lower fundamental. Give the other a little roomabove it by attenuating on the competing instrument. For example, if the drums occupy the lowest range, dip a little above the fundamental for the bass to occupy.
• Mud, roughly 200–400 Hz. This builds up across every group at once. A smallcut on the melodic instrument group often clears more than cuts on each track.
• 2–5 kHz on the Vocal and Melodic Instrument Groups. If the vocal is hard tounderstand, try cutting a little in this 2-5kHz range on the melodic inst. group before you boost the vocal.
Keep in mind that louder almost always sounds better, at least initially. If an EQ boostraises the level, you may think it's better when it's just louder. Match the output levelafter every move so you're judging tone, not volume.
The plugin I’m using here affects the tonal balance via EQ for all four groups in one window and can level-match automatically. But this works with any EQ, aslong as you match levels yourself.
Watch the video to learn more >
Most people compare their full mix to a reference full mix. That tells you the overall toneis off, but not which part is off.
A better way is to listen part by part. How loud is their vocal against the drums? Howmuch low end sits in the bass versus the kick? Some tools can split a reference intostems and show you this directly, which makes it a lot faster.
One important note for picking a reference: find something with a similar arrangementand genre. It’s common to pick a song that’s of a similar genre, but be sure that thedensity of the reference’s arrangement is similar.
I’ll use a reference as a starting point here - then make some adjustments. If you use a plugin that makes automatic adjustments, keep in mind you may need to pull things back slightly - consider any automatic changes as a starting point and then make changes by ear.
Here’s how the track sounds now, with each bus equalized, the stem’s relative overall levels altered, and tonal eq changes.
Watch the video to learn more >
A static EQ may work well one moment in the song. If a part changes, like a vocal thatgets bright when the singer pushes, use a dynamic EQ on that range so it only workswhen it needs to.
Then, check the stereo image. Keep the low end centered. A gentle low cut on the sidechannel, between 40 and 100 Hz, keeps the bass solid in mono and on small speakers.
Watch the video to learn more >
A great mix exported wrong still causes problems.• Take any limiter or loudness maximizer off the mix bus.• Leave headroom: highest peak at -3 dB or lower.• Export at full session resolution, that being 44.1kHz, 24-bit or higher, and as WAVor AIFF. No MP3s or other lossy formats.• Don't normalize on export.• Leave a few seconds before and after so fades and tails aren't clipped.• Listen to the exported file once, start to finish, before you send it. Ensure that you’ve made all needed changes before sending it to theengineer.