Fairlight · AI dialogue · Studio
Voice Isolation in Fairlight
Voice Isolation is Blackmagic's DaVinci Neural Engine tool for keeping spoken voice intelligible while suppressing competing background and room sound. It is a DaVinci Resolve Studio feature and is most useful when the goal is a clean human voice rather than a detailed remix of the surrounding environment.
Keep the human voice, push competing sound away
1. What it does
AI isolation for recorded human voice
Blackmagic introduced Voice Isolation as a Studio Neural Engine process for removing loud unwanted sounds from existing voice recordings. Blackmagic's own examples include traffic, machinery, air conditioning, room sound, echo and reverb, and current 2026 production material still shows Voice Isolation being used to salvage dialogue with street noise and reverberant rooms.
The design target is dialogue. It is not a general-purpose source separator for music, Foley or ambience. If you need to keep and rebalance the environment instead of mostly suppressing it, Dialogue Separator is the more appropriate branch of the Fairlight AI toolset.
Identify whether the problem is noise, room/reverb, level, sibilance or something else.
Start conservatively and listen for consonants, breaths and tonal instability.
Judge the processed voice against the picture, ambience and music — not solo only.
Use EQ, dynamics or Dialogue Leveler for the problems Voice Isolation does not solve.
2. Practical use
Do not turn isolation into a one-click maximum setting
Location dialogue
Interviews and production dialogue where traffic, HVAC, machinery, rain or a reflective room competes with speech are natural candidates.
Start conservatively
Blackmagic production material notes that even low settings can remove enough background while avoiding artifacts. More processing is not automatically better.
Listen to consonants
Check word attacks, breaths, quiet syllables and the tonal body of the voice. Separation artifacts often become obvious there before they show up on meters.
Judge in context
A voice can sound impressively dry in solo and still feel unnatural against picture, music and room tone. A/B the full scene regularly.
3. What it does not replace
Isolation is one stage of dialogue repair, not the entire chain
Once unwanted background has been reduced, the voice may still need level work, tonal correction, sibilance control or conventional dynamics. Do those jobs with the tools designed for them instead of asking Voice Isolation to solve everything.
Dialogue Leveler
For inconsistent dialogue level and perspective — not for separating voice from a noisy environment.
De-Esser
For excessive vocal sibilance. Do not use AI isolation as a substitute for a targeted sibilance processor.
EQ & Dynamics
For tone, compression, gating and final control after the recording is clean enough to shape musically.
Noise Reduction
For conventional noise-profile repair when the unwanted sound is a stable identifiable noise bed.
4. Voice Isolation vs other tools
Choose the processor by what you need to preserve
Next: preserve more of the environment
Use Dialogue Separator when the room itself matters
Dialogue Separator exposes voice, background and ambience as separately adjustable elements, making it a better fit when the goal is rebalance rather than simply a cleaner isolated voice.
Voice Isolation is checked against Blackmagic's original Studio 18.1 release documentation, the current DaVinci Resolve 21 manual availability page, and Blackmagic production stories from 2025–2026 showing continued real-world use for dialogue cleanup. Older parameter-specific documentation is used only where Blackmagic explicitly documented behavior. The current single-Amount control layout shown here is anchored to the supplied live Resolve 21 screenshot.