Electrical hum is often a narrow, steady tone with harmonics—not the same problem as room noise, wind, keyboard clicks, or competing voices. Keep the original and judge every change by spoken clarity, not silence.
Quick answer
Preserve the original, identify whether the sound is a stable narrow hum, and test a representative difficult sentence. Use a focused de-hum or notch-style repair before increasing broad noise reduction. Compare the same sentence, pause, and cut at matched volume.
- Stable narrow tone: test focused de-hum first.
- Broad fan-like layer: use cautious steady-noise cleanup.
- Clicks, crackles, and interruptions: repair locally.
- Background speech and severe echo: seek a better source or rerecord key words.
Diagnose the sound before choosing a fix
A steady low tone with related harmonics can be electrical hum. A broad low wash can be rumble, wind, handling, or vibration. Adobe distinguishes DeHum around common 50-Hz and 60-Hz ranges from Reduce Rumble, which targets lower-frequency noise.
- Listen for a fixed tone and harmonics.
- Check whether it shifts or crackles.
- Preview what is removed when the editor supports it.
Use a preview-first workflow
Select a normal sentence, a quiet sentence, a pause, and the noisiest hum. Confirm the tone is stable, try focused repair first, then compare Original and Cleaned at matched volume. If speech becomes hollow, metallic, watery, thin, or less intelligible, use less processing or return to the original.
Audacity notes that lower reduction lowers the risk of losing desired sound, while loud or variable noise may be difficult to remove satisfactorily.
Where NoiseVanish fits—and where it does not
NoiseVanish is intended to reduce steady background noise around speech, including hum, fan wash, air conditioning, traffic beds, and some wind. A short preview can help assess whether a stable layer falls while dialogue remains natural.
It is not a promise to identify the electrical source, remove every harmonic perfectly, repair clicks, or reconstruct words another sound covered. Use focused de-hum or notch-style repair when a narrow hum is the main problem.
Prevent hum in the next recording
Use short safe tests with a different cable, microphone, light, charger, camera input, or power source one at a time. Keep audio cables away from power supplies where practical, use maintained equipment, and monitor with headphones. Do not alter electrical wiring unless qualified.
Identify the hum pattern before you process it
Start with a five- to ten-second pause where nobody speaks, then listen again during a sustained vowel. A mains-like hum usually stays at roughly the same pitch while the speaker pauses, whereas traffic, wind, handling noise, and a nearby appliance often drift in level or texture. Headphones make this easier to hear than laptop speakers. If your editor provides a spectrum display, look for one narrow fundamental line with evenly spaced harmonics above it. That pattern is a useful clue, not a diagnosis of the electrical cause.
Do not assume every low sound is a 50 Hz or 60 Hz problem. Camera preamps, grounding loops, LED lights, chargers, dimmers, HVAC motors, and vibration can add nearby tones or a changing buzz. Write down where the recording was made, which devices were powered, and whether the sound is present in every clip. That record helps you choose a repair and can prevent repeating the same problem on the next shoot. It also stops a broad noise-reduction setting from being used on the wrong kind of noise.
- Compare a silent pause and a spoken sentence before selecting a repair.
- Look for a stable narrow line and regularly spaced harmonics when a spectrum view is available.
- Treat a moving buzz, intermittent crackle, or handling vibration as a different troubleshooting problem.
Work from the fundamental upward
A focused repair normally begins with the lowest audible hum frequency, then checks whether its harmonics still make the recording sound boxy, buzzy, or thin. Removing only one line may leave higher tones that are more noticeable once the fundamental is lower. Removing a wide band around all of them, however, can take useful warmth and consonant detail out of speech. Make a small change, bypass it, and compare the same sentence at matched playback volume. Louder output can sound cleaner even when it is less natural, so level matching matters.
If a notch or de-hum control lets you adjust width, begin narrowly. Widen it only when the tone remains obvious and the voice still sounds full. Avoid adding several aggressive processors at once: a camera's built-in voice isolation, an editor's de-hum, a broadband denoiser, and a final enhancer can each remove a little speech texture. Once stacked, it is difficult to know which step created the metallic result. Save versions or duplicate the track so every decision is reversible.
Use a short, repeatable listening test
Choose three checkpoints before changing settings: the noisiest spoken sentence, a normal sentence, and a pause or edit point. Apply the proposed repair to a short copy of the clip, then switch between Original and Cleaned without changing the monitor level. Listen first for whether words are easier to understand, then for collateral damage: weak S and F sounds, flattened breaths, a hollow chest tone, or pumping that rises and falls between phrases. A cleaner-looking waveform is not the goal if the presenter becomes tiring to listen to.
It is useful to take a brief break before the final comparison. Your ears adapt quickly to a constant hum and to an overprocessed voice. After returning, play the audio through headphones and ordinary speakers, and watch a cut or title transition in the actual video. Some repairs sound acceptable in a solo preview but reveal a sudden room-tone jump at an edit. The best setting is usually the least noticeable intervention that improves dialogue enough for the intended viewer.
- Check the same timestamps before and after rather than comparing different sentences.
- Match playback level before deciding that one version is better.
- Keep the untreated source and a clearly named cleaned version until export is approved.
Know when a general noise cleanup is useful
A speech-focused cleanup pass can be useful after a light de-hum when the recording also contains a broad, steady bed such as fan wash, air conditioning, camera hiss, or distant traffic. In that situation, address the narrow tone first and use the lowest cleanup amount that improves the remaining bed. NoiseVanish is most useful for this steady background layer around otherwise understandable dialogue. Use a preview and judge the difficult sentence, not only an empty pause.
General cleanup is not a substitute for a targeted fix when a single electrical line dominates the track. Nor can it restore syllables that were covered by a loud buzz, click, handling impact, or another speaker. If the hum changes each time a device turns on, or if it is accompanied by crackling and dropouts, the safest answer may be local editing, a different source recording, or rerecording a short voiceover. Make that limitation clear if someone expects a one-button restoration.
Prevent the same problem on the next shoot
Prevention is easier than aggressive repair. Before an important recording, capture a short headphone-monitored test with the camera, microphone, lights, charger, and computer arranged as they will be for the real take. Change one item at a time: unplug a charger, move an audio cable away from a power adapter, try a different light, switch to battery power where appropriate, or use another microphone input. This controlled approach is safer and more informative than changing several things at once.
Do not attempt electrical-wiring repairs unless you are qualified to do so. For a recurring production setup, label the cable and power arrangement that produced a clean test, keep connectors maintained, and record a few seconds of room tone at the start of each session. That room tone will not magically repair speech, but it gives an editor a reference for the space and makes abrupt edits easier to hear. A clean capture also leaves more room for gentle post-production when unexpected noise remains.
Final export check
Before exporting, audition the beginning, middle, and end of the finished video instead of only the repaired clip. Confirm that music, captions, cuts, and loudness changes do not make the hum return or make the cleaned dialogue feel detached from the scene. Keep the original media and your editable project version. If a client, teammate, or future editor needs to revisit the choice, they should be able to compare the repair with the source rather than inherit an irreversible file.
Frequently asked questions
Is hum always 50 Hz or 60 Hz? Those are common references, but recordings can include harmonics, shifted tones, mechanical noise, or other interference.
Can broad AI noise reduction remove hum? It may reduce a steady layer, but a narrow hum may need focused treatment and strong processing can thin speech.
Why does hum remain? The recording may also include harmonics, a changing source, rumble, or fan noise.
Related NoiseVanish guides
Sources and update notes
This article was reviewed on September 28, 2026. Product interfaces and third-party features change; check the linked official documentation before relying on exact tool behavior.