This post explains why AI-generated tracks tend to sound quiet, what loudness actually means on streaming platforms, and the practical options for fixing it on a Mac.
The missing step: mastering
AI music generators compress an entire production pipeline — composition, performance, mixing — into a single generation. What you download is best thought of as a mix, not a finished master.
Mastering is the final processing stage applied to a mix before release. It typically involves raising the overall loudness to a competitive level, controlling peaks so the track doesn't clip or distort after encoding, and smoothing the tonal balance so the track translates across speakers and headphones. Commercial releases have all been through this stage. A raw AI export usually hasn't — which is exactly why the difference jumps out in back-to-back listening.
Loudness, LUFS and true peak in two minutes
Perceived loudness is measured in LUFS (Loudness Units relative to Full Scale), based on the ITU BS.1770 standard. Two facts matter for anyone releasing music.
First, major streaming platforms normalize playback loudness. Spotify, for example, adjusts tracks to −14 LUFS during playback on supported devices, and Premium listeners can switch between Loud (−11), Normal (−14) and Quiet (−19) settings (Spotify's official documentation).
Second, Spotify's own mastering recommendation is to target −14 LUFS integrated and keep true peaks below −1 dBTP (or below −2 dBTP if you master louder than −14), so that lossy encoding doesn't introduce distortion (same source).
"If platforms normalize loudness anyway, why bother mastering?"
A fair question — with three concrete answers.
Normalization isn't everywhere. On Spotify, the web player and some third-party devices don't apply loudness normalization (Spotify support). Files you share directly, play in DJ sets, or drop into videos aren't normalized at all. In those contexts, a quiet master simply sounds quiet.
Normalization only matches average loudness. It doesn't fix uncontrolled peaks, harsh or dull tonal balance, or inconsistency between sections. A master that merely gets "turned up" by the platform still sounds less dense and less controlled than one that was properly limited and balanced.
And quiet masters aren't boosted without limits: for softer tracks Spotify applies positive gain but leaves headroom to protect lossy encodings — so an extremely quiet master may still end up below the platform target.
Three ways to master a Suno track on a Mac
Option 1: Do it yourself in a DAW. Logic Pro, Ableton Live or even GarageBand plus a limiter will get you there. Maximum control, but you'll need to learn metering (LUFS, true peak), limiting, and EQ. If you enjoy audio engineering, this is a rewarding path — but it's a genuine skill with a genuine learning curve.
Option 2: Cloud AI mastering services. Upload your track to a web service and download a mastered version. These work well and require no engineering knowledge. Trade-offs: your audio gets uploaded to someone else's servers, processing depends on your connection, and pricing is typically subscription-based or per-track.
Option 3: A local mastering app on your Mac. The same drag-and-drop simplicity, but processing happens entirely on your machine: no upload, no waiting on a queue, and it works offline.
Where Tonekai fits (full disclosure: it's our app)
We build Tonekai, an AI mastering app for macOS, so option 3 is the one we can speak to concretely.
Tonekai's workflow is a single drop: drag an audio file onto the app, and it analyzes the track and adjusts loudness, balance and peaks automatically. You can A/B compare the original and the mastered version before exporting. Everything runs locally on Apple Silicon — your audio never leaves your Mac. The app is free to try, and a one-time Pro Unlock (no subscription) adds WAV/AAC export and unlimited mastering. It requires macOS 14 or later.
We won't claim it replaces a professional mastering engineer — nothing automatic does. But for getting an AI-generated track from "raw export" to "release-ready loudness and polish" in under a minute, it's exactly the tool we wanted for our own tracks.
Download Tonekai on the Mac App Store — free to try.
A simple workflow for Suno tracks
- Download the best quality you can. On Suno, WAV downloads are available on Pro and Premier plans; the free plan provides MP3 (Suno help center). WAV is preferable for mastering, but a high-bitrate MP3 is workable.
- Master the file — in your DAW, a cloud service, or locally (with Tonekai: drop the file, wait a few seconds, A/B compare).
- Check the result against a commercial reference track at matched volume.
- Export and release. For streaming, a master around −14 LUFS with true peaks under −1 dBTP follows Spotify's own guidance.
Frequently asked questions
Can mastering fix generation artifacts? No. Mastering polishes what's there; it can't remove glitches, garbled vocals or noise baked into the generation. If the source has artifacts, regenerate first, then master the best take.
Is MP3 good enough as a source? Workable, but WAV avoids stacking a second lossy encode on top of the first. If you plan to distribute to streaming platforms, WAV in → WAV out is the safer chain.
Do I need different masters for different platforms? For most independent releases, one well-controlled master around the common streaming targets is fine. Chasing per-platform loudness targets adds complexity with little practical benefit.
The bottom line
Your Suno track doesn't sound quiet because AI music is inherently worse — it sounds quiet because it hasn't been mastered. Whether you learn to do it in a DAW, use a cloud service, or keep it local with a Mac app, adding that one step is the single biggest jump in perceived quality an AI music creator can make.