DAVINCI RESOLVE AUDIO TROUBLESHOOTING
DaVinci Resolve Voice Isolation Not Working?
If Voice Isolation is missing, greyed out, doing almost nothing, making speech sound unnatural, or behaving differently after export, the cause is much easier to find once you know which of those problems you actually have. Start by matching your symptom below, then work through the checks in order.
Drop your recording here
or click to browse
We'll clean a short preview first. No payment required.
MP4, MOV, MP3, WAV
Hear the difference
Play the same moment with and without AI cleanup. The audio changes, but the video stays the same.
What’s wrong with Voice Isolation?
“Not working” covers several different failure modes. Find the row that matches what you are seeing — the second column is where to look first, not a guaranteed diagnosis.
Is Voice Isolation available in your version of DaVinci Resolve?
Check this first, because it can solve the problem immediately. DaVinci Resolve comes in two editions — the free DaVinci Resolve and the paid DaVinci Resolve Studio — and they do not include the same features.
Blackmagic Design describes Voice Isolation as one of the DaVinci Neural Engine AI tools, and states that the Neural Engine powers features found exclusively in DaVinci Resolve Studio . In practice that means:
- Voice Isolation requires DaVinci Resolve Studio.
- In the free edition the control is expected to be absent — that is not a broken installation.
- Reinstalling the free version will not add a Studio-only feature.
If you already run Studio and the control is still missing, the edition is not the issue — move on to the version and context checks below. Edition features can change between releases, so if anything here does not match what you see, treat Blackmagic’s current documentation as the source of truth.
Check your Resolve version
Where controls live and how audio behaves can shift between Resolve releases. A tutorial or forum screenshot from a different version can make a control look “missing” when it has simply moved, and an older installation may not match the current interface at all.
- Check which Resolve version you actually have installed.
- Compare it against Blackmagic's current documentation rather than an old thread.
- Update only if it suits your workflow.
One caution: if you work on production projects, updating mid-project can affect plugins and project behaviour, so do not update blindly just to chase a feature. Confirm the feature and its requirements first.
Check whether you’re processing the clip or the track
In Fairlight, effects can be applied to an individual clip or to a whole track . That flexibility is useful, but it is also a common source of confusion: it is easy to add Voice Isolation at one level while listening for a change at another.
This matters because:
- A clip-level effect only touches that one clip, not the others on the same track.
- A track-level effect touches everything routed through that track.
- Clip and track processing can coexist, so you may be applying more than you think.
- If you expected one clip to change on its own but processed the track, every clip is affected instead.
Confirm where the effect actually sits, and where you are listening, before deciding it is broken. Choose the level that matches what you intend to process, and check you are not unknowingly stacking both.
Check mono, stereo, and multichannel audio
A surprising share of “Voice Isolation is broken” reports turn out to be an audio-layout problem rather than an effect problem. Source media varies a lot in how its channels are arranged:
- Mono audio
- Stereo audio
- Several embedded channels
- Camera audio with multiple channel assignments
- Interview media with lav and room mics on separate channels
- 5.1 or other multichannel audio
Different channel layouts can make an effect behave in ways you do not expect — for example, if the speech you want is not on the channel you assume. There is no single universal fix, so work through it as a diagnostic:
- 1
Inspect the clip's actual channel configuration..
- 2
Identify which channel contains the speech you want to keep..
- 3
Make sure the timeline and track configuration match the intended source..
- 4
Test Voice Isolation on a simple mono or stereo speech clip..
- 5
Compare that against the problematic media..
This tells you whether Voice Isolation itself is failing or the source configuration is responsible.
Test Voice Isolation on a simple clip
A controlled test is one of the most useful things you can do, because it separates a source or workflow problem from an effect problem in a couple of minutes:
- 1
Take a short 10–30 second speech recording..
- 2
Use straightforward mono or stereo audio..
- 3
Place it in a simple, clean timeline..
- 4
Apply Voice Isolation on its own..
- 5
Temporarily avoid any other audio effects..
- 6
Listen to the before and after..
If the test works
The problem is likely specific to your real project: the source media, its channel layout, the timeline configuration, the effect chain, or the clip/track setup.
If the test does not work
Look at the fundamentals instead: your Resolve edition and version, whether the effect is bypassed, and your software or project configuration.
Disable other audio processing temporarily
Audio in Resolve can pass through several stages before you hear it. If the result is unexpected, temporarily remove or bypass everything except Voice Isolation, then test it alone:
- EQ
- Compressor
- Noise gate
- Other restoration tools
- Third-party plugins
The goal is not to claim these effects are incompatible with Voice Isolation — they are not. The goal is to find out which stage is causing the unexpected result. Once you know, you can rebuild the chain deliberately.
Processing order matters too. Several aggressive processors can produce very different results depending on their order and settings — a hard gate before isolation, a strong EQ cutting speech frequencies, or heavy compression amplifying leftover background can all change the outcome. There is no single correct chain for every recording, so the reliable approach is simple:
Test Voice Isolation by itself first. Then add the other processing back one stage at a time, listening after each one.
Voice Isolation is on, but I barely hear a difference
If the effect is clearly enabled but the audio hardly changes, one of these is usually why:
The source is already relatively clean
If there is little background noise to begin with, do not expect a dramatic change.
The unwanted sound is not the kind it handles well
Speech isolation struggles with another person talking, loud music, severe reverberation, clipping, sudden impact sounds and complex overlapping audio.
The voice itself is weak
If the speaker is far from the microphone and the noise is nearly as loud as the voice, clean separation becomes much harder.
The effect is applied somewhere other than expected
Re-check clip vs track placement and whether the effect is bypassed.
Other processing changes the result
EQ, compression, gates, restoration effects or plugins elsewhere in the chain can alter what you end up hearing.
Voice Isolation makes the voice sound robotic
This is one of the most common complaints, and it comes from a real trade-off rather than a bug. Isolation becomes difficult when the wanted voice and the unwanted sound overlap strongly, so separating them can leave artifacts such as:
- Robotic speech
- A metallic tone
- An underwater or watery sound
- Missing consonants
- Unstable, shifting ambience
- Unnatural breath sounds
These artifacts are more likely when:
- the background noise is loud;
- there is strong room echo;
- several people talk at once;
- music overlaps the speech;
- the microphone clipped;
- the recording is already heavily compressed;
- isolation is pushed aggressively.
This is not unique to Resolve. It is a general limitation of source separation and speech enhancement: the harder the voice and noise are to tell apart, the more the voice is affected. If the result sounds processed, easing the strength usually helps more than pushing it further.
Voice Isolation cannot fix microphone clipping
A lot of wasted effort comes from asking speech isolation to repair something it was never meant to fix. It helps to keep two different problems apart:
- Background noise: unwanted sound captured alongside usable speech, such as a fan or steady hiss.
- Clipping: the original signal exceeded the recording range and became distorted, so information was lost at capture.
Once speech has clipped badly, detail is already gone. Voice Isolation or any noise reduction can change the recording, but it cannot reconstruct clean speech that was never captured. If clipping is your real problem, no amount of isolation will make the audio pristine — recognising that early saves a lot of re-tuning.
It sounds right in Resolve but wrong after export
Treat this as its own diagnostic case — there is no single cause. When the timeline sounds right but the exported file does not, the effect you rely on may not be in the path that is actually being rendered. Areas worth inspecting:
- Whether the intended effects are active in the final audio path
- Timeline vs render configuration
- The output audio bus
- Channel mapping
- Export format
- Whether the correct track or bus is being rendered
A controlled test narrows it down quickly:
- 1
Export a short section rather than the whole timeline..
- 2
Compare it directly with timeline playback..
- 3
Disable unrelated audio processing..
- 4
Confirm the correct output channels and bus..
- 5
Repeat until the export matches what you hear..
Voice Isolation disappeared after an update
If the control was there before and now seems gone, several explanations are more likely than a specific bug:
- The interface or the effect's location changed between versions.
- A different edition is now installed than you expect.
- The current project or audio context is set up differently.
- The selected clip or track is configured in a way that hides the option.
- A tutorial you are following refers to a different Resolve version.
Re-check your edition, your version, Blackmagic’s current documentation, and the audio context you are working in before concluding anything is broken.
When Voice Isolation may not be the right tool
Voice Isolation is optimised around speech. That makes it the wrong choice when your goal is to keep the rest of the sound intact — for example when you want to preserve:
- Music
- Room ambience
- Environmental sound
- Multiple simultaneous voices
- Natural acoustic character
Some noise types also respond better to more targeted treatment than to broad speech isolation:
- Electrical hum: a de-hum tool aimed at the specific frequency
- Low rumble: filtering or EQ on the low end
- Pauses with room noise: possibly a gate
- Steady broadband noise: general noise reduction
For a single dominant noise type, a focused guide can help — such as how to remove hum from audio or how to remove hiss from audio.
Clean the audio separately and bring it back into Resolve
If you decide to clean the file outside Resolve, the flow is straightforward:
- 1
Export or gather the audio or video source you want to clean..
- 2
Process it with Noisero..
- 3
Preview the cleaned result before you commit..
- 4
Bring the cleaned file back into your Resolve project..
Noisero accepts both video and audio. If you upload a video such as an MP4 or MOV, it cleans the audio and leaves the picture exactly as it was, so you can drop the returned file straight back onto your timeline; if you upload audio, you get the cleaned audio back to reconnect to your clip.
Which background noises Noisero can help with
Cleanup works best on steady, consistent sound. Noisero is aimed at exactly that kind of background noise:
- Fan noise
- HVAC and air conditioner noise
- Hiss
- Hum
- Steady room noise
- Distant traffic and general background
For a specific problem, these focused guides go deeper: how to remove fan noise from audio, remove air conditioner noise, remove traffic noise, or remove background noise from a video online.
Resolve Voice Isolation vs separate noise cleanup
These are not competitors so much as different workflows. Which one fits depends on what you are doing, not on one being universally better.
Voice Isolation in Resolve
- You want to stay inside the Resolve timeline.
- Speech enhancement is part of the full audio mix.
- A detailed Fairlight workflow is appropriate.
A separate cleanup tool
- You just want to clean a source file.
- You want to test another result quickly.
- You would rather not debug a complex timeline or effect chain.
If you are troubleshooting other editors too, the same logic applies in Premiere Pro, CapCut, Audacity and Adobe Podcast Enhance Speech.
