loudness.ai
Recording clinic / Hiss

Microphone hiss: find the cause and fix it

Find where the “shhh” enters your recording, fix the next take, or rescue a file you cannot record again. Start with the symptom that matches your setup.

Only in my headphones

Check playback before altering the recording.

Follow this path

In every recording

Separate room noise, mic placement and input settings.

Follow this path

I already have the file

Try cleanup, then check whether soft words survive.

Follow this path

Start with this one-minute mic test

Save a 15-second test: a few seconds of room sound, one ordinary sentence, then a quieter sentence. Say “The first and last words should both be clear.” Use your normal recording app or record a sample with Mic Test, and note the selected microphone.

The saved file hisses elsewhere

Play the downloaded file on another device at a normal voice volume. If the same hiss travels with it, continue with the recording checks below.

The saved file is clean elsewhere

Keep the file unchanged. Compare another headphone output or another pair of headphones. Processing the file will not repair noise added by playback.

Locate the hiss with three controlled takes

Keep the same sentence and normal speaking effort. Save each take separately; do not overwrite the baseline. Leave automatic gain and suppression unchanged while comparing positions. If you disable an effect to investigate it, make that a separate test.

TakeChange only thisRead the result
A · BaselineUse your usual mic position and settings.This is the file every later take is compared with.
B · RoomStop a nearby room fan briefly, if practical, without moving the mic. If it changes the noise, use the fan-noise recording guide to test a better microphone position. Leave computer cooling running.Less hiss at the same voice level points to an acoustic contribution. If unchanged, continue; this does not prove the mic is faulty.
C · DistanceRestore take A, then move closer. Reduce input gain only if needed to avoid clipping.Match voice playback level with A. Less background relative to speech makes placement a useful fix. More bass or breath pops means you need a little more space or an off-axis position.

Hiss, hum or crackling? Listen to the pattern

What you hearWhat to check first
Steady, airy hissRoom noise, input level and the recording chain. Repeat the test with nearby fans off when practical.
Low hum or buzzUse the microphone hum troubleshooting guide to check connected devices, cables and power. This has a different sound from a broad hiss.
Intermittent clicks or cracklesA loose connection or a problem that appears when a cable moves. Try a known-good cable with the input muted.
Harsh sound on loud wordsInput peaks. Lower the recording gain and repeat the loudest phrase.

You can compare the sound with Shure’s recorded examples of common microphone noises. A sound pattern is a starting point, not proof that one part is faulty. Compare the saved file and a second setup before replacing equipment.

Check distance and input gain together

Move the microphone a little closer to your voice and repeat the sentence. Keep it slightly off to one side if breath sounds hit the capsule. Leave enough space to speak naturally.

Adjust the input gain so your loudest words have room below the maximum. Then compare both takes at similar voice volume. Changing distance and gain together can improve the balance between voice and room sound.

For a microphone connected to an interface or mixer, check that you selected a microphone input. Shure’s explanation of microphone and line levels explains why a line input expects a stronger signal. Turning up a mismatched input can make hiss more obvious. USB microphones do not use that same external mic-versus-line connection.

Why turning the volume down is not a hiss fix

Imagine a voice measured at −24 dBFS and a pause at −48 dBFS using the same RMS meter. Lowering the whole file by 12 dB gives −36 and −60 dBFS. Both changed equally: the 24 dB gap stayed the same. These are worked numbers, not measurements from your microphone.

Original

Voice −24 · Pause −48 → gap 24 dB

Volume down 12 dB

Voice −36 · Pause −60 → gap 24 dB

Raise the playback volume to hear the voice as loudly as before and the hiss comes back too. This is why the distance test and the listening examples use comparable voice levels. Do not choose a take just because its meter or its waveform looks smaller.

Listen for three different failure points

Listen to three different voices with added hiss, then compare each recording after Noise Removal. These are real audiobook narrators. The hiss was added for this demonstration; it was not captured by their microphones. The spoken passages are kept intact, including their pauses.

The voice levels are matched for comparison. Keep your headphone volume comfortable and unchanged. Listen during a pause, then listen to the next word: does the voice stay clear as the background changes?

1. Garth Comira — Broad, steady hiss. Listen to the spaces between phrases and the ends of words.

Garth Comira — Before: voice with added hiss
Download WAV
Garth Comira — After: Noise Removal
Download WAV
Inspect the before-and-after spectrogram
Garth Comira: hiss spectrograms, before above and after below, with identical time, frequency and colour scales.
Look between the spoken phrases: the broad coloured background becomes much darker after processing. The brighter shapes belong mainly to the voice. Listen to the word endings before deciding the result is better.
Open full-size chart

2. Heather Barnett — Brighter, stronger hiss. Compare the softer parts of the voice. Check whether consonants remain easy to hear.

Heather Barnett — Before: voice with added hiss
Download WAV
Heather Barnett — After: Noise Removal
Download WAV
Inspect the before-and-after spectrogram
Heather Barnett: hiss spectrograms, before above and after below, with identical time, frequency and colour scales.
The stronger hiss fills much of the upper chart, including the pauses. The lower chart shows less background energy, but the voice still needs a listening check: especially softer consonants.
Open full-size chart

3. Anders Lankford — Softer, air-like hiss. Listen for background noise under the speech, as well as in pauses. Check that word endings are not cut short.

Anders Lankford — Before: voice with added hiss
Download WAV
Anders Lankford — After: Noise Removal
Download WAV
Inspect the before-and-after spectrogram
Anders Lankford: hiss spectrograms, before above and after below, with identical time, frequency and colour scales.
Compare the space around each phrase rather than judging the brightest colour alone. A darker background is useful when the words remain clear at the same voice level.
Open full-size chart

Each “after” file is the processed version of its matching “before” recording. Turning the whole recording down would lower the voice and hiss together. Noise reduction changes that balance, but it can also change the voice: choose by speech clarity, not by the quietest pause alone.

Reduce hiss in a recording you already have

For adjustable cleanup, follow Audacity’s Noise Reduction instructions: select a noise-only pause, open Noise Reduction and get a noise profile. Select the passage to treat, reopen the effect and preview a modest reduction. Use Residue to hear what would be removed; recognizable words there mean you should reduce the strength or sensitivity.

After cleanupDecision
Words stay clear; a little hiss remainsKeep this candidate and judge it in the finished recording at normal volume. Chasing total silence may make it worse.
Soft consonants disappear or the voice warblesReturn to the original. Use a gentler setting in an adjustable editor, or keep the original if it communicates better.
The important words were already buriedRecord those lines again if possible. Keep the microphone position and room tone consistent with the surrounding passage.

Measure a pause without mistaking it for silence

In the detailed Mic Test analysis, select a pause with room sound and no speech or music. Compare that section across takes made with the same setup. An edited silent gap cannot tell you how noisy the microphone was.

A lower background reading is useful only alongside a clear voice at a comparable level. Use the LUFS Meter to compare recording levels and peaks. There is no single noise reading that identifies the faulty device or guarantees a good recording.

How to read the before-and-after graphs

A spectrogram shows how sound changes over time. Read left to right through the recording; higher frequencies are higher up. Bright yellow means a stronger signal at that frequency. Purple and black show weaker sound.

Yellow does not automatically mean noise: speech can be yellow too. Start with a pause, then follow the sound into the next word. Both panels use the same time, frequency and colour scales and come from the exact audio files below.

What to compareWhat to do with it
Colour spread across a pauseCompare the background in the same pause before and after.
The patterns during soft wordsPlay that phrase and check consonants and word endings. A darker chart alone is not a quality score.
Noise comes back in the next takeRepeat the distance and input checks, changing one thing at a time.

The hiss charts show 0–8,000 Hz; the hum charts zoom in to 0–800 Hz. Open a chart at full size for detail. A black area means little energy on this colour scale, not proof of perfect silence.

Recording credits

Recording credits: The Ashiel Mystery, read by Garth Comira; Sense and Sensibility, read by Heather Barnett; The Age of Chivalry, read by Anders Lankford. Audio from the LibriSpeech collection, adapted by Loudness.ai with added hiss, noise reduction and level matching. The source recordings are licensed under CC BY 4.0.

Sign in to Loudness.ai