Skip to main content

Audio File Normalization

Audio File Normalization

Normalizes an audio file using two passes of FFmpeg's loudnorm filter, implementing EBU R128 loudness normalization.
This ensures consistent perceived loudness across audio files while preventing clipping and controlling dynamic range.

How It Works​

The node performs a two-pass normalization:

  1. First pass: Measures the loudness statistics of the input file (input_i, input_lra, input_tp, input_thresh) using the loudnorm filter with print_format=json.
  2. Second pass: Applies normalization based on the measured statistics, targeting the specified Integrated Loudness, Loudness Range, and True Peak. This ensures precise adjustment without introducing distortion.

Target Integrated Loudness​

Parameter: TargetLoudness
Units: LUFS
Default: –14 LUFS

Controls the overall perceived loudness of the audio. A LUFS (Loudness Units Full Scale) value defines the average volume level across the track.

  • Music (default): –14 LUFS is a good balance for albums, streaming, and digital releases.
  • Other broadcast content: –24 LUFS is typical.
  • Podcasts or voice content: –14 to –16 LUFS is common to maintain consistent voice levels.

Setting this too high can make music sound harsh; too low can make it quieter than other tracks in a playlist.

Target Loudness Range​

Parameter: LoudnessRange
Units: LU
Default: 8 LU

Defines how much variation in loudness is allowed across the program. Smaller values reduce the dynamic range, making quiet and loud parts closer in volume.

  • Music albums: 7–10 LU preserves dynamics while keeping overall loudness consistent.
  • Background music or voice-heavy tracks: 4–6 LU makes the audio more uniform.
  • High-dynamic content (classical, cinematic music): 11–20 LU maintains strong dynamic contrasts.

Maximum True Peak​

Parameter: TruePeak
Units: dBTP
Default: –1.5 dBTP

Sets a ceiling for the highest instantaneous peak in the audio signal to avoid clipping.

  • Music (default): –1.5 dBTP prevents clipping while allowing a healthy peak headroom.
  • Broadcast or streaming platforms: –2.0 dBTP is commonly used.
  • Setting this too low slightly reduces overall loudness but improves safety margin for playback devices.

Practical Tips​

  • Two-pass normalization is more accurate than single-pass because it first measures the actual audio levels.
  • For albums or playlists, keep TargetLoudness, LoudnessRange, and TruePeak consistent across all tracks.
  • For high-dynamic music genres (classical, jazz), consider slightly higher LoudnessRange.
  • For podcasts or voice tracks, reduce LoudnessRange to make speech more even and intelligible.
  • If fields are left blank (or when upgrading older flows), the node will automatically use the defaults: –14 LUFS, 8 LU, –1.5 dBTP.

Outputs​

  1. Audio file normalized and saved to a temporary file. This file can then be passed to downstream nodes for encoding, tagging, or packaging.