Online Mastering Tool

How Online Audio Mastering Works: The Processing Chain Explained

Mastering is the final stage of audio production before distribution. It takes a finished mix and applies a sequence of processing steps: tonal shaping, dynamic control, peak limiting, and loudness normalization. The goal is a file that sounds consistent, competitive, and within the technical specifications of every major streaming platform. This tool applies that same four-stage chain inside your browser, entirely offline, using the Web Audio API.

The chain runs in a fixed order: EQ first, then compression, then limiting, then a loudness measurement pass that calculates the exact gain offset needed to hit your target LUFS. If a positive gain offset is required (the track needs to be louder), the limiter runs a second time to catch any new peaks introduced by the boost. Your audio file never leaves your device at any point in this process.

Stage 1: EQ and Tonal Shaping

The default EQ settings apply three gentle adjustments using standard biquad filter shapes. A low shelf at 200 Hz adds +1.5 dB of warmth to the low end. A mid peak at 3 kHz adds +1 dB of presence, which helps vocals and guitars cut through on small speakers. A high shelf at 8 kHz adds +1.5 dB of air, the term engineers use for the brightness and clarity that makes a master sound open rather than dull.

These are starting points, not universal corrections. A mix that already has sharp highs will sound brittle if you add another +1.5 dB of high shelf. Use the bypass toggle to compare the EQ against the raw mix. If your track already sounds well-balanced in the top end, reduce the high shelf to 0 dB or cut it slightly. The EQ stage here is intentionally conservative. Professional mastering engineers often make adjustments of 0.5 to 2 dB per band across a full session, not broad cuts and boosts.

Stage 2: Compression and Dynamic Control

The compressor uses the browser's native DynamicsCompressorNode, which implements a standard soft-knee feedforward compressor. The default settings are 2:1 ratio, -20 dB threshold, 10 ms attack, and 250 ms release. At these settings, the compressor acts more as a dynamic glue device than a heavy gain reduction unit. A 2:1 ratio means that for every 2 dB the signal exceeds the threshold, only 1 dB comes through. That is a relatively gentle compression curve compared to the 4:1 to 8:1 ratios common in mixbus compression.

The 6 dB soft knee setting smooths the transition into compression so it does not create an audible step at the threshold. This matters for mastering because aggressive hard-knee compression on a full mix often creates pumping artifacts, where the whole track breathes in and out visibly with the kick drum. A soft knee and a moderate ratio on a well-mixed track sounds transparent. To verify that the compression is adding value, listen to the A/B preview after processing with the compressor enabled, then disable it and reprocess to hear the difference.

Stage 3: True Peak Limiting

Every major streaming platform specifies a maximum true peak level. Spotify, Apple Music, YouTube, and Tidal all recommend -1 dBTP. Amazon Music specifies -2 dBTP. True peak is different from standard peak because it accounts for inter-sample peaks, which are peaks that occur between samples during the digital-to-analog reconstruction process and during lossy codec encoding. A file that reads 0.0 dBFS on a standard DAW meter can encode to AAC or Ogg Vorbis and produce a true peak of +0.3 to +0.5 dBTP, causing audible distortion on playback.

The limiter in this tool uses a look-ahead buffer of approximately 5 milliseconds. Look-ahead means the processor reads ahead in the audio stream before outputting each sample, which allows it to reduce gain before a peak arrives rather than reacting to it after the fact. This prevents the clipping artifacts that a hard clipper introduces at high input levels. The release time controls how quickly the gain returns to unity after a peak has passed. Shorter release times (10 to 50 ms) can cause pumping artifacts on material with frequent peaks. Longer release times (100 to 300 ms) sound more transparent but may cause audible level ducking on dense, loud material.

Stage 4: LUFS Loudness Normalization

After stages 1 through 3 are rendered, the tool measures the integrated LUFS of the processed audio using the ITU-R BS.1770-4 algorithm. Integrated LUFS is the average loudness across the entire file, with gating applied to exclude very quiet sections. The standard defines two gating thresholds: an absolute gate at -70 LUFS, which removes near-silence, and a relative gate at 10 LU below the ungated mean, which removes quiet passages that would otherwise pull the average down.

Once the integrated LUFS is measured, the tool calculates the gain offset needed to hit your target. If your processed audio measures at -16.2 LUFS and your target is -14 LUFS, the tool applies +2.2 dB of gain and then runs the limiter a second time to catch any resulting peaks. If your processed audio measures at -13 LUFS and your target is -14 LUFS, the tool applies -1 dB of gain (no re-limiting needed, since reducing gain cannot introduce new peaks).

The platform targets in this tool reflect the normalization behavior of each service as of July 2026. Spotify normalizes to -14 LUFS in its default "Normal" mode. Apple Music uses -16 LUFS through its Sound Check system, which is why masters intended primarily for Apple Music often benefit from slightly more headroom and dynamic range. YouTube normalizes to -14 LUFS. TikTok and Instagram use -16 LUFS with adaptive normalization tied to the xHE-AAC codec, though neither platform publishes exact specifications. For a full breakdown of platform targets and how streaming royalties connect to listener behavior, see the LUFS Loudness Meter and the Streaming Royalty Calculator.

What This Tool Does and Does Not Do

This tool applies a fixed four-stage processing chain to a stereo mix. It does not perform mid-side processing, multiband compression, stereo widening, or harmonic excitation. It does not use AI analysis to detect genre-specific spectral targets. It does not compare your track against a reference master. These are all features of professional mastering software like iZotope Ozone, Izotope Neutron, or service platforms like LANDR and eMastered.

What this tool does well is apply consistent, technically correct loudness normalization to a mix that is already well-balanced. If your mix has obvious frequency problems (a muddy low-mid buildup, a harsh 3 to 5 kHz presence peak, a thin low end), a 3-band EQ with gentle default settings will not fix those problems. Mastering EQ works on a mix the way finishing polish works on a car: it refines what is already there. It does not repair underlying structural issues.

For tracks that are going to radio, sync licensing, or physical release, hire a professional mastering engineer. For demos, social media clips, reference tracks, and self-releases where budget is a constraint, this tool gives you a technically compliant master in under a minute. You can verify the output against platform specifications using the LUFS Loudness Meter after downloading.

Output Format Choices

WAV 16-bit is the standard delivery format for most digital distributors. It is lossless, universally supported, and produces files roughly 10 MB per minute of stereo audio. WAV 24-bit preserves more dynamic headroom and is preferred for archival storage or delivery to engineers who will do further processing. Both WAV formats are lossless, meaning no audio quality is lost from the processing chain output.

MP3 output is lossy. Encoding a processed audio file to MP3 introduces compression artifacts, particularly in the high-frequency content above 16 kHz. For streaming distribution, most platforms transcode your uploaded file to their preferred codec anyway (Spotify uses Ogg Vorbis, Apple Music uses AAC), so the starting format matters less than the LUFS and true peak specifications. For social media uploads where the platform will re-encode the file again, a 320 kbps MP3 is a reasonable choice that keeps file sizes manageable without meaningful quality loss. For everything else, use WAV.

To check how your output compares against specific distributor delivery requirements, the Sample Rate Calculator covers bit depth, sample rate, and file size specifications for major distributors. For guidance on the full production chain from raw session to release-ready master, see the AES TD1008 streaming loudness recommendation, which is the document that established the industry-wide convergence on -14 LUFS as the standard loudness reference for music streaming.

A Note on Browser-Based Processing Speed

Processing speed depends on the length of your audio file and the performance of your device. On a modern laptop with a 5-minute stereo WAV at 44.1 kHz, the full processing chain including two render passes typically completes in 10 to 30 seconds. Longer files, 24-bit inputs, or lower-end devices will take longer. The tool processes the entire file offline using OfflineAudioContext, which means it runs as fast as your CPU allows rather than waiting for real-time playback to complete. You do not need to wait for the track to play through in real time before downloading.