Audio leveling

Bad leveling is one of the fastest ways to make a video feel unfinished. A voice that is too quiet gets lost. Music that is too loud fights the message. A sudden SFX hit can make viewers turn the volume down.

Good leveling is almost invisible. The viewer does not think about the mix; they just understand the video.

• Voice is too quiet compared to music.
• Music jumps between sections.
• SFX are much louder than the rest of the video.
• One voiceover clip is louder than another.
• The final export sounds quieter than expected after upload.
• Cuts feel abrupt because there are no fades.

Audio leveling is about balance inside the video: voice against music, SFX against narration, clip against clip.

Loudness normalization is about the overall loudness of the final export. Platforms may adjust this after upload, but leveling still matters because a normalized bad mix is still a bad mix.

There is no universal perfect level, but a practical short-form mix usually keeps speech clearly above background music, avoids clipping, and leaves enough headroom before export.

Use exact dB/LUFS numbers as guidelines, not rules. The most important test is whether the voice is easy to understand on phone speakers.

Fuse gives music, voiceover, SFX, and original video audio their own lanes. You can adjust levels separately, add fades, mute or solo lanes, and use auto-duck when music needs to move under speech.