There is no single correct level, because voices, music and playback devices vary. These starting points will get you close quickly; then trust your ears on several devices.
Starting levels
- Under continuous narration: music roughly 15 to 20 dB below the voice. Speech should be effortless to follow on a phone speaker.
- Under interviews: often lower, or no music at all during the most important answers.
- Between lines and in montages: bring music up to near the voice level, then back down when speech returns.
Use ducking
Ducking lowers music automatically whenever someone speaks. Most editors have it built in: Premiere Pro’s Essential Sound panel, DaVinci Resolve’s Fairlight ducking, and auto-ducking options in several mobile editors. Set a gentle reduction with smooth fades so the music does not pump noticeably.
Make space with EQ
The ear relies heavily on the midrange, roughly 1 to 4 kHz, to understand speech. A gentle dip of a few dB on the music in that range helps the voice sit on top without turning the music down further.
Choose music that wants to be underneath
The easiest fix is choosing the right track. Look for steady dynamics, no lead melody in the vocal range and no lyrics. When generating, add "background bed for narration, no lead melody, even dynamics" to your prompt.
Final loudness
After mixing, set overall loudness to the platform you publish on. Many streaming and video platforms normalise to around -14 LUFS; Apple recommends about -16 LUFS for podcasts. Our mastering guide covers targets in more detail.
Check on three devices
Listen on headphones, laptop speakers and a phone. Phone speakers are the harshest test: if the voice is clear there, it will be clear everywhere.
Written by the Sonvela editorial team with AI drafting assistance, then checked against the sources linked in the text. Spotted an error? Tell us and we will correct it.