Audio Normalizer

About Audio Normalizer

This online audio normalizer helps you standardize audio specifications for different platforms and use cases. Convert between mono and stereo, adjust bitrate for file size control, and change samplerate for compatibility. Perfect for optimizing audio for podcasts, YouTube, TikTok, music streaming, ringtones, and telephony systems. See also our Audio Recorder and Audio Joiner.

What does audio normalization actually do?

Normalization scans an audio file, measures its loudness or peak level, and applies a single uniform gain so the file matches a chosen target. It does not change dynamic range, EQ, or timing — every sample is multiplied by the same factor. Peak normalization sets the loudest sample to a chosen ceiling (often -1 dBTP or 0 dBFS); loudness normalization (LUFS-based) sets the overall perceived loudness, ignoring brief peaks. The result is that quiet files become louder and over-hot files become quieter without altering their character. This makes a podcast season, an album, or a YouTube upload sound consistent track-to-track. Normalization is non-destructive in this tool: the original samples are recoverable by applying the inverse gain.

What is the difference between peak normalization and LUFS (loudness) normalization?

Peak normalization just rescales the file so its highest sample reaches a target dBFS or dBTP value — it does not consider how loud the file actually sounds. Two tracks both peak-normalized to -1 dBFS can still differ by 10 LUFS in perceived loudness because one might be a dense pop master and the other a sparse acoustic ballad. LUFS normalization measures integrated loudness across the whole file using ITU-R BS.1770-4 and EBU R128 weighting, which mimics human ear response. To deliver consistent playback across an album or podcast feed, always normalize to LUFS, not to peak. Streaming platforms (Spotify -14, YouTube -14, Apple Music -16, broadcast EBU -23) all use LUFS — peak-only normalization is a legacy CD-era practice.

What LUFS target should I aim for?

Pick the target that matches your delivery destination, because most platforms will down-adjust louder material and turn up quieter material to their own internal reference. For Spotify, YouTube, Tidal, and Amazon Music, normalize to -14 LUFS integrated with -1 dBTP ceiling. Apple Music uses -16 LUFS. Broadcast TV in Europe (EBU R128) requires -23 LUFS with a -1 dBTP ceiling; US broadcast (ATSC A/85) uses -24 LKFS. Podcast distributors (Apple, Spotify) typically expect -16 LUFS mono or -19 LUFS stereo. Audiobook (Audible / ACX) requires -23 to -18 LUFS, peaks below -3 dBTP, and noise floor under -60 dBFS. Mastering louder than -8 LUFS will be turned down on every major platform, wasting your dynamic range.

Will normalization cause clipping or distortion?

Normalization itself never introduces clipping — the algorithm checks the peak first and limits the gain so no sample exceeds the chosen ceiling. However, raising LUFS while keeping peaks below a true-peak ceiling may be impossible without a limiter: a sparse track with high peaks but low average loudness cannot reach -14 LUFS by gain alone, because the peaks would clip first. In that case the normalizer must either skip the LUFS target (peak-limited) or apply a brick-wall limiter to tame the peaks before boosting. True-peak (dBTP) measurement uses oversampling to catch inter-sample peaks that ordinary dBFS meters miss — these can clip lossy codecs even when sample peaks look fine. Set ceiling to -1 dBTP for safety with MP3/AAC encoding.

Should I normalize each track individually or as an album?

It depends on the artistic intent. Track normalization sets every song to the same LUFS, so a quiet ballad ends up as loud as a dense rocker — convenient for shuffle play but destroys the dynamic arc the artist may have intended. Album normalization measures the entire album as one continuous program, applies one common gain, and preserves the loud-quiet relationships between tracks — the way the artist sequenced them. For most pop and electronic releases, use track normalization to match streaming expectations. For classical, jazz, film soundtracks, and concept albums, use album normalization. Spotify offers a user toggle between the two modes; YouTube applies track normalization only. When in doubt, deliver in album mode and let platforms choose.

What is dBTP (true peak) and why does it differ from dBFS sample peak?

dBFS sample peak measures the level of individual discrete samples in the file. True peak (dBTP) measures the level of the analog waveform that would be reconstructed between samples during digital-to-analog conversion or transcoding. Because the analog wave passes between sample points, its actual peaks can exceed the digital sample peaks by 1-3 dB. A file showing -0.1 dBFS sample peak may produce +2 dBTP after MP3 encoding, causing audible clipping on consumer playback systems. True-peak meters use 4x or 8x oversampling per ITU-R BS.1770-4 to detect these inter-sample peaks. Always set your normalization ceiling using dBTP, not dBFS — -1 dBTP for streaming, -2 dBTP for safety on lossy codecs, -3 dBTP for podcast and broadcast where peak headroom matters.

How is integrated, short-term, and momentary loudness different?

These are three timescales in the EBU R128 / ITU-R BS.1770 specification. Momentary loudness is measured over a sliding 400 ms window — useful for catching individual transients and brief peaks. Short-term loudness uses a 3-second sliding window and reflects what you would call "the loud chorus" or "the quiet verse." Integrated loudness measures the entire program in one number, gated to exclude silences below -70 LUFS absolute and -10 LU relative to ungated loudness, so it represents the overall perceived loudness of the file. For normalization targets, always use integrated. For delivery compliance (broadcast, podcast), also check max short-term and max momentary. Loudness Range (LRA) measures the spread between loud and quiet sections — high LRA means dynamic, low LRA means consistently loud.

Why do my normalized files sound quieter than commercial releases?

Commercial pop and EDM releases from 1996-2015 were mastered far louder than streaming standards — often -8 to -6 LUFS integrated with heavy limiting, the so-called "Loudness War." When you normalize your master to -14 LUFS for Spotify it will play back at the platform's preferred level, but if you A/B against a 2010 commercial track it may sound quieter because that older track was mastered too hot and Spotify turns it down dynamically. The fix is psychological, not technical: compare against modern reference tracks mastered for streaming (post-2017 releases), and trust the LUFS target. If you want more apparent loudness within the same LUFS budget, work on transient sharpness, midrange clarity, and stereo width — not on the volume knob.