Speech Separation

Sometimes you want to isolate the speaker's voice from background music or ambient noise in your video. StarZero Agents' speech separation feature uses AI to split a video clip's audio into two separate tracks: one with just the vocals (speech), and one with everything else (music, ambient sounds, background noise).

What It Does

When you run speech separation on a video clip, StarZero Agents:

1. Analyzes the clip's audio track
2. Separates it into vocals (speaking voices) and background (music, ambient noise, sound effects)
3. Creates two new audio clips on the timeline
4. Mutes the original video clip's audio (sets volume to 0%)

The two new audio clips are linked to your video clip, so they move and delete together — keeping everything synchronized automatically.

When to Use It

Speech separation is perfect for:

• Isolating a speaker's voice when there's distracting background music or noise
• Removing background music while keeping the dialogue
• Adjusting the music separately from the speaker's voice (like ducking music under dialogue)
• Rearranging speech segments while maintaining continuous background music — great for ads or content where you want to reorder what's said without the music cutting awkwardly
• Creating clean audio for content where you want full control over the voice and background mix

How to Separate Audio

Just ask in the chat, pointing to the specific video clip:

"Separate the audio on clip 2"

"Split the vocals and background on the interview clip"

"Isolate the speaker's voice from the music in clip 5"

The AI will process the audio (this typically takes 10-60 seconds depending on the clip's length) and create two linked audio clips.

Controlling the Volume

By default, both the vocals and background tracks are set to 100% volume (their original levels). You can specify different volumes when you separate:

"Separate clip 3's audio with vocals at 100% and background at 30%"

"Split audio on clip 1, keep the voice at 80% and mute the background"

Volume ranges from 0% (silent) to 200% (double volume), just like regular audio controls.

Linked Clips

The vocals and background audio clips are linked to the original video clip. This means:

• Moving the video automatically moves both audio clips with it
• Deleting the video also deletes the audio clips
• Trimming isn't linked — you can still adjust in/out points independently if needed

Linked clips make it easy to rearrange your timeline without worrying about audio getting out of sync.

Adjusting After Separation

Once the audio is separated, you can treat the vocals and background clips like any other audio:

• Change their volume individually to get the perfect mix
• Mute one track to hear only the voice or only the background
• Adjust in and out points to trim the audio if needed
• Re-separate if you want to try different volume settings (though it will create new clips)

Requirements

To use speech separation, your video clip must:

• Have its volume above 0% (you can't separate audio from a muted clip)
• Contain actual audio in the source media (video-only files won't work)

The AI will let you know if there's an issue with the clip you're trying to process.

Tips for Best Results

• Background music at 20-30% usually works well when you want subtle music under dialogue
• Vocals at 100% keeps the speaker's voice clear and prominent
• Mute the background entirely if you just want clean dialogue with no music or noise
• Preview the separated tracks before moving forward — sometimes the separation isn't perfect with very complex audio mixes

What's Happening Behind the Scenes

StarZero Agents uses AI trained to distinguish human speech from other sounds. There's no special setup — just ask, and the separated tracks arrive in a moment.

What's Not Supported

Speech separation is powerful, but has some limitations:

• Can't separate multiple speakers into individual tracks — all voices go into the vocals track together
• Not perfect with very complex mixes — if there's heavy reverb, overlapping sounds, or extreme effects, the separation may have artifacts
• Can't re-combine separated audio — once separated, the tracks are independent. You can delete them and turn the video clip's volume back up if you want to start over.

For most video content — interviews, vlogs, presentations, podcasts-to-video — speech separation gives you clean, usable results that make your audio mix sound more professional.