Start by naming the audible problem: unstable vocals, harsh highs, distortion or poor clarity. The right processing depends on that problem and the available source, rather than a detector score alone.

Preserve the best available source

Suno lists MP3 downloads across its plans and website WAV downloads for Pro and Premier subscribers. Options can vary by song and device. Keep the original download and work on a copy.

Avoid successive lossy encodes. A WAV or FLAC export preserves the processing result without adding another MP3 stage; it cannot restore information already lost from a compressed source.

Listen to three passages

Choose a busy passage, an exposed vocal and a quieter phrase ending. Record the timestamps. Cymbals, consonants, reverb tails and short attacks help reveal changes that are easy to miss in the chorus.

First compare at the same listening level. A louder version may appear more detailed without resolving the problem.

Match the action to the problem

These are starting points, not a universal mastering recipe. Strong high-frequency reduction can hide a problem while removing useful detail.

Measure, then decide by listening

The free inspector measures integrated loudness and true peak, among other properties. Save reports for both source and output. It does not assign an overall quality rating or measure vocal fidelity.

When trying Cleaner, use the same listening passages. Processing that suits one song may be too strong for another. Return to the original when the compromise does not suit the track.

Export and archive

Keep the working master, delivery version and processing settings. Encoding a degraded source as a high-bitrate MP3 does not automatically recover missing bandwidth. Read our MP3 and WAV guide and loudness guide for the next steps.

Prepared with AI assistance. Sources below.

Sources and documentation