Why Your Track Sounds Different on Every Streaming Platform — What to Know Before You Upload

If you've finished a mix, sent it to mastering, got it back sounding polished and loud, uploaded to a distributor — and then played it on Spotify and thought it sounded flat or thin compared to other tracks — this article is for you.

The answer is loudness normalization. Every major streaming platform adjusts the playback volume of every track to hit a loudness target. Your file isn't broken. But if your master was built without accounting for it, you're either losing quality at playback or you've been optimizing for a standard that streaming made obsolete.

What loudness normalization actually does

When you upload a track to Spotify, Spotify doesn't play it at whatever level you exported at. It measures the integrated loudness of your file and applies a gain offset to bring playback to their target level.

The file itself is not re-encoded or modified for this. The adjustment happens at the playback layer — the same way turning a volume knob down doesn't change what's on the record. If your track is louder than the platform's target, it will be played back quieter. If it's quieter than the target, it plays at its natural level (Spotify does not apply upward gain to quiet tracks by default).

Users can turn normalization off in their Spotify settings. Most don't. You should master with normalization on as the assumed listening context.

Platform targets

Here are the approximate integrated loudness targets for the major platforms as of 2026. These are the levels platforms aim to normalize playback to:

If you're mastering for general streaming release with no specific platform priority, −14 LUFS integrated is the practical baseline. It meets Spotify's default target and is close enough to the others that a single master works cleanly across all of them.

What LUFS actually measures

LUFS stands for Loudness Units relative to Full Scale. It's a perceptual measurement — it attempts to approximate how loud something sounds to a human ear, not just how high the waveform peaks. Two tracks can have identical peak levels but very different perceived loudness. LUFS accounts for this.

There are three measurement windows you'll see in mastering tools:

• Integrated LUFS — the average loudness across the entire program (the whole track from start to finish). This is the number platforms use for normalization decisions.
• Short-term LUFS — average loudness over a rolling 3-second window. Useful for monitoring dynamic shifts during mastering — choruses, drops, breakdowns.
• Momentary LUFS — average over a 400ms rolling window. Shows the loudness of individual moments — a snare hit, a vocal phrase, a transient spike.

Most mastering tools (iZotope Ozone, FabFilter Pro-L 2, Nugen Audio MasterCheck, Youlean Loudness Meter) display all three in real time. The number that determines how platforms treat your track is integrated LUFS — the full-program average.

The over-limiting problem

Here is the core issue with mastering loud for streaming: if you push your master to −8 or −9 LUFS through heavy limiting, Spotify will reduce playback gain by approximately 5 to 6 dB to bring it to −14 LUFS. On the surface, that sounds like a wash.

The problem is what that 6 dB reduction reveals.

Heavy limiting works by clamping transients, reducing dynamic range, and using the consistent level to create an impression of density and energy. At the original loud level, these processing artifacts — the pumping on kick hits, the harshness on cymbals, the smeared transient response — are partially masked by the sheer volume. They feel like loudness, not distortion.

Turn the whole thing down 6 dB, and the listening context changes. Those same artifacts sit against a quieter background. The pumping becomes apparent. The high end sounds congested. The kick punch is gone. What felt loud and exciting at −8 LUFS sounds flat and dense after normalization to −14 LUFS.

A track mastered with care at −14 LUFS will sound more dynamic, more detailed, and more alive than the same track crushed to −8 LUFS and normalized back down.

This is the central irony of mastering loud for streaming: the gain you added through limiting is removed by normalization, but the damage done by that limiting stays.

True Peak

Every mastered file for streaming should have a True Peak ceiling of −1.0 dBTP or lower. Here's the reason.

When streaming services encode your WAV for delivery, they use lossy codecs — Spotify uses OGG Vorbis, Apple Music uses AAC, YouTube uses AAC or Opus. During encoding, the codec can create inter-sample peaks that exceed your original file's peak level. A file that measures exactly 0 dBFS peak can encode to a file that clips during playback.

Setting True Peak to −1.0 dBTP gives the codec enough headroom to complete the encoding process without introducing distortion. Some engineers use −0.5 dBTP for extra safety, particularly for tracks with significant high-frequency content — fast transients and bright material are most susceptible to encoding-induced clipping.

True Peak is a separate measurement from sample peak. Standard limiters show you sample peak. You need a True Peak meter to set this correctly — most modern mastering limiters include one.

What a streaming-ready master looks like

Putting it together, here's the practical target for a master built for streaming distribution:

• Integrated loudness: −14 LUFS (or −16 LUFS if prioritizing Apple Music). If you want a single master for all platforms, target −14 LUFS — it normalizes cleanly on Spotify and YouTube, and Apple Music will apply its own normalization on top.
• True Peak: ≤ −1.0 dBTP. Measured with a True Peak meter, not standard sample peak.
• Dynamic range: preserved. Limiting used to control occasional peaks and add polish, not to maximize perceived loudness for an era when loudness maximization no longer helps you.
• Format: 24-bit WAV at the original sample rate (44.1 kHz or 48 kHz). Distributors will handle conversion from there.

Does this make the loudness war irrelevant?

Mostly, yes — for streaming. The loudness war of the 1990s and 2000s made commercial sense at the time: louder songs stood out on radio, and CDs had no automatic normalization. Producers and labels pushed masters as loud as possible, accepting significant dynamic range loss in exchange for perceived competitive advantage.

Streaming normalization eliminates that advantage. A track mastered at −8 LUFS and a track mastered at −14 LUFS will play back at the same perceived volume after normalization. There is no loudness edge to gain.

What remains is the quality difference. The track that was mastered with headroom will sound better — more dynamic, more punchy on transients, better stereo imaging, cleaner high end — because the limiter didn't have to work as hard. The track that was crushed for loudness will sound worse at normalized playback levels, not better.

The loudness war had one winner: dynamics. Streaming just took a while to enforce the result.

One caveat worth noting: if your music is primarily intended for live DJ use, film sync, or licensing contexts where tracks are played without streaming normalization, the calculation changes. But for consumer streaming releases by independent artists, optimizing for −14 LUFS integrated is the correct approach in 2026.