Make a Quiet Recording Loud — Properly (Normalize vs Boost)
A recording at −29.4 LUFS. The volume booster will safely add 1.7 dB; the normalizer asks for 15.40. The difference is one row of the loudness meter — and it is explained here, free and with nothing uploaded.
Length: 3:25Published:
What you'll learn
Measure first — LUFS, not the waveform
Peak to loudness is the number that decides everything
Why the booster’s safe maximum is only +1.7 dB here
The normalizer asks for a destination, not a number
Landed −14.90 LUFS with the limiter engaged
When the booster is still the right tool
In this video
A recording at −29.4 LUFS. The volume booster will safely add 1.7 dB; the normalizer asks for 15.40. The difference is one row of the loudness meter — and it is explained here, free and with nothing uploaded.
If a recording is too quiet, the obvious fix is a volume booster — and on a recording with one loud moment in it, the booster will refuse to add more than a decibel or two, because it is protecting that one moment rather than your voice. This video measures a real 28-second meeting recording on the loudness meter (−29.4 LUFS, true peak −2.65 dBTP, peak to loudness 26.7 dB), shows what the volume booster does with it, and then shows the normalizer working out the gain from a target instead: +15.40 dB to reach −14 LUFS, limiter engaged, landed at −14.90. The point is not that one tool is better. A booster reasons about peaks; a normalizer reasons about loudness; and knowing which question you are asking is the whole skill. Everything runs inside your browser, so the recording is never uploaded, and the network log is shown at the end.
Image Tool Hub — 200+ free tools for images, video, audio, PDFs and Office documents, in English, Spanish, French, German, Portuguese and Indonesian. Everything runs on your device.
Full transcript
Every word spoken in the video, so you can read it, search it or skim it.
This recording is far too quiet, and the obvious fix is a volume booster. On this file the booster will safely give you one point seven decibels. It needs fifteen. Here is why — and what to use instead.
Three tools, one recording, about two minutes. By the end you will know which of the two you want, and how to tell in five seconds.
−29.4 LUFS
LUFS is loudness the way an ear hears it, averaged over the whole file — not the height of the tallest peak. Streaming platforms want you around minus fourteen, so this is a long way down.
Peak to loudness 26.7 dB
Maximum without clipping: +1.7 dB
Asked +12.0 dB · applied +11.20 dB
Apply +15.40 dB to reach −14 LUFS
Landed −14.90 LUFS · limiter engaged
That is the whole difference, and it is not that one tool is better. The booster reasons about peaks, so one loud second decides what it will do. The normalizer reasons about loudness, so it ignores the thud and holds it instead.
Use the booster when you already know the answer — lift this one clip by three. Use the normalizer whenever the file has to match something else: a podcast, a voiceover, or a folder of interviews all recorded at different levels.
And all three of those ran right here. This is the browser’s own record of everything it fetched while it happened — not one request carried the recording anywhere.
All three are free and none of them needs an account. Measure first, then normalize — and I will see you in the next video.
Subscribe for new tutorials
New walkthroughs every week. Free tools, nothing uploaded.