What Mastering Can’t Do
Two orders came in last week asking for things a master can't deliver. Both customers got the same straight answer: here's where the limit is, and here's what I can do inside it.
I think that answer is worth more than any before-and-after clip I could post. A before-and-after shows you what I did to one song. The limit tells you what I can and can't do to yours, and you should know that before you pay anyone to master anything.
"Put each instrument in its own space"
A customer last week sent a finished stereo mix and asked for a wide, immersive image. Every instrument in its own space. Real front-to-back depth.
It's a fair thing to want. Most of it isn't a mastering job.
A stereo master works on one finished file. Everything in the song is already printed into two channels. Mastering can bring forward separation that already exists in the mix and make it easier to hear. It can't pick up the guitar and move it somewhere else, because by the time the song is a stereo file, the guitar is welded to everything around it. Repositioning an instrument means going back to the stems and placing it again. That's mixing, and it happens before the master.
If you want that kind of placement, the work starts with stems: the vocal, drums, bass, and everything else as separate files. Depending on how your song was made, you may have them or you may not. If all you have is the stereo file, then the stereo file is the song, and the job is to make the most of what's already in it.
So here's what I actually did inside that limit.
I used Mid/Side EQ on the Side channel only. The Side signal is whatever differs between left and right, which is where the edges of the stereo field live. Shaping it changes how the outer parts of the mix sit without touching the Mid, and the Mid is where the vocal, the kick, and the bass are. Those stayed planted in the center.
I left the stereo enhancer off on purpose. The EQ was already doing the widening, and putting an enhancer on top of it would have been width stacked on width. I added no reverb and no delay either. Ambience added at the master stage doesn't create depth between instruments. It washes over all of them at once, including the parts that were already clear.
The finished master came out 3.9 LU louder than the source, and the loudness range didn't change. The song got louder without getting flatter. The quiet passages still sit below the loud ones by the same margin they did when the file arrived.
Width is the easiest thing to overdo on a master. It sounds exciting on the first pass through headphones. Push it too far and the master starts to sound processed, the center thins out, and the whole thing collapses badly the moment someone plays it in mono. A single smart speaker on a kitchen counter is mono. So are plenty of phone speakers.
The customer got the width that mix could honestly support.
"Send it back with and without vocals"
A new customer asked for their songs mastered two ways: once with vocals, once without.
Mastering doesn't remove vocals, and it isn't stem separation. Whatever WAV you send is what gets mastered. If the vocal is in the file, the vocal is in the master.
If you want an instrumental master, make the instrumental version first and send that WAV. It gets mastered as its own track. That's the right way to do it anyway, because the instrumental is a different balance. The space the vocal used to fill is empty now, and the master should be built for what's actually in the file.
I don't fault anyone for asking. It's a reasonable assumption if nobody ever explained where mastering sits in the chain. The song gets made, then mixed, then mastered, then released, and each step works on whatever the step before it handed over. Mastering comes late in that line. It can only work with the file in front of it.
The rest of the list
The same rule covers the other requests I see. A mix balance problem, like a vocal buried under the guitars or a bass that's too loud, is a mix fix. Every master includes one revision, and a revision adjusts the master. Here's where I draw that line in practice. If the whole master feels a little bright, or the low end needs to come down across the song, that's a revision, and it's mine to fix. If the snare needs to come up and nothing else should move, that's the mix. On a stereo file I can't turn one instrument up without turning up everything that shares its frequencies. When a revision request turns out to be a mix request, I do what the master allows and tell you plainly where the limit is.
And if I listen to a file and find a problem mastering can't meaningfully fix, you hear that from me before any work starts.
None of this shows up in a before-and-after clip. You can't hear the enhancer I left off or the reverb I didn't add. Most of what you're paying a mastering engineer for is judgment about what a file can and can't become.
A mastering engineer who says yes to everything isn't listening.
Josh, Founder, JG BeatsLab
Want real ears on your track, including an honest answer about what the master can and can't do? Mastering for AI Music
