First, the premise: platforms level everything anyway
Start with the fact that surprises most people. According to Spotify's official documentation, playback is normalized to -14 dB LUFS (per the ITU 1770 standard), and Premium listeners can choose -11, -14, or -19. Which means: if you crush your master into a brick to sound bigger than commercial releases, Spotify simply turns it down at playback. What survives isn't the loudness — it's the side effects of chasing it: flattened dynamics and distortion.
Being too quiet has real costs too. Per the same page, when quiet tracks are turned up, Spotify applies a limiter (engaging at -1 dB, 5 ms attack, 100 ms decay) to prevent clipping. Your track getting shaped by someone else's limiter is exactly what a deliberate master avoids — that's the practical reason mastering belongs in your workflow at all.
Check 1: Loudness — aim for a balanced master around -14 LUFS
Spotify's own recommendation is explicit: "Target the loudness level of your master at -14dB integrated LUFS," keeping true peak "below -1dB TP," and if you master louder than -14, "keep True Peak below -2dB to avoid extra distortion" (same source). Some genres customarily master hotter, and that's a legitimate choice — but the platform itself is telling you that "louder wins" is not how streaming works anymore.
Fresh Suno exports tend to land several dB below that neighborhood, which is why raw AI tracks sound noticeably small next to commercial releases once they hit streaming — we covered the mechanics in why Suno and Udio tracks sound quiet. This is a measure-and-fix job, not a guess-and-hope job: put a LUFS meter on the file and work to a number.
Check 2: True peak — leave headroom before the transcode
Streaming platforms don't play your file as-is; they transcode it to lossy formats (AAC, Ogg) for delivery. A master slammed against 0 dBFS can produce inter-sample peaks that exceed the ceiling during that conversion, adding distortion you never heard on your own machine. That's precisely why Spotify recommends keeping true peak below -1 dB TP for lossy formats. The check is simple: run a true-peak meter over your final file, and if it's over, redo the peak control at the mastering stage. This is the one item your ears will usually miss — only a meter catches it reliably.
Note the order of these first two checks matters: loudness first, then peaks. If you fix peaks and then push loudness, the push can create new overs and you're back where you started. A proper mastering pass handles the two together — raising the integrated loudness while keeping the true-peak ceiling intact — which is exactly the part that's tedious to do by hand and routine for purpose-built tools.
Check 3: Format — never stack lossy on lossy
The delivery rule is one sentence: don't add another lossy conversion. Spotify prefers FLAC (WAV also accepted), 44.1 kHz or higher, delivered at the master's native bit depth — and explicitly asks that you don't process or downconvert before delivery: "We handle all format conversions." DistroKid accepts WAV, FLAC, AIFF, MP3, M4A, and WMA, noting 16-bit / 44.1 kHz WAV is typical.
On the Suno side: WAV downloads are Pro / Premier only (via the website — the Free plan's download button produces MP3s), and per the official FAQ, Free-plan trial downloads are personal, non-commercial use only. So in practice, the shortest path to a release is: spend one download on the WAV, keep every later step local and lossless, and deliver the final master as WAV or FLAC. If all you have is an MP3, loudness and peak fixes still work — but the compression artifacts baked into it won't go away, and converting it to WAV doesn't bring back what the encoder discarded.
Check 4: Inspect with your ears — the things meters don't see
Once the numbers are right, finish with listening — and listen to the whole track once, start to finish, at least one time. Meters summarize; they won't flag a rough vocal moment at 2:41, an ending that cuts off a beat early, or a fade that lands awkwardly. Check the silence and fade at the head and tail of the track. A/B against the original at matched volume — anything sounds "better" when it's simply louder, so matching levels is the only honest comparison. Then play it on small speakers and a phone, because that's where most of your listeners actually are. We've written up the full no-meter routine in five checks you can do with just your ears. If what you hear wrong is a mastering problem, another local pass is free. If it's a generation problem — buried vocals, a muddy arrangement — no amount of post-processing will fix it; use what mastering can and can't fix as the referee.
Where Tonekai fits
We make Tonekai, a mastering app for macOS — so here's the honest statement of what our tool does in this checklist, and nothing more. Tonekai is checks 1 through 3: drop in your downloaded WAV and on-device AI measures and adjusts loudness, tone balance, and peak control, with a matched-volume A/B against the original built in. Processing is entirely local — your unreleased track never leaves your Mac. It's free to try, and Pro is a one-time unlock, not a subscription. Since a release is rarely your last, it's worth noting that track two costs nothing extra.
Download Tonekai on the Mac App Store (free to try)
FAQ
Should I hit -14 LUFS exactly? No need to chase the exact number. With normalization in play, a balanced-sounding master with true-peak headroom beats a numerically perfect one. Spotify's -14 LUFS / -1 dB TP guidance is a good starting point; make the final call with a matched-volume A/B.
Do I need a separate master per platform? Reference levels differ slightly between services, but in a normalized world the practical answer is one well-balanced master with peak headroom, delivered everywhere. Spotify itself asks for your master "exactly as it was mastered."
Will mastering make my release perform better? Not in the plays-and-algorithm sense, no. What mastering fixes is quality: not sounding small next to other releases, and not distorting in the transcode. Be skeptical of any tool — ours included — that promises more than that.
The checklist, in one breath
Four checks before you upload: loudness (measure, aim around -14 LUFS integrated), true peak (below -1 dB TP), format (stay WAV/FLAC end to end, no extra lossy hops), and an ear pass (matched A/B, real-world speakers). None of it needs special hardware, and all of it happens on your Mac, on a file you already own. You spent a download getting this track out — spend twenty minutes making sure it arrives the way you meant it.