How to get the Perfect AI Master...

Suno can make mastering easier even if you don't have great ears.

Mastering is one of the hardest pieces of the audio chain. You used to need great ears trained to listen to 1000 things at once before AI came into the picture. No… I’m not talking about the AI mastering services that have been around for years. Oh no!!!

With the release of their new v4.0 Suno unveiled something magical called “REMASTER”. Their marketing tells you that it’s all about bringing things up to par in v4.0. However, that is NOT the best way to use REMASTER. Instead of just remastering once you should remaster three times so you get six tracks to choose from. Then you download the original and the best two remasters to finish up in your DAW. Wait!!! - I’m getting ahead of myself…

A Brief History of Mastering

Mastering is the last thing that you do to the audio before you push it out to the masses. Therefore, it’s purpose has changed over the years. Back in the early days of stereo on vinyl they could get into a problem where the V cut bounced the needle too hard and caused a skip due to bass buildup. It was especially bad if the Bass wasn’t centered.

Then they fixed that problem and disc jockey’s were slapping vinyl down for radio play. But if your song was quieter than the previous it felt less energetic. So they started creating a loudness race in the 70s and 80s. By the 90s every radio station had a leveler to keep things basically at the same volume. So the loudness wars were less prevalent on radio after that. But then came the MP3s and it started all over again.

The problem is that if you crank a sound too much it distorts the speakers. Eventually it mathematically can’t increase any more and the sound clips causing rather nail biting distortion in the upper mid-range. Obviously this becomes a problem for the streaming platforms. So just like the radio stations before them they started creating leveling compression algorithms that they use to finesse the best sound.

That has lead to standards of -14 LUFS with -1 dBFS which remains the standard of loudness we try to go for. But that can be achieved by either light compression or in many cases just turning down the mix. So for anyone distributing through streaming platforms you can mostly let their internal stuff bring things up to standard as long as you don’t crank it much past -1 dBFS. If your eyes are getting glossy because I’m getting too technical don’t worry I’m done…

So why do we need mastering today?

I’ve told you the brief story above so that I can illustrate how mastering has changed over the years. But the one constant is trying to make the audio sound as polished as possible. To do that and get a richer sound most engineers do double tracking. It’s frustrating to lay down and it’s even more frustrating to mix. That’s why there are so many doubling plugins and stomp boxes now. Let me just say, it was a pain!

The other thing that we ended up trying to do is create subtle differences with the stereo image between the left and right ears which results in the track sounding bigger with more depth. And we would usually drive the transistors in a compressor, mixing desk, tube amp or even the physical magnetics of tape to add harmonics to glue everything together. But when recording went digital in the 90s mixing engineers stopped giving it extra harmonics. That’s one of the reasons modern music has harsh highs…

As a result of all of this we forced then need for Character Mastering or Stem Mastering which both amounts to the master engineer applying the final lay-up of glue to make everything sit in the mix. A lot of the final mastering devices created in the last 30 years are to bring character back that was previously put in by consoles and tape machines. So now instead of taming harsh frequencies or avoiding a needle bounce in the vinyl groove or trying to make our songs louder than our neighbor we are back to the heart of mastering which is the polish of audio.

How does Suno help us then…?

Remember those three audio files? It’s as simple as putting the original and two remasters and mixing them more or less left right and center. Then you just apply a little bit of compression and light reverb separately to each and you end up with an amazing mix every time. These tracks are basically in phase with one another but they do have micro timing variations so everything is double. You can do this with individual stems or with the whole stereo mix. But don’t go crazy…

One time I tried to mix 12 remasters together and it became difficult to tell which sound source produced what. Really 3-5 tracks is all that is needed. (I often add some extra vocals split out…) For the best results mix one side watching that your master doesn’t bump past the -1 bBFS mark. Then you mute that side and do the other. You should be able to get a sound fairly close in about 30 minutes after a little practice.

Remember I mentioned reverb? For some reason Suno always does things more dry than you actually want. You don’t need a lot. Reverb is like salting food - a lot goes a long way. But you can use the free stuff like Galatic or the Plate from Air Windows to get things quite nice. (Dampen the highs so you don’t get too much build up in that range…)

So there you have it.

This gives me a better results than all my outboard gear in less time.

Let me know what your experiments return…

Reply

or to participate.