I saw a post recently that said, “Stop comparing your mixes to industry mixes.” At first I didn’t totally agree with it, because I think referencing great records is one of the best ways to get better at mixing. But the more I thought about what he was actually saying, the more I liked the idea.

When you compare the record you made in your bedroom or studio to a major label release, you’re not really comparing one mix to another mix.

You’re comparing your work to an entire production system. There might’ve been an amazing recording engineer, vocal producer, producer, mixer, mix assistant, mastering engineer, A&R, artist, manager, and probably a bunch of other people listening and giving notes along the way. By the time you hear the finished record, it has gone through a lot of really good ears.

That doesn’t mean you shouldn’t reference it. You absolutely should. Great records can teach you where the vocal sits, how much low end you can get away with, how bright something can be without becoming annoying, how loud a record can feel while still having punch.

But there’s a difference between referencing someone else’s record and trying to turn your record into theirs. I think this distinction is going to matter even more as AI gets better. We’re quickly getting to a place where almost anybody can make something technically “correct.” You can have AI help write the song, make the beat, tune the vocal, mix it, master it. Eventually, technical competence is probably going to become pretty cheap. Taste won’t.

Your decision to make the vocal a little too loud might actually be what makes the record cool. Maybe your mix is darker than the reference. Maybe the bass takes up way more space than a mastering engineer would normally want. Maybe there’s something technically wrong with the mix that emotionally feels right.

That’s your taste. It should be ok sometimes for that to be the final record.

For years mastering has had this mystique around it, like you send your mix off to someone with better ears and better equipment and they somehow fix everything you did wrong. Obviously, a great mastering engineer can hear things most people won’t hear, and there are absolutely mixes that need help.

But I don’t think the goal of mastering should be to give you a different version of your mix.

The mix already contains thousands of decisions. You decided how loud the vocal should be. You decided how hard the drums should hit. To me, mastering should ask, "What is this mix trying to do, and how can we help it do that better?"

Maybe the low end needs to feel a little fuller. Maybe the top needs to open up. Maybe there’s a frequency masking the vocal. Maybe you can get another couple dB of level without killing the punch. Maybe nothing needs to change except the level.

The point isn’t to make the mix more “correct.” It’s to make sure all the frequencies and dynamics that communicate the emotion of the record are full, present and translating the way they should.

That philosophy is a huge part of how we think about HyFi. We don’t want HyFi to remix your record. We want it to understand what’s already there and finish it. And when we do A/B tests against professionally mastered records, that’s really what I’m listening for. Not whether HyFi made the most dramatic change. Sometimes the best master is the one where you barely notice what changed until you bypass it. The record just feels finished.