The Master Where I Did Almost Nothing

A track came into the shop this month that tested part of my job. It arrived from a composer working on a large theatrical project, a full-length work with dozens of songs, and the source was better than almost anything that crosses my desk. It arrived dynamic and clean, with a dark, intimate character that was obviously deliberate.

The temptation on a track like that is to do something anyway. You are being paid. Visible work justifies the invoice. Every plugin in the chain is an argument for its own existence, and "I barely touched it" sounds, to the wrong ear, like "I barely worked."

Here is the entire master. A high-pass filter to clear subsonic energy no playback system was going to reproduce, the kind of rumble that eats headroom and gives nothing back. Two corrective EQ cuts, both smaller than a decibel. Compression light enough that it reduced less than a decibel at the loudest moments and did nothing most of the time. A loudness lift of less than one LU. That's it. The dynamics the composer built survived nearly untouched, because those dynamics were the point of the piece, and my job was to bring the track to a professional level without spending them to get there.

The cut you'll never hear

The detail I care most about is a move I didn't make.

There was a third EQ cut I evaluated, dialed in, and listened through several times. On paper it addressed a resonance. In practice, the resonance turned out to be character, not a problem. With the cut engaged the track sounded more correct and less like itself. So I turned the band off. It's still in the session, disabled, and it's written up in the client's Mastering Dossier with the reasoning attached.

That entry matters more than the moves I kept. Restraint documented is different from work not done. The client can see the option was found, tested, and declined on purpose, which is a different thing entirely from it never being noticed.

Restraint is the skill

Anyone can make a track louder. Software does it in seconds. And an algorithmic mastering service will always do something to your song, because doing something is what it exists to do. It has no setting for "this needed almost nothing," and it will never send you a note saying so.

Knowing when a track needs almost nothing, and being able to prove that with measurements instead of vibes, is what ears are for. If your track is genuinely good, a heavy-handed master makes it worse in exactly the places it was good. The dynamics get flattened. The character gets corrected away.

So here's a better question to bring to any mastering service, human or automated. Most people ask what will be done to their song. Ask instead how the engineer will know what not to do, and how you'll find out what was considered and rejected. If the answer is a shrug, you're paying for processing and hoping judgment comes with it. It usually doesn't.

Mastering here is $50 a track, flat, and every master comes back with the full Dossier, including the moves that were considered and declined. On this particular master, the Dossier was most of the product.

The best work on this track is a cut nobody will ever hear.

Josh, Founder, JG BeatsLab