Home Vocal Recording and Mixing

How to fix harsh vocals in REAPER (with stock plugins)

By StudioRescue Editorial

Published July 17, 2026 · Updated July 17, 2026

“Harsh” is that hard, edgy, fatiguing quality that makes a vocal feel like it is poking at your ears, especially on louder words and on cheaper monitors or earbuds. It is one of the most common complaints about home-recorded vocals, and like mud it is a tone problem, not a talent problem. You do not need a better mic or a treated room to make a real improvement. You need to find the upper-mid range that is spiking and bring it down.

Before you start, it helps to know exactly which problem you have, because “harsh” gets used for three different things that live in three different places. This article is about edge and bite in the upper mids. Here is how to hear it, confirm it, and tame it in REAPER using only stock plugins.

What harshness actually is

Every voice needs some energy in the upper mids. That range, very roughly 1.5 kHz to 5 kHz, is where presence and intelligibility live, and it is the part of the vocal that helps it cut through a mix. Harshness is what happens when there is too much of it, or when one narrow spot in it is spiking far above the rest.

It helps to separate three problems that people all call “harsh,” because the fix is different for each:

• Edge and bite (roughly 2–5 kHz). A hard, aggressive, in-your-face quality on vowels and loud notes. This is what most people mean by harsh, and it is what this guide fixes.
• Sibilance (roughly 5–9 kHz). Piercing “s,” “sh,” and “t” sounds. That is a de-esser job, not a plain EQ cut, because it comes and goes with the words.
• Nasal honk (roughly 1–2 kHz). A pinched, boxy, talking-through-your-nose tone. That is a separate resonance, covered in our nasal vocals guide.

A few things commonly push a home vocal into harsh territory:

• Bright or cheap microphones. Many budget condensers have a boosted presence range, which flatters a dull room but stacks up fast on an already-bright voice.
• Room reflections. Hard, untreated surfaces reflect high-mid energy straight back into the mic and exaggerate the edge.
• Too much presence EQ. Reaching for an upper-mid or high-shelf boost to add “clarity” is the fastest way to tip a vocal into brittle and harsh.

Test it before you touch anything

Guessing at EQ is how good vocals get over-processed. Confirm where the harshness lives first.

The fix, step by step in REAPER

REAPER ships with ReaEQ, an unlimited-band parametric EQ, and ReaXcomp, a multiband compressor. Between them you can handle both constant harshness and harshness that only spikes on loud notes. Add ReaEQ to the vocal track’s FX chain first.

1. Cut the harsh band with ReaEQ

Take a ReaEQ band, set it to a normal band (peaking) shape, and place it at the frequency you found during the sweep, usually somewhere in 2–5 kHz. Start with a cut of about 2–4 dB and a medium width. A good rule of thumb is boost wide, cut narrow, so keep the width tighter than you would for a warmth boost, but not so narrow that it sounds like a hole. Listen in the full mix, not just soloed. You want the edge to soften while the vocal still sounds clear and present, not dull or distant.

2. If it is only harsh on loud notes, use ReaXcomp

Sometimes the vocal sounds fine on quiet phrases and only turns harsh when the singer pushes. A fixed EQ cut then makes the quiet parts dull just to fix the loud ones. This is where a dynamic approach helps. Add ReaXcomp and narrow one band to cover roughly your problem range, for example about 2–5 kHz, then set the threshold so that band only compresses when the harshness spikes. The result is an EQ move that backs off automatically when it is not needed and only clamps the edge on the loud words.

3. Re-check in the mix

Harshness is partly relative to everything else. A vocal that sounds edgy alone can settle once the rest of the mix supports it, and a vocal that seemed fine soloed can turn brittle over bright cymbals and synths. Always make the final call with the full arrangement playing, and on more than one set of speakers or headphones if you can. Toggle the processing on and off to be sure you are improving things, not just changing them.

What EQ will not fix

Taming harshness moves tone around, but it cannot undo everything.

• It will not repair a recording that clipped or distorted on the way in. Digital clipping adds its own hard, gritty edge, and no EQ cut removes it. That has to be re-recorded at a safer level.
• It will not fix sibilance on its own. If the problem is the “s” and “sh” sounds coming and going, a static cut up high will dull the whole vocal. Use a de-esser instead.
• It cannot fully remove strong room reflections baked into the take. EQ can soften the tone they add, but it does not remove the reflections themselves.

If it still sounds off

Quick troubleshooting order when the vocal is not improving:

1. You cut too much. If the voice went dull or distant, raise the cut back toward 0 dB and try a smaller reduction, or widen the band slightly.
2. It is sibilance, not edge. If the pain is on the “s” and “t” sounds around 5–9 kHz, a fixed cut will not track it. Reach for a de-esser.
3. It is nasal, not harsh. If the tone is more honky and pinched than piercing, the resonance is lower, around 1–2 kHz. See the nasal vocals guide.
4. It is distortion. If soloing reveals crackle or grit on the loudest words, the take likely clipped. No EQ setting fixes that; re-record with more headroom.