You bounce a track, drop it into a playlist next to a release you love, and yours sounds small. Thinner, further back, less confident. The instinct is to reach for the limiter and push another 3 dB into it, and for a while that feels like progress. Then it stops translating. The kick loses its punch, the cymbals get crispy, and the loudness you fought for evaporates the moment a streaming service touches it. So how loud should your master actually be? The honest answer is a range, not a single magic number, and it depends on where the track is going to play.
Why your master sounds quiet next to commercial releases
The first thing to understand is that a released master almost never sounds quiet because the producer forgot to turn it up. It sounds finished because the low end is controlled, the mid-range is clear, and the transients still have shape. When your track sounds small next to it, the gap is usually balance, not level. A muddy 200 Hz to 500 Hz region, a sub that wanders 4 dB hot, or a kick that has no defined click around 2 to 5 kHz all read as "smaller" even at the exact same integrated loudness.
Producers chase this with the limiter because pushing the level is the one move you can do in ten seconds. The trouble is that limiting a mix that is not balanced just makes the problems louder. If you want to close the gap properly, the work happens upstream in the mix, and the master only puts a final polish on something that was already solid.
Streaming normalizes to around -14 LUFS (what that means for you)
Here is the part that changes the whole calculation. Spotify and YouTube normalize playback loudness to around -14 LUFS integrated, and Apple Music aims slightly lower at about -16 LUFS. If you master your track to -6 LUFS by slamming a limiter, the platform turns it back down to roughly -14 LUFS for the listener anyway. You did not get louder. You got the same playback level, minus the dynamics you crushed to get there.
So for a streaming-first release, mastering hot past -14 LUFS buys you nothing on the platform and costs you transient impact. That does not mean you should master to exactly -14 LUFS and stop. Real releases sit in a window, not on a line, and a master that measures -9 or -10 LUFS with the level coming from a well-controlled low end and a punchy kick will survive normalization with its character intact. The number to stop worrying about is the headline LUFS figure. The number to obsess over is whether the track still hits after the platform turns it down.
Club masters play by different rules (dynamics over peak loudness)
If your track is built for a room with a real PA, the streaming math does not apply. A club system is not normalizing anything, and the engineer riding the booth has plenty of headroom on the desk. What matters there is not raw integrated loudness but dynamics and transient impact: a kick that slams with space around it, a low end that moves air without turning to mush, and a master that has not been brickwalled into a flat wall of sound.
Many club-oriented masters sit looser than streaming-optimized pop, often a few LU more dynamic, precisely because that breathing room is what makes a kick feel physical on an 18-inch driver. Push a techno master too hard and you sacrifice the transient that the whole genre is built on. The loudest record in the room is rarely the one that moves the floor. The one that moves the floor is the one whose kick still has a sharp leading edge when it is playing at 110 dB.
What hundreds of released masters actually measure
Rather than guess, it helps to look at what real releases do. TrackSensei calibrated its loudness targets on a corpus of 563 released tracks across seven electronic categories, including house, tech-house, and four tiers of techno. The first lesson from that data was that there is no single correct LUFS value. Real masters spread across a band, so the targets were set to the 25th to 75th percentile window of actual releases rather than to one number. If your master lands inside that window for your genre, it is in good company. If it sits well below it, it will read as quiet even after normalization, and if it sits well above, you are almost certainly crushing dynamics other producers in the genre chose to keep.
The second lesson was about where loudness comes from. The body of a master lives in the 60 to 250 Hz bass band far more than in the pure sub below 60 Hz. Masters that feel big are not the ones with the most 30 Hz rumble, they are the ones with confident weight in the 60 to 250 Hz region where the speaker can actually reproduce it and the ear can actually feel it. If your track feels thin, the fix is usually to build body in that bass band, not to add sub you cannot hear on most systems anyway.
True peak under -1 dBTP, and why it matters after codecs
Loudness is only half the master. The other half is the true peak ceiling, and the safe number here is firm: keep true peak under -1 dBTP. The reason is what happens after you export. When your WAV gets encoded to a lossy format like MP3 or AAC for streaming, the codec reconstructs the waveform and can overshoot the sample peaks it started from. A master that hit exactly 0 dBFS on the samples can clip past full scale once it has been through an encoder, and that clipping is audible as harsh, crackly distortion that was not in your bounce.
Setting your limiter or true peak ceiling to -1 dBTP leaves enough margin for the codec to do its reconstruction without going over. Use a limiter with true peak detection turned on, not just sample peak, because the two readings can differ by more than a dB on bright, transient-heavy material. One dB of headroom costs you nothing perceptually and protects the master on every platform it lands on.
Balance beats volume: less limiting, more loudness
The producers who get loud, clean masters are usually the ones limiting the least. That sounds backwards until you realize that a balanced mix already sits at a higher average level before the limiter ever touches it, because no single frequency band is spiking and triggering gain reduction. Tame the 200 Hz to 500 Hz mud, control the sub so it is not eating 4 dB of headroom on every kick, and your track gets louder on its own. Then the limiter only has to catch the last 2 or 3 dB instead of papering over a mix that was never under control. When a single master bus genuinely cannot fix the problem, because the kick and bass are fighting for the same 2 dB of headroom, stem mastering gives you more control: bounce a handful of grouped stems instead of one stereo file and you can balance and limit the low end against the rest before it ever reaches the ceiling.
If you are not sure whether your loudness problem is a mix issue or a master issue, that is a question a meter can answer faster than your ears can after a long session. Running a bounce through AI track feedback will tell you in numbers how your integrated loudness and true peak compare to released tracks in your genre, and whether your sub or low mids are stealing the headroom you are trying to fight for with the limiter. It is also worth understanding where mastering feedback ends and mix feedback begins, because the two solve different problems and confusing them is how producers spend an hour on the wrong stage.
So, how loud should your master be? Loud enough to sit inside the range real releases occupy in your genre, with true peak under -1 dBTP, and not a single dB louder than the point where your kick starts losing its punch. Aim for balance first and the loudness mostly takes care of itself. When you want a number on where your own bounce actually lands, drop it into the analyzer and let the data tell you whether you are short, in the pocket, or over the edge.