audio-toolsHow to Normalize Audio Loudness (Peak vs RMS Explained)
Make quiet tracks louder and even out volume swings the right way. A clear guide to peak vs RMS normalization, and when each one is the tool you actually want.
6 mins read
Audio Normalizer, Level Volume Online with Web Audio
Drop a recording that's too quiet, too hot or all over the place and watch it settle to a clean, consistent level. Real peak and RMS normalisation with a before/after waveform, a live preview and lossless WAV or MP3 export: all in your browser, never uploaded.
100% browser processingFree · no sign-up
Loading tool…
What is the audio normalizer?
The Audio Normalizer lifts a too-quiet recording to a clean target level using peak or RMS normalisation to an exact dBFS ceiling like −1 or −3. Drop a voice memo, interview or song, compare the before/after waveform, then download WAV or MP3. It decodes and re-levels every sample on the native Web Audio engine, so nothing leaves your device.
The GigAI Audio Normalizer runs entirely on the native Web Audio engine your browser already ships with, so it can decode a clip, measure it, and re-level every sample without ever sending your audio to a server. Voice memos recorded across a room, interview tracks where one speaker is buried, podcast segments cut from different sessions, and songs exported too quiet all share the same problem, their level is wrong for where you need to play them, and reaching for a full DAW just to raise the volume is overkill. Drop your file, choose how to normalise (peak, which lifts the track so its single loudest moment just reaches your ceiling without clipping, or RMS, which raises the average energy so the whole thing feels louder), set a target in dBFS, and the tool computes the exact gain and applies it to the entire signal in one pass. When it finishes you get the original waveform above the normalised one, a preview player to confirm it sounds right, and the measured peak and RMS before and after, then download a lossless 16-bit WAV or an MP3 encoded locally in your browser. Because everything is client-side there is no upload wait and no privacy tradeoff: personal recordings, unreleased mixes and confidential interviews never leave your machine. Be honest about what normalisation is: it applies one gain factor to make a quiet track louder or a hot track safe. It is not multiband compression, it can't un-clip audio that was already recorded too hot, and peak/RMS targets are not the same as the broadcast LUFS loudness spec that streaming platforms measure. Close the tab and every sample is gone.
- Difficulty:
- Medium
- Typical time:
- ~30s
- Processing:
- 100% browser processing
Last updated
How to use the audio normalizer
- 1
Drop your audio
Drag an MP3, WAV, M4A, OGG or FLAC onto the workspace, or click to browse. It's decoded locally and you'll see its waveform, duration and channel layout immediately.
- 2
Choose mode and target
Pick Peak (maximise headroom without clipping) or RMS (lift perceived loudness), then set a target ceiling in dBFS. −1 is a safe default for a loud, clean result.
- 3
Normalize
Press Normalize. The whole track is re-levelled on your device in one fast pass. No engine download, no upload, no waiting on a server.
- 4
Compare, preview & download
Read the before/after peak and RMS, see both waveforms, play the preview to confirm it sounds right, then download a lossless WAV or a browser-encoded MP3, or start over with a different target.
What Audio Normalizer includes
Peak or RMS, your call
Peak normalisation lifts the track so its loudest sample just kisses your ceiling: perfect for rescuing a quiet recording without clipping. RMS raises average energy so the whole clip feels louder. Pick the one that fits the job.
Set an exact dBFS target
Aim at a precise ceiling: a common −1 dBFS for a safe max, −3 dBFS for streaming headroom, or wherever you need. The tool computes the gain to hit it and applies it to every sample in a single clean pass.
Before / after waveform
See the original envelope stacked above the normalised one, so the level change is obvious at a glance. Not a vague 'processed' badge. The shape is preserved. Only the amplitude scales.
Preview before you download
The normalised result plays right in the page on a waveform player so you can confirm it sounds balanced and nothing distorts before you commit, the safety net a command line never gives you.
Lossless WAV or local MP3
Export a clean 16-bit PCM WAV to keep every sample for further editing, or an MP3 encoded entirely in your browser for a smaller share-ready file. No exotic codec loads for either.
100% private, nothing uploaded
Decoding, measuring and re-levelling all happen on your device with the native Web Audio engine. Voice memos, interviews and unreleased mixes are never uploaded, stored or logged.
Why use our audio normalizer
Rescue a recording that's too quiet
A voice memo taped across a room or an export that came out low is unusable at normal volume. Peak-normalise it to −1 dBFS and it lifts to full, clean level in one pass. No hunting through a DAW.
Even out a batch to a shared level
Segments cut from different sessions rarely match. Normalising each to the same target gets them into the same ballpark so listeners aren't reaching for the volume knob between clips.
No upload wait, no privacy tradeoff
Server normalisers make you upload the file first and trust them with it. Here the audio never leaves your browser, so there's no upload bar and nothing to trust: safe for private or confidential recordings.
Hear it before you keep it
The before/after waveform and the preview player let you confirm the level is right and nothing clipped at your chosen target before you download. No re-doing a bad export.
Built for the way you work
From quick one-off fixes to daily workflows, see how people put this tool to use.
- Podcaster / Interviewer
Level a quiet interview track
One guest recorded far quieter than the host. Normalise the buried track up to a matching ceiling so both voices sit at a comfortable, consistent level before you mix.
- Musician / Producer
Bring a low export up to level
A bounce came out several dB under where it should be. Peak-normalise it to −1 dBFS to get a full-scale, clean master reference without re-exporting from the session.
- Content creator
Match voiceover clips before editing
Voiceover takes recorded on different days drift in level. Normalise each to the same RMS target so they cut together smoothly instead of jumping in loudness.
- Privacy-conscious user
Boost a sensitive recording locally
A confidential memo or a personal voice note that's too quiet shouldn't be uploaded to a stranger's server to be levelled. Do it entirely in your browser so nothing ever leaves the device.
Supported formats
Accepts MP3, WAV, M4A, AAC and OGG, and produces WAV and MP3, all processed locally in your browser.
Input formats
- MP3
- WAV
- M4A
- AAC
- OGG
- FLAC
Output formats
- WAV
- MP3
Frequently asked questions
Recommended tools
Audio Trimmer
Trim audio in your browser with a real waveform: drop an MP3, WAV, M4A or OGG, drag a selection to keep exactly the part you want, preview it, then download a lossless WAV or a compact MP3. Sample-accurate and 100% private. Nothing is uploaded.
Audio Studio
MP3 Converter
Convert any audio to MP3 in your browser, WAV, M4A, AAC, OGG, FLAC and more to a clean, universally-playable MP3 with a bitrate you choose. Preview the waveform, encode locally with lamejs, then download. Nothing is uploaded.
Audio Studio
Audio Recorder
Record from your microphone right in the browser: live level meter and timer, trim the ends on a waveform, then export a clean WAV or MP3. Nothing is uploaded. Your audio never leaves your device.
Audio Studio
Audio Merger
Merge multiple audio files into one in your browser. Drop MP3, WAV, M4A or OGG clips, drag to set the order, add an optional crossfade to smooth the joins, then download a single lossless WAV or compact MP3. 100% private. Nothing is uploaded.
Audio Studio
Ringtone Maker
Make a ringtone from any song in your browser. Drop an MP3, WAV, M4A or OGG, drag a selection over the waveform to grab the hook (up to ~40 seconds), add a smooth fade-in and fade-out, preview it, then download an MP3 or M4A/AAC. 100% private. Nothing is uploaded.
Audio Studio
AAC Converter
Convert any audio to AAC in your browser, WAV, MP3, M4A, OGG, FLAC and more to a compact, high-efficiency.aac file that plays natively on iPhone, Android, YouTube and modern video apps. Preview the waveform, encode locally, then download. Nothing is uploaded.
Audio Studio
Guides & articles
- audio-tools
Common problems, solved
Hit a snag? Here are quick fixes for the issues people run into most.
The output barely got louder.
Your source probably already peaks near the ceiling, so there's little headroom to gain in Peak mode. The loudest sample is already close to the target. If it still feels quiet, switch to RMS mode to lift the average energy, and check your target isn't set lower than the source's existing peak.
The normalised audio sounds distorted or pumped.
That usually means RMS mode drove the peaks into the hard limiter on already-loud material. Lower the target a few dB, or switch to Peak mode for a completely transparent gain with no limiting. If the distortion was in the original file, normalisation can't remove it: the source itself is clipped.
Nothing happened. It sounds the same.
If the track is effectively silent, there's nothing to normalise and the tool leaves it unchanged rather than amplifying noise to full scale. Otherwise the source may already be at your target level. Try a different mode or a slightly higher ceiling and compare the before/after peak and RMS readouts.
Get the most out of it
Start with Peak at −1 dBFS, it's the safe, transparent way to bring almost any quiet clip up to a full, clean level.
Reach for RMS when a clip is loud enough at its peaks but still feels quiet overall (lots of dynamic range). It lifts the average, not just the loudest moment.
Always play the preview before downloading, it's the fastest way to catch a target that pushed RMS material into audible limiting.
Normalising several clips to the same target gets them close, but for a truly matched set, listen and nudge each one, perceived loudness isn't purely a number.
Keep a lossless WAV as your master if you'll edit further. Only encode to MP3 as the final step to avoid compounding lossy generations.
What's new
Recent updates and improvements to the audio normalizer.
Initial release, native Web Audio peak and RMS normalisation with an exact dBFS target, a stacked before/after waveform, a preview player, measured before/after peak & RMS, and one-click lossless WAV download.
Added in-browser MP3 export with a bitrate selector (encoded locally, no engine download), plus clearer honest guidance distinguishing peak/RMS normalisation from LUFS loudness targets.
Refined the RMS hard-limiter behaviour to reduce audible pumping on hot material, improved the silent-track guard, and added drag-and-drop plus reset-and-retry to the workspace.
Keep exploring
Problems we solve
Definitions
From the blog
Explore categories
Compare formats
Common tasks
Your privacy is built in
Your audio is decoded, measured and re-levelled entirely in your browser using the native Web Audio engine. It is never uploaded to any server. Nothing is stored, logged or transmitted, so you can safely normalise personal, confidential or unreleased recordings. WAV and MP3 export both happen on-device with no exotic codec download. There's no upload step and no cloud copy, so close the tab and every sample is gone.
- Runs in your browser
- No uploads
- Nothing stored
Ready to try the audio normalizer?
Free, private and instant. Audio Normalizer runs right in your browser.