Audio Normalizer
How to Normalize Audio Online
- 1
Upload your audio file
Drop your file or click to browse. MP3, WAV, FLAC, OGG, AAC, M4A, WMA, and WebM are all supported. Nothing is uploaded — the file stays in your browser.
- 2
Review the analysis
The tool automatically analyzes your file — showing duration, sample rate, channels, and current peak and RMS levels in dBFS. This tells you how far off your target your audio currently is.
- 3
Choose a LUFS target
Select a preset for your platform — Spotify and YouTube use -14 LUFS, Apple Podcasts recommends -16 LUFS, broadcast TV follows EBU R128 at -23 LUFS. Or enter a custom value.
- 4
Normalize and download
Click "Normalize Audio." The two-pass process first measures your audio precisely, then applies the correction. Download the normalized file when processing is complete.
Understanding Audio Loudness
LUFS — Perceived Loudness
Loudness Units Full Scale. The standard metric for perceived loudness, weighted to match human hearing. Streaming platforms enforce LUFS targets to make all tracks sound equally loud to listeners, regardless of original recording level.
Peak — Maximum Amplitude
The highest sample value in the entire waveform. Peak tells you the loudest instant moment in the file. Keeping peak below 0 dBFS (true peak below -1 dBFS) prevents clipping, but peak alone does not reflect perceived loudness.
RMS — Average Signal Level
Root Mean Square — an average of all sample values over time. RMS gives a rough sense of overall loudness and correlates better with perceived volume than peak alone. LUFS is a more accurate, frequency-weighted evolution of RMS.
Why Normalize?
Inconsistent loudness is jarring — a podcast episode recorded quietly sounds half as loud as the next. Platforms like Spotify apply automatic gain adjustments that can make loud masters sound worse. Normalizing to target levels avoids both problems.
LUFS Targets by Platform
Different platforms enforce different loudness standards. Normalizing to the target before upload prevents the platform from applying its own (often lower quality) gain correction.
| Platform | Target LUFS | Notes |
|---|---|---|
| Spotify | -14 LUFS | Loudness normalization enabled by default |
| YouTube | -14 LUFS | Applied after upload; louder content gets turned down |
| Apple Music | -16 LUFS | Sound Check feature targets this level |
| Apple Podcasts | -16 LUFS | Recommended by Apple for podcast submissions |
| Broadcast (EBU R128) | -23 LUFS | European standard for TV and radio; legal requirement in many countries |
Frequently Asked Questions
- What is LUFS?
- LUFS stands for Loudness Units Full Scale. It is the industry standard measurement for perceived loudness over time, used by streaming platforms, broadcasters, and podcast hosts to ensure consistent volume across content. Unlike peak level, LUFS reflects how loud audio actually sounds to human ears.
- Why does normalization use two passes?
- The first pass measures the actual loudness of your audio file — how loud it really is, not just the peak. The second pass uses those measurements to apply a mathematically precise gain correction. Two-pass processing gives far more accurate results than guessing the gain in a single pass.
- Will normalization reduce audio quality?
- The impact is minimal. Normalization applies a linear gain change, which is lossless in concept. The only quality loss comes from re-encoding the audio, which we do at high quality settings. For most content — speech, podcasts, casual recordings — the difference is imperceptible.
- What's the difference between Peak and LUFS?
- Peak level is the highest amplitude sample in the waveform — the loudest instant moment. LUFS is a weighted average of loudness across the entire file, reflecting perceived loudness over time. A file can have a high peak but still sound quiet overall. Streaming platforms normalize by LUFS, not peak.
- Is my audio uploaded to a server?
- No. SonoCut processes everything entirely in your browser using WebAssembly. Your audio file never leaves your device, travels over a network, or touches any server. This makes it safe for sensitive recordings — meetings, client calls, medical audio, personal voice memos.