How to master a Suno song so it sounds finished

Suno songs come out of the model sounding almost done. Almost is the problem: on good headphones the top end can feel glassy, on a phone the chorus can shrink, and the loudness was never aimed at a streaming service. Mastering is the last small set of moves that fixes those tells. You do not need a studio for it.

1. Start from the best file you have

If Suno offers a WAV download on your plan, use it. An MP3 has already thrown away detail, and every later step works with what is left. If you only have the MP3, that is fine. It will still improve.

If you can download separate parts (stems), keep them too. A separate vocal lets you turn the voice up later without touching the drums.

2. Soften the sharp edge

Many generated songs carry a brittle bite between roughly 3 and 6 kHz: sibilant S sounds, fizzy cymbals, a vocal that feels like it is scratching. A gentle cut there, plus a de-esser on the voice, takes the edge off without making the song dull. The goal is less fatigue, not less air.

3. Clear the mud

Low-mid haze, roughly 200 to 500 Hz, is what makes a big chorus sound small on a laptop or phone speaker. A small dip there lets the bass and the vocal both come through. Big cuts make the song thin, so listen as you go.

4. Glue, only if it needs it

Compression evens out how far the loud parts jump away from the quiet parts. Many AI songs are already squeezed tight, and squeezing them again makes them flat and tiring. Measure first. If the song is already dense, skip this step.

5. Set the loudness, then hold the peaks

Spotify and YouTube play songs at about −14 LUFS integrated. If your file is louder, they turn it down. If it is quieter, it sounds weak next to the songs around it. Set the song to −14 LUFS, then use a limiter so the loudest peak stays at −1 dB. That headroom keeps the MP3 and AAC versions the services make from clipping.

The one-button version

Orbit Waves does steps 2 to 5 in one press. Drop the file, press Make it sound finished, and it listens to this song for the sharp edge, the mud, jumpy loudness, and a lopsided stereo image. It makes the moves it needs, skips the ones it does not, and writes a sentence telling you what it changed. The file is set to −14 LUFS and the ceiling stays at −1 dB.

Then listen. Press Hear the original, then Back to the mix. If the vocal should sit higher, split the song into stems and move the Voice slider. Save WAV 24-bit for a distributor, MP3 to share.