How to stop the stock clips' own sound getting into your mix
Tell Vizard Agent to mute the source clips unless their sound is doing something. Downloaded and stock footage carries whatever was recorded on the day — traffic, a room, a stranger talking — and it arrives under your narration by default. It checks what each clip actually holds, mutes the ones adding nothing, and measures whatever survives against the voice.
What is the short version?
Nobody chooses the ambience in a stock clip. It comes with the file, several clips each bring their own, and the result under a voiceover is a wash of unrelated rooms that sounds like a fault rather than a texture.
- Go to Vizard Agent with the footage.
- Say the source audio should be muted by default.
- Ask which clips have sound worth keeping.
What do you need before you start?
The clips and a view on what the soundtrack is. Vizard Agent needs to know whether the video is narration over music, in which case most source audio is noise, or whether the footage's own sound is part of the point.
- The clips. Stock, downloaded or supplied.
- What carries the audio. Narration, music, or the footage.
- Any clip whose sound matters. Name it.
- Whether the video is spoken over throughout. It usually is.
- What the mix should sound like. Clean, or textured.
What do you type into Vizard Agent?
Give Vizard Agent the default and then ask it for the exceptions. "Mute the stock footage sounds" is already a complete instruction, and asking which clips are worth un-muting is what turns a blanket rule into an actual decision about each one.
Prompt
Variants worth knowing:
- "With the source clips muted." The default, stated.
- "Tell me if any is worth keeping." The exceptions.
- "So the voice leads." What the mix is for.
What does Vizard Agent actually do?
Here is the shape of it on a real narrated video assembled from mixed sources. The muting itself is just a rendering decision; all of the useful work happens before that, in Vizard Agent finding out what is actually sitting on each audio track.
- Measures each clip and checks whether it has audio at all.
- Checks the footage for spoken content it would be a shame to lose.
- Listens to what the native audio actually is.
- Mutes the clips whose sound adds nothing.
- Keeps any clip whose sound carries the moment.
- Measures the narration's loudness.
- Measures the music separately.
- Sets the duck level by comparing an active narration window against the music.
- Renders with the source clips muted.
- Checks the waveform for gaps or an abrupt ending.
Step two is the one worth insisting on. Muting everything is easy and occasionally throws away a line somebody actually said on camera, so the check for spoken content comes before the blanket rule rather than after it.
Step three matters because "has audio" and "has useful audio" are different findings. A clip can carry a full track of nothing but air conditioning, and that is exactly the sound that makes a finished mix feel muddy without anybody being able to say why.
Step eight is what makes the mix defensible rather than a guess. Comparing a window where the narration is speaking against the music underneath it produces a number, and a number is something you can adjust deliberately later.
Step ten is the last thing worth asking for. Muting several source tracks changes what the mix is carrying, and a bed that was masked by all that ambience can turn out to have gaps in it that nobody could hear before, so Vizard Agent looks at the waveform of the result rather than assuming the removal was clean.
What does the result look like?
A clean mix: your narration, your music, and silence where four unrelated rooms used to sit under everything. Vizard Agent tells you which clips it muted and which it kept, so a moment you did want to hear is one instruction away rather than a rebuild of the whole thing.
When does this not work well?
Muting by default is the wrong rule when the footage itself is the point, and Vizard Agent will say so. Documentary material, reaction clips and anything where the recorded sound is evidence all want their native audio kept, and there the default should be reversed.
- Documentary footage. The sound is the record.
- Reaction or interview clips. Someone is speaking.
- Action footage. The impact sells the shot.
- Music-led edits. There may be nothing to mute.
- Clips with only partial audio. Muting some is inconsistent.
How do you fix a result that came back wrong?
Name the clip and say what you wanted to hear from it. Because Vizard Agent made the decision per clip rather than applying one setting globally, un-muting a single one is a small change that leaves the rest of the mix exactly as it was.
- "Keep the sound on the third clip." Un-muted and levelled.
- "It sounds too dry now." Add ambience deliberately instead.
- "The music is too loud." Re-measured against the voice.
- "Someone speaks in that shot." Kept, and ducked around.
How does Vizard Agent compare to doing it yourself?
By hand the clips go on the timeline with their audio attached, because that is how they arrive. The mix gets balanced by ear afterwards, which trims the worst of it without anybody ever deciding which rooms are in the video.
| By hand | Vizard Agent | |
|---|---|---|
| The default | Source audio comes along | Muted unless it earns it |
| What is checked | The mix | What each clip carries |
| Spoken content | Lost with the mute | Found before muting |
| The duck level | Set by ear | Measured against the narration |
| The record | None | Which clips were muted |
Common questions
Why is stock audio a problem? Because several clips each bring an unrelated room, and they stack.
Will it mute everything? Only what adds nothing. Vizard Agent checks each clip first.
What if someone speaks in a clip? Vizard Agent looks for that before muting, and keeps it.
Can I keep some ambience? Yes, though Vizard Agent finds one deliberate bed cleaner than four unrelated rooms.
Does it check every clip? Yes, including whether the clip has any audio at all.
Will the video sound too dry? It can. Ask Vizard Agent to add one deliberate bed instead.
How is the music level set? By measuring it against a window where the narration is speaking.
Can it mute only part of a clip? Yes. Vizard Agent can keep the few seconds where the sound matters.
Does this apply to my own footage? Vizard Agent runs the same check, though the answer is more often to keep it.
What about sound effects? Those are added deliberately, so Vizard Agent leaves them alone.
Will it tell me what it muted? Yes. Vizard Agent reports it clip by clip, so you can reverse any of them.
Does muting change the picture? No. Vizard Agent keeps this to the audio only.
Can I set this as a habit? Yes. Say it with each brief and Vizard Agent applies it.
Why did my mix sound muddy? Because it contained several rooms nobody chose to put in it.