Audio Normalizer / Compressor
Compress audio dynamics before peak or LUFS normalization, compare playback, and export local PCM16 WAV.
About this audio normalizer / compressor
Compress audio dynamics before peak or LUFS normalization, compare playback, and export local PCM16 WAV Processes streaming audio and video buffers directly in the browser using HTML5 Media and Web Audio APIs with zero server uploads.
Audio and video decoding pipeline
The execution engine processes input tokens and applies transformations in the browser sandbox. Specifically, Choose local audio, set a sample peak target from −20 to 0 dBFS, then process. Compare the original and result with the two players; download the normalized WAV. Select integrated loudness and a target from −24 to −14 LUFS (default −14). The tool measures gated K-weighted loudness, applies one constant gain, and reports the achieved result. Sample peak mode remains available. Optionally enable dynamic compression before the selected peak or LUFS normalization. Set threshold −60 to 0 dB (default −24), ratio 1 to 20 (default 4), and attack/release 0 to 1 second (defaults 0.003/0.25). A native compressor uses a fixed 30 dB knee. The selected normalization then supplies the final gain.
Worked media transformation example
Scenario: Example file: /examples/audio-stereo.wav; Download the example file and select it in the file input. Set the visible controls to: Target peak (dBFS) = -1 Click “Process locally”. Download the result and inspect: czoa-audio-volume-normalizer.wav — PCM16; 48000 Hz; 2 channels; 72000 frames; peak ≈ 0.89125
Sample input:
Processing: Finds the largest absolute sample across all channels. For target dBFS t, target amplitude is 10^(t/20); every sample is multiplied by target amplitude / input peak.
Illustrative output:
Buffer duration & codec support limits
Operational constraints and validation boundaries: Files must be smaller than 100 MB (100,000,000 bytes). Processing decodes at 48 kHz and exports PCM16 WAV. Sample peak mode retains all decoded channels. Original playback and input codec support depend on the browser. Loudness mode accepts mono/stereo at 48 kHz, at least 400 ms and at most 256 MiB of decoded PCM. Measurement uses 400 ms blocks with 100 ms hops and −70 LUFS absolute/−10 LU relative gates. Gain is limited by sample headroom; a high crest factor can prevent reaching the target. LUFS normalization applies constant gain; optional compression is a separate first step. No true-peak limiting. Compression requires mono/stereo at 48 kHz and decoded PCM no larger than 256 MiB. Active compression adds a 50 ms render tail; ratio 1 bypasses it exactly with no tail. Browser compressor behavior can vary. LUFS input loudness refers to the compressed audio when enabled. This does not provide true-peak limiting.
Audio and video rendering in Audio Normalizer / Compressor depends on browser hardware and supported media codecs. Verify playback compatibility on your target platforms.
In-memory audio-video processing boundary
File buffers and document assets in Audio Normalizer / Compressor are processed in-memory via local browser APIs with zero file uploads to CZOA servers.
HTML5 Media & Web Audio standards
W3C Web Audio API specifications and HTML5 media container standards:
Audio Normalizer / Compressor verification · Audit standards
How to use it
- Enter, paste, or select your input data into the Audio Normalizer / Compressor workspace controls.
- Review parameter fields, units, or options configured for Audio Normalizer / Compressor.
- Click the action button or observe immediate live calculations rendered by Audio Normalizer / Compressor.
Frequently asked questions
Audio Normalizer / Compressor+
Choose local audio, set a sample peak target from −20 to 0 dBFS, then process.
Original file+
Compare the original and processed WAV with the two players; starting one pauses the other. Changing the file, mode, either target or any compression setting clears the previous result.
Normalized PCM16 WAV+
Files must be smaller than 100 MB (100,000,000 bytes). No true-peak limiting.
Normalization mode+
An input sample peak of 0.2 and target of 0 dBFS give a gain of 5. The result is quantized to PCM16; silence cannot be normalized. For input −20 LUFS and target −14 LUFS, the requested gain is +6 dB (about 1.995×).