VIDEOMONTAZHERDAVINCI RESOLVE · VIDEO · AI

FAIRLIGHT / TARGETED REPAIR / 07.02

De-Esser / De-Hummer

Two different faults. Two different decisions.

A sharp consonant, a steady electrical tone and broadband hiss can coexist without becoming one problem. Describe the event first, then test an appropriate tool on a short scene.

  • 01Separate symptoms from causes
  • 02Avoid adding both effects by default
  • 03Preserve words and tone
FAIRLIGHT / INTERACTIVE STUDIOTEACHING MODEL
REPAIR / TARGETB / MODEL
SOURCE / SYNTHETICB / MODELTARGET 58 / MODEL 45%

Move the target region and compare the source and illustrative result.

An event within a word.↗

Locate a consonant that jumps out of the sentence. Do not treat all high-frequency content as a fault.

Check: the word remains intelligible.

Change the controls to explore the diagram. No audio playback, native DSP or connection to Resolve.

Sources and scope · 7 October 2026

Purposes were checked against Blackmagic’s official page; controls against the primary Resolve 21 manual text and illustrations on the independent iMaster mirror. Text and images are distinguished: a missing button is not drawn in. The teaching records are an authored review method, not the plug-in algorithm. Listening tests and verification of a particular installed build were not performed.

Read the source notes ↓
On this page

01 / UNDERSTAND & APPLY

Start with an event, not a plug-in name

Choose a troublesome passage and a nearby clean phrase. Listen to the same source without repair inserts: existing processing may be making the problem more conspicuous. Note the position, channel and the actual distraction—a consonant, persistent tone, broadband hiss or isolated click.

Record an observation rather than assigning a cause from a single clue. A narrow spectral peak may belong to a voice or instrument. A persistent tone during a pause is useful evidence, not proof of a faulty cable. Preserve the source and use a test copy.

ObservationWhat to inspectFirst direction
Sharp s, ts or sh soundsIs the event tied to a consonant?De-Esser
A steady tone and related harmonicsDoes it persist in pauses and other takes?De-Hummer
Broad persistent noiseCan it be distinguished from useful sound?Noise Reduction
An edit click, impact or overloadWould editing or another take be more appropriate?Separate investigation

Reference: Blackmagic Design — Resolve 21 Reference Manual · Fairlight FX, pp. 4101–4102 (iMaster mirror)

02 / UNDERSTAND & APPLY

De-Esser: keep the consonant, reduce excessive sharpness

This effect targets excessive sibilance in dialogue and vocals. Start with several words at different intensities, not only the loudest s. The consonant carries information: making it disappear is not the goal. For a single outlier, consider a local clip adjustment first.

The full plug-in provides Frequency, Range, Amount and Reaction Time. The documented Range options are Narrow Band, Wide Band and All High Frequency. These are neither De-Hummer controls nor three quality levels. The separate De-Esser guide explains them individually.

Reference: Blackmagic Design — Resolve 21 Reference Manual · Fairlight FX, pp. 4101–4102 (iMaster mirror)

03 / UNDERSTAND & APPLY

De-Hummer: tonal hum and its harmonics

De-Hummer addresses hum, including electrical interference. The manual describes fixed 50/60 Hz choices and a variable fundamental. Choosing 50 Hz solely from the shooting location is not a substitute for listening: the disturbance may have another structure or may not be mains hum at all.

For an assumed 50 Hz fundamental, early multiples are 100, 150 and 200 Hz; at 60 Hz they are 120, 180 and 240 Hz. This explains harmonic spacing; it is not a measurement of your file or a map of every internal filter. Slope balances treatment between the fundamental and harmonics; it is not a conventional high-pass slope in dB/oct.

Use Listen to Hum only to check whether useful low-frequency content is being removed with the disturbance. Return to normal output and inspect the voice, music and ambience. This overview helps select a tool; the full De-Hummer controls belong to its own page.

Blackmagic Design · iMaster · external imageClick to enlarge ↗
De-Hummer — reference illustration

De-Hummer

The illustrated Variable mode and 50 Hz value do not identify the frequency of your interference. The graph is not a measurement of your recording.

Image source ↗

De-Hummer

513 × 571 · Blackmagic Design · iMaster · external image

The illustrated Variable mode and 50 Hz value do not identify the frequency of your interference. The graph is not a measurement of your recording.

Open original ↗

Reference: Blackmagic Design — Resolve 21 Reference Manual · Fairlight FX, pp. 4101–4102 (iMaster mirror) · Blackmagic manual illustration — De-Hummer (independent hosting)

04 / UNDERSTAND & APPLY

When the recording has more than one problem

Sibilance ends with the word while hum may persist through the pause. The clues are not mutually exclusive: a recording can contain hiss, hum and sharp consonants at once. Do not force one insert to solve all of them. Define a separate task for each event and compare one change at a time.

Quieter noise with new holes in the speech is not a finished repair. Check another take, channel or gentler setting. A repair effect cannot guarantee recovery of information missing from an overloaded or truncated recording.

Reference: Blackmagic Design — Resolve 21 Reference Manual · Fairlight FX, pp. 4101–4102 (iMaster mirror)

05 / UNDERSTAND & APPLY

Clip, track or shared bus

A clip is a useful scope for a one-off defect. A track insert suits genuinely consistent material. Clips sharing A1 do not necessarily share microphone sensitivity, levels or sibilance.

Before processing a shared bus, identify its sources. De-essing a speech-and-music sum may affect cymbals, ambience or another voice. Start with the individual source when uncertain. Recheck after moving settings to another scope because the processor input has changed.

ScopeUseful forRisk
ClipOne take or phraseTone changes at edit boundaries
TrackA consistent recording chainDifferent voices receive the same setting
BusDeliberate processing of a sumUnintended sources are altered

Reference: Blackmagic Design — Resolve 21 Reference Manual · Fairlight FX, pp. 4101–4102 (iMaster mirror)

06 / UNDERSTAND & APPLY

Order: compare rather than follow a recipe

After a high-frequency EQ boost, recheck sibilance. If compression makes quiet hum more noticeable, compare repair before and after it. These are reasons for an A/B test, not a rule that repair must always come first. Reordering also changes the input to subsequent processing.

Keep the source passage, scope, listening level and other processing unchanged. Do not move an insert while also changing frequency and Amount. Retain the original setting, document the change and check a neighbouring scene.

Reference: Blackmagic Design — Resolve 21 Reference Manual · Fairlight FX, pp. 4101–4102 (iMaster mirror)

07 / UNDERSTAND & APPLY

Listen to what is removed—then return to the speech

The manual describes Listen to Ess Only and Listen to Hum only as monitoring the removed component. This is diagnostic monitoring, not an unprocessed reference or the final mix. Recognisable syllables, bass notes or ambience in that signal warrant checking the cost of the repair.

Useful material in the removed component does not automatically quantify damage; normal output must also be evaluated. Turn diagnostic monitoring off before checking the mix. Inspect a quiet ending, loud phrase, pause and edit, then listen to the whole scene with music.

Reference: Blackmagic Design — Resolve 21 Reference Manual · Fairlight FX, pp. 4101–4102 (iMaster mirror)

08 / UNDERSTAND & APPLY

Observation record: what is actually known

This is not a recording analyser. Record whether harshness follows consonants, a steady tone persists in pauses, and broadband noise is present. Not checked is available for each clue and is not treated as no. Several clues produce several investigation directions, not a mandatory effects chain.

Enter the target and event position. Changing the target or clues clears the previous listening assessment while keeping the note. The JSON carries your observations; it neither diagnoses a noise source nor changes the project.

REPAIR / OBSERVATIONS

Describe the clue. Keep the uncertainty.

No file upload, automatic noise detection, cause diagnosis or processing order. Unknown is not no. This record remains only in the open page until you export it.

Correct the fields: target (1–160 characters); position (1–160 characters)

09 / UNDERSTAND & APPLY

Troubleshooting and handoff

Save the test project and a short record of the insert location, task and unresolved issues. A preset name alone is not enough for handoff: sources, order and insert availability matter. Reopen the project and check the same passage again.

SymptomNext check
Speech lisps after de-essingReturn to normal output; compare Amount and Range on whole words.
Vocal body disappears after de-hummingInspect the fundamental and removed component; compare another passage.
Only noise seems audibleCheck Listen Only and the intended instance, then routing.
One voice improves, another worsensSeparate sources and processing scope.
Preset loaded but sound differsCheck input, order, bypass and subsequent inserts.

Reference: Blackmagic Design — Resolve 21 Reference Manual · Fairlight FX, pp. 4101–4102 (iMaster mirror)

PRACTICE / NO AUTOPILOT

A small task. A checkable result.

Work in a test copy, listen before and after, and write down what you actually observed. The checklist contains no pre-approved results.

  1. Find two distinct events

    In a timeline copy, mark a sharp consonant and suspected hum. If the second fault is absent, record that.

  2. Keep an untreated comparison

    Identify the source and scope. Listen without the chosen insert at the same monitoring level.

  3. Test one tool

    Change one parameter at a time. Do not add another effect merely because it sits nearby in the menu.

  4. Restore normal output

    Inspect the removed component, then turn Listen Only off. Check articulation, vocal body and ambience.

  5. Record a scoped conclusion

    Save the conclusion for this scene: benefits, damage and unverified details. Repeat after reopening the project.

Download the practice record · TXT ↓

FAQ / BEFORE THE NEXT MOVE

Questions that matter in practice

Should I always use both effects?

No. One addresses sibilance, the other tonal hum. Add processing for an observed task and evaluate it separately.

Reference: Blackmagic Design — Fairlight / Clean Up and Repair Sound

Will De-Hummer remove any background noise?

No. It addresses hum and related tonal components. Broadband noise, reverberation, wind and edit clicks need separate investigation.

Reference: Blackmagic Design — Resolve 21 Reference Manual · Fairlight FX, pp. 4101–4102 (iMaster mirror)

Why not choose 50 or 60 Hz from the recording country alone?

The switch position does not establish the interference frequency. Inspect pauses and useful sound; not every low tone is mains hum.

Reference: Blackmagic Design — Resolve 21 Reference Manual · Fairlight FX, pp. 4101–4102 (iMaster mirror)

Can a high-frequency cut replace de-essing?

EQ continuously changes its selected region, whereas De-Esser targets sibilance. Compare consonants and neighbouring sounds; darkening the entire phrase may be undesirable.

Reference: Blackmagic Design — Resolve 21 Reference Manual · Fairlight FX, pp. 4101–4102 (iMaster mirror)

Does the page record diagnose the noise source?

No. It holds entered clues and suggests investigation directions. No audio is uploaded or spectrum measured.

When can I accept the repair?

After checking the specific source and normal output, preserving words and desired tone, and recording unresolved issues. A positive self-report on the site does not replace listening.

REFERENCES / 2026-10-07

What this guide is based on

Purposes were checked against Blackmagic’s official page; controls against the primary Resolve 21 manual text and illustrations on the independent iMaster mirror. Text and images are distinguished: a missing button is not drawn in. The teaching records are an authored review method, not the plug-in algorithm. Listening tests and verification of a particular installed build were not performed.

Blackmagic Design — Fairlight / Clean Up and Repair Sound ↗

Primary source for the purposes of De-Esser, De-Hummer and Noise Reduction and clip/channel placement. Read 7 October 2026; not a test of an installed build.

Blackmagic Design — Resolve 21 Reference Manual · Fairlight FX, pp. 4101–4102 (iMaster mirror) ↗

Primary Blackmagic text on an independent mirror; De-Esser, De-Hummer and Common Controls read. Range names, Reaction and removed-signal monitoring checked. The Listen to Ess Only text/image discrepancy is explicitly noted.

Blackmagic manual illustration — De-Hummer (independent hosting) ↗

Inspected Frequency, 50 Hz / 60 Hz / Variable, Amount, Slope and Listen to Hum only. It depicts settings, not a measurement of user audio.