Finding the Problem Isn't the Same as Knowing How to Fix It
You found the artifact. Now you've got four different tools that could theoretically address it, a static notch, a dynamic EQ, a low-pass filter, spectral repair, and reaching for the wrong one does real damage even when your diagnosis was completely correct. Knowing what's wrong and knowing which tool fixes it without collateral cost are two separate skills.
Static notch: when the problem holds still
A narrow bell cut at a fixed frequency and fixed depth, always on. This is the right tool when the artifact is constant, a resonance that rings the same way regardless of what's happening musically around it. Start narrow and shallow, 2 to 3 dB, and only deepen after confirming with an A/B that it's actually helping rather than just feeling like progress because you did something.
The failure mode: using a static notch on something that isn't actually constant. If the problem only shows up in certain sections, a notch that's always engaged is cutting real content the rest of the time for no benefit.
Dynamic EQ: when the problem comes and goes
A cut that only engages when the target frequency crosses a threshold, otherwise it stays out of the way. This is the right tool for anything intermittent, an artifact that's only audible in specific moments or specific sections, warbling bands that swing in level rather than holding steady.
The advantage over a static notch is real: you get correction exactly when it's needed and none of the cost when it isn't. The tradeoff is setup complexity, you need to actually identify when the problem occurs before you can target it properly.
Low-pass: the broad tool, and usually not the first move
A filter that removes everything above a set frequency, no exceptions. This is the blunt instrument of the group, and it's tempting precisely because it's simple, one filter, problem gone. But if there's any genuine, gradually-declining real content in the region you're filtering, and there often is, a low-pass takes that real detail out along with the artifact.
Reserve this for cases where a narrower approach genuinely isn't working, or where you've confirmed through an A/B that there's nothing worth preserving in that range. Even then, treat it as a deliberate trade-off you're choosing after comparing, not a default reach.
Spectral repair: for isolated, specific moments
A tool that lets you edit a specific time-and-frequency region directly, rather than applying a change across the whole track. This is the right choice for a brief, isolated event, a single moment of shimmer, one bad transient, something that happens once or twice rather than persisting throughout. It's more surgical than anything else on this list, and more time-intensive per instance, which is exactly why it's the right tool for rare problems and the wrong one for anything happening constantly.
The actual decision
Ask two questions before choosing. Is the problem constant or does it come and go? If constant, notch. If intermittent, dynamic EQ. Is it happening throughout the track or in one or two specific moments? If throughout, one of the two above. If isolated, spectral repair. Low-pass only enters the conversation once you've confirmed, through an actual comparison, that there's nothing worth saving in the range you're about to remove entirely.
The tool doesn't matter until the diagnosis is right. But the diagnosis doesn't finish the job either, it just tells you which of these four conversations you're actually having.
— Josh Founder, JG BeatsLab
