FAQ / Consistent loudness on a long narration
How do I keep loudness consistent across a long YouTube narration?
Aim for an integrated loudness around -14 LUFS for YouTube (-16 LUFS is a safe default that ports well to most platforms). Get it consistent by recording with a fixed mic distance, balancing per-paragraph gain on outliers, and running the loudness-normalized merge before you export.
Pick a target and stick with it
YouTube normalizes audio toward roughly -14 LUFS for playback. If you master quieter than that, YouTube will not boost you, so your video will play quieter than the next one in autoplay. If you master louder than that, YouTube will turn you down and your dynamic range will suffer. -14 LUFS is the sweet spot for YouTube specifically. -16 LUFS is a safer default if the same file is going to multiple platforms (Vimeo, Spotify video, podcast feeds).
Three places loudness drifts
- Mic distance drift inside a session. You lean forward when excited, lean back when reading something dry. Same paragraph, different level.
- Day-to-day drift across sessions. You sit a touch closer Tuesday than Monday. Whole paragraphs come out hotter.
- Performance drift. A whispered passage and a punchy one inside the same paragraph can be 10 dB apart on the meter.
The first two are gear and habits. The third is intentional and you actually want to keep it.
Use per-paragraph gain to fix outliers
If you scan the merge-screen waveforms and one paragraph is visibly quieter than the rest, nudge its gain slider up a couple of dB on the editor screen. The merge picks the new value up. Easier than re-recording, faster than a DAW pass.
Per-paragraph gain in VoiceOverAndOver is a clean digital boost - it does not change the room sound or noise floor, just the level. For a paragraph that is quiet because of room noise, boost works. For a paragraph that is loud because you clipped on a plosive, re-record. Boost will not save a clipped paragraph.
Let the merge do the heavy lifting
The merge step in VoiceOverAndOver normalizes the merged file toward your chosen LUFS target. That covers the slow drift between paragraphs that no manual adjustment realistically can. Your job is to keep the input close to right; the merge tightens it the rest of the way.
Listen on phone speakers
The final consistency test is not your studio monitors. It is a phone speaker. If a paragraph fades out on the phone, it will fade out for the half of your audience that watches on phones. That paragraph wants more gain (or a re-record, if the issue is room noise).