You have finished the mastering chain and the track sounds loud and polished in your studio. Before you upload it, you need to check your master against the real conditions it will face: lossy streaming codecs, phone speakers, car systems and loudness normalisation. Skipping this step is how good mixes end up sounding distorted or thin on release day. This guide is a repeatable pre-flight checklist that takes about ten minutes and catches the mistakes that are almost impossible to hear when you are too close to a project.
Measure loudness and true peak first
Numbers do not lie the way tired ears do. Load a dedicated loudness meter such as Youlean Loudness Meter on your master bus and play the whole song from start to finish. You are checking three values:
- Integrated LUFS across the entire track. Streaming platforms normalise around -14 LUFS integrated, so know where you sit. If you mastered a pop single to -9 LUFS, expect Spotify to turn it down by roughly 5 dB; that is fine as long as it was a deliberate choice and the track still sounds good turned down.
- True peak (dBTP). This must sit at or below -1 dBTP. Inter-sample peaks appear when your WAV is transcoded to AAC or Ogg, and a master that peaks at 0 dBFS can clip on playback even though it looked clean in your DAW. If your true peak is above -1, lower your limiter ceiling.
- Loudness range (LRA). This tells you how dynamic the track is. A hyper-compressed master shows an LRA of 3 or less and often sounds fatiguing; most well-mastered singles land between 4 and 9 LU.
Trust these meters over your ears for level, because ten minutes of A/B listening will convince you that louder is better every single time.
Check your master for codec and mono problems
Streaming does not play your WAV. It plays a compressed copy. To hear what listeners hear, encode a test copy at 128 or 256 kbps AAC or MP3 and listen back. Watch for two failure modes: harsh, swirly artifacts on cymbals and vocal sibilance, which usually means your master is too bright or too loud, and a squashed transient response, which means the limiter is working too hard. If the encoded copy sounds noticeably worse, back your limiter off by 1 to 2 dB and re-check.
Next, collapse the track to mono. A huge amount of real-world listening happens on single-speaker phones, Bluetooth pills and laptop speakers that sum to mono. When you fold to mono, listen for anything that disappears or gets weak, especially wide stereo effects, hard-panned guitars or a bass that was spread with a stereo widener. If the vocal or kick loses power in mono, you have a phase problem in the mix, not the master, so go back a step. A reference tool like ADPTR MetricAB has one-click mono and correlation views that make this fast.
Compare against a reference at matched loudness
The single most useful check is a loudness-matched A/B against a professionally mastered song in the same genre. Import a reference track, match its integrated LUFS to yours so volume does not bias you, then flip between them on your chorus. You are listening for tonal balance, not loudness: is your low end boomier, your top end harsher, your midrange more cluttered? Broad differences of more than a couple of dB in a frequency region are worth fixing with a gentle 1 dB master EQ move rather than a big one. If you find yourself wanting to add 4 dB anywhere, the fix belongs in the mix.
Also check spectral balance visually. A linear-phase analyzer or the tonal balance view inside iZotope Ozone overlays your track against a target curve for the genre, which is a helpful second opinion when your ears are fatigued. Do not chase the curve exactly, use it to catch obvious excesses.
The final playback checklist
Before you export the deliverable, run through this quick list on the actual master:
- Integrated LUFS is in your intended target range for the genre and platform.
- True peak is at or below -1 dBTP with the limiter engaged.
- No audible clipping, pumping or distortion on the loudest chorus.
- The intro fade-in and outro fade-out are clean, with no clicks at the file boundaries and at least a few hundred milliseconds of silence trimmed correctly.
- Nothing important vanishes in mono.
- The AAC or MP3 encode still sounds like the master.
- You listened on at least three systems: monitors, headphones and a phone speaker.
Export a 24-bit WAV at your project sample rate for delivery and keep an archived copy of both the master and the pre-master mix. If any of these checks fail and you cannot fix it quickly, it is worth getting a second set of ears, and an AI service can act as a fast reference point; our eMastered review covers when that is worth using. For the full mastering workflow that feeds into this checklist, see our guide to mastering a mix at home, and grab the free metering utilities in the Violet Recording tools hub. Checking your master properly is the difference between a release that travels and one that only sounds right in the room it was made in.
