Audio Equalizer that fixes, not muddies

An equalizer is the most misunderstood control in audio software. It looks like a tone control, so people use it like one: everything up, everything down, or a dramatic curve copied from a video. Then the recording sounds worse and nobody knows why. The useful version is simpler and less exciting.

It changes balance, not quality

Every sound is a stack of frequencies. A voice has weight around 100–200 Hz, body near 500 Hz, intelligibility between 2 and 4 kHz, and air above 8 kHz. An equalizer turns the volume of those slices up or down. It cannot restore a frequency that was never recorded, remove a room completely, or turn a phone memo into a studio take. What it can do is stop one slice from hiding another.

Why cuts beat boosts

Boosting raises the signal and the noise already living in that band. Cutting removes both, and it frees headroom for the rest. This is why an experienced engineer reaches for a small cut at 250 Hz before touching the treble: the mud is masking the clarity, so removing the mud reveals the clarity that was already there.

There is a second reason. The ear is very sensitive to narrow, large boosts — they sound like a filter. Wide, shallow cuts are almost invisible. If a move is obvious, it is usually too big.

A practical pass on speech

1. Cut everything below 70–80 Hz. Nothing useful lives there in a voice recording, but plenty of rumble does.
2. Cut 2–4 dB around 250–350 Hz if the speaker sounds boxy or "inside a cardboard box".
3. Lift 2–4 kHz by 2 dB only if words still get lost. More than that turns the voice harsh.
4. Cut 8–10 kHz if the "s" whistles. A narrow cut is better than a broad one here.
5. Compare at the same loudness. A brighter version always seems better until you match the level.

When EQ is the wrong tool

If the problem is hiss, traffic or a fan, filtering will not fix it: the noise shares the same frequencies as the voice. Use the background noise remover. If the whole file is too quiet, use the normalizer. If the room rings, the reverb tool cannot remove it — reverb is added, never subtracted.

Exporting without losing the work

Choose WAV when the file goes back into an editor or a video project. Choose MP3 when the destination is a phone, a messenger or a podcast host. Avoid converting MP3 → WAV → MP3 repeatedly; each lossy generation removes a little more of the top end, and no equalizer can put it back.

Finally, keep the original. A good EQ pass is one you can undo.