The Loudness War — Why Your MP3 Collection Sounds Uneven
Since the mid-1990s, music has gotten progressively louder. Mastering engineers pushed levels higher and higher, competing for attention on radio and in shuffled playlists. This "loudness war" means that MP3 files from different eras and genres can differ by 15+ dB in perceived volume:
| MP3 Source | Typical Loudness | Why |
|---|---|---|
| 2010s–2020s pop/EDM | -6 to -10 LUFS | Peak loudness war — heavily limited masters |
| 2000s rock/hip-hop | -8 to -12 LUFS | Loud mastering, moderate dynamic range |
| 1990s grunge/alternative | -12 to -16 LUFS | Before the worst of the loudness war |
| 1980s–early 90s | -14 to -18 LUFS | Pre-loudness-war mastering, more dynamics |
| Classical / jazz | -20 to -28 LUFS | Wide dynamic range preserved intentionally |
| Podcast / spoken word | -14 to -20 LUFS | Varies by production quality |
That's a potential 20+ dB spread between your quietest jazz record and loudest EDM track. Without normalization, converting these to WAV carries the volume chaos straight through.
Which LUFS Target to Choose
The converter above offers three normalization presets. Each targets a different use case:
| Preset | Target | True Peak | Best For |
|---|---|---|---|
| Podcast (-16 LUFS) | -16 LUFS | -1.5 dBTP | Spoken word, Apple Podcasts, audiobooks |
| Streaming (-14 LUFS) | -14 LUFS | -1 dBTP | Music playback, DJ prep, Spotify/YouTube level |
| Broadcast (-23 LUFS) | -23 LUFS | -1 dBTP | EBU R128, European TV/radio, film production |
Platform reference: Spotify and YouTube normalize to -14 LUFS. Apple Music targets -16 LUFS. European broadcast follows EBU R128 at -23 LUFS. Amazon Music and Tidal use -14 LUFS. Netflix dialogue sits at -27 LUFS (dialogue-gated).
Most common choice: Use Streaming (-14 LUFS) for music playlists, DJ prep, and general playback. Use Podcast (-16 LUFS) for spoken word and audiobooks.
How Loudness Normalization Works
When you upload an MP3 file and select a normalization preset, Convertio performs two operations in a single step:
- MP3 → WAV decoding: The lossy MP3 audio is decoded to uncompressed PCM (WAV)
- Loudness normalization: FFmpeg's
loudnormfilter measures the audio's integrated loudness in LUFS and applies a constant gain adjustment to reach the target
The normalization is linear — a single gain value applied uniformly across the entire file. It does not compress dynamics, alter frequency response, or change the stereo image. A true peak limiter prevents clipping if the gain boost would push peaks above the safe ceiling.
LUFS vs dB vs RMS: LUFS (Loudness Units relative to Full Scale) measures perceived loudness using a K-weighted frequency curve that matches human hearing. Unlike peak dB or RMS, LUFS accounts for the fact that our ears are more sensitive to midrange frequencies (1–5 kHz) and less sensitive to sub-bass. Two files at the same LUFS will sound equally loud, even if their peak levels differ.
DJ Workflow: Normalize Before Mixing
DJs deal with the loudness problem constantly. A set list might include tracks from five different decades, three different labels, and a dozen different mastering engineers. Without normalization, gain-riding between tracks is a constant headache.
Normalizing your MP3 collection to WAV at -14 LUFS before a gig solves this:
- Consistent gain staging: Every track starts at the same perceived loudness. Your channel faders and gain knobs stay near zero
- No lossy re-encoding: MP3-to-MP3 normalization means decoding and re-encoding — generation loss. WAV output preserves the audio without further compression artifacts
- Hardware compatibility: WAV plays natively on Pioneer CDJ, Denon SC6000, and all professional DJ hardware without format support issues
- Headroom for live EQ: Starting at -14 LUFS gives you room to boost bass or mids live without clipping the output
For club environments where you want extra headroom for live EQ work, consider -16 LUFS instead — it provides 2 dB more room before clipping while keeping tracks consistent with each other.
Car Audio & Portable Playlists
In a car, volume jumps between tracks are not just annoying — they're a safety issue. Reaching for the volume knob at highway speed is a distraction. Normalizing your driving playlist to a single LUFS target eliminates these jumps entirely.
Recommended approach for car audio:
- Target: Streaming (-14 LUFS) for most systems
- Format: WAV gives you the cleanest output, though file sizes are larger (a 4-minute track is ~40 MB)
- Storage: A 64 GB USB drive holds 1,500+ normalized WAV tracks — enough for most driving playlists
For background music in retail stores, restaurants, and events, the same principle applies. Normalize all tracks to -14 LUFS so staff never need to adjust the volume between songs.
DAW Editing & Production
When importing MP3 files into a DAW (Audacity, Logic Pro, Reaper, Pro Tools), normalizing to WAV first gives you a consistent starting point:
- Sample libraries: If you're building a sample library from MP3 sources, normalizing ensures every sample plays at the same volume when triggered
- Podcast editing: Interview recordings from different guests and microphones. Normalize to -16 LUFS for Apple Podcasts compliance before arranging in your timeline
- Video editing: Background music from different sources mixed with dialogue. Normalize music to -18 or -23 LUFS so it sits below the dialogue track
- Transcription: Consistent volume across files means your speech-to-text tool performs reliably on every recording
Why WAV for production? WAV is uncompressed PCM — every DAW reads it natively, there's no decoding overhead, and any further processing (EQ, compression, effects) is done on pristine audio. Normalizing to WAV avoids the quality loss of MP3-to-MP3 re-encoding.
Normalization vs Dynamic Compression
These two audio processes are often confused, but they do fundamentally different things:
| Feature | Loudness Normalization | Dynamic Compression |
|---|---|---|
| What it does | Adjusts overall volume | Reduces dynamic range |
| How | Single constant gain (like a volume knob) | Time-varying gain (loud parts get quieter) |
| Dynamics | Fully preserved | Reduced — quiet and loud parts get closer |
| Use case | Match volume across tracks | Control volume within a single track |
| Reversible? | Yes (apply opposite gain) | No (information is lost) |
Convertio uses only normalization. Your audio's natural dynamics — the difference between the quietest whisper and loudest crescendo — are preserved exactly as the artist intended. The only change is the overall volume level.
True Peak vs Sample Peak
Digital audio stored as samples can produce inter-sample peaks — peaks between samples that exceed the highest sample value. This happens because a DAC reconstructs a continuous waveform from discrete samples, and the reconstructed wave can overshoot.
True Peak (dBTP) measurement accounts for these inter-sample peaks by oversampling the signal at 4x. The converter's limiter uses True Peak limiting to prevent clipping on any playback system:
- Streaming preset: True Peak ceiling at -1 dBTP (matches Spotify, YouTube requirements)
- Podcast preset: True Peak ceiling at -1.5 dBTP (extra safety margin for Apple Podcasts)
- Broadcast preset: True Peak ceiling at -1 dBTP (EBU R128 requirement)
This means your normalized WAV files will play without distortion on any system, from cheap earbuds to high-end studio monitors.