Mastering for Streaming: Loudness and Format Guide

Spotify publishes a mastering target: -14 dB integrated LUFS, with true peaks below -1 dB TP. The Apple Music, YouTube, Amazon Music, and Tidal materials reviewed in August 2026 did not state a LUFS target; a 2019 Deezer community response relays a -15 dB LUFS playback-normalization target and warns that it may change. Deliver the lossless format your distributor requires, at 24-bit when that is the native source resolution.

Spotify publishes its mastering figures in its artist documentation. A 2019 Deezer community response also relays a -15 dB LUFS playback-normalization target and warns that it may change. For Apple Music, YouTube, Amazon Music, and Tidal, use the platform materials cited below and keep any unconfirmed number labeled as such.

This guide separates what the platforms document from what has been passing around unsourced, gives you an industry recommendation that does have a primary source behind it, and covers the file specs worth getting right. For where mastering sits in the wider release plan, see How to Plan a Music Release: Step-by-Step Checklist.

How Loudness Normalization Works

Loudness normalization means a service measures how loud a track is, then adjusts playback level so tracks sit closer together. Most of the major services offer some version of it. What differs is the target each one uses, how much of it the service puts in writing, and whether it applies to music at all.

Spotify is specific. Its artist documentation states that it adjusts tracks to -14 dB LUFS under ITU-R BS.1770. It may apply positive gain to softer masters, but preserves 1 dB of headroom for lossy encodings and can stop below -14 LUFS when the track's true peak leaves insufficient headroom. Premium listeners can choose Loud at -11 dB LUFS, Normal at -14 dB LUFS, or Quiet at -19 dB LUFS. Normal is the default setting, and Premium listeners can switch normalization off.

The rest of the field publishes less. Apple's Apple Digital Masters document describes Sound Check, which determines the loudness of a track, stores it in the file's metadata, then uses that metadata to raise or lower the volume of each track. It names no target number. Amazon and Deezer document a listener-side toggle, and Tidal's published support articles cover quality tiers and audio formats.

What that means for your master. A song mastered at -8 LUFS was compressed hard to get there. When a service turns it down to sit with everything else, the compression stays behind. What is left is a flatter version of the same song, played at the same level as a master that kept the distance between its quiet and loud moments.

Loudness stopped being the lever. That distance is.

What Each Platform Publishes

The table below is what survived a source check. Each figure is attributed to where it was found, and every "None found" means the check came back with nothing.

Checked in August 2026 against each platform's own material. That means Spotify's artist help article on loudness normalization, Apple's Apple Digital Masters document and its provider delivery pages, and YouTube's encoding specifications for music videos. It also means Amazon's Amazon Music help pages, Tidal's support library, and Deezer's support article on volume normalization. If a service publishes a mastering figure somewhere that check missed, treat that source as better than this table.

Where the Apple Music Number Comes From

"-16 LUFS for Apple Music" appears in a large amount of mastering advice. It does not appear in Apple's own documentation. The Apple Digital Masters document explains Sound Check, walks through clipping, sample rate conversion and the afclip tool, and gives no LUFS or LKFS figure anywhere in it.

What Apple does publish about level is a headroom recommendation. Players oversample digital data on the way to analog, and Apple's document notes that if the original data sits at 0 dBFS, oversampling can result in clipping. Its recommendation: "leave at least 1 dB of headroom in order to avoid such clipping." Apple also asks providers to avoid clipping the signal, and provides afclip so you can check a file for it.

YouTube is a similar case. Its published encoding specifications for music videos cover codec, sample rate, bit depth and bitrate, with no loudness figure. YouTube's help page for stable volume, the setting that reduces variation in playback level, says it is on by default when watching videos and turned off for YouTube Music and official music videos.

An Industry Recommendation With a Primary Source

The Audio Engineering Society publishes a loudness recommendation for streaming distributors: AESTD1008.1.21-9, dated September 24, 2021. It is a real document with a real number, and it is worth reading before you borrow from it.

The document also recommends that maximum true peak level not exceed -1 dBTP at the codec input of lossy-encoded streams, and that integrated loudness stay above -20 LUFS so quiet material does not disappear on players with limited gain. Normalizing an album's loudest track to -14 LUFS, it notes, typically lands the album's integrated loudness near -16 LUFS.

Now the scope line. AES states that the document "is intended for use by distributors of Internet audio streams and on-demand audio files," and says plainly that it does not make recommendations for the production side. It tells services where to set playback. It does not tell you where to set your limiter.

So What Do You Master To

Use -14 LUFS integrated with true peaks at -1 dB TP or lower as a Spotify-oriented starting point, not a universal delivery requirement. Spotify publishes that mastering guidance, and -14 LUFS also matches AES's distributor recommendation for an album's loudest track. Deezer's platform community reported a -15 dB LUFS playback-normalization target in 2019 but warned that it may change; that figure should not be treated as a current mastering instruction. If your master lands louder than -14 LUFS, Spotify advises true peak below -2 dB TP.

For classical and other wide-ranging material, the AES document treats it as its own case, using an album whose loudest track sits at -18 LUFS as its example. Material built around the distance between soft and loud passages does not belong at a pop target.

Understanding LUFS

LUFS stands for Loudness Units relative to Full Scale. It measures perceived loudness over time, accounting for how human hearing responds to different frequencies. LKFS is the same unit under a different name.

Integrated LUFS: The average loudness of the entire track from start to finish. This is the number platforms use for normalization.

Short-term LUFS: Loudness measured over a 3-second window. Useful for checking section contrast (verse vs. chorus).

Momentary LUFS: Loudness measured over a 400ms window. Shows instantaneous peaks in perceived loudness.

True Peak: The actual maximum amplitude of the waveform, including inter-sample peaks that occur between measurement points. This matters because digital-to-analog conversion can create peaks above 0 dB that can clip and distort if they exceed the playback chain's available headroom, even if your meter shows headroom.

Most DAWs have built-in loudness meters, or you can use free plugins like Youlean Loudness Meter or the meters built into iZotope, FabFilter, or Waves mastering tools.

True Peak and Why It Matters

Standard peak meters measure sample values. When a digital file is converted to analog audio through headphones, speakers, or a phone, the reconstructed waveform can exceed the measured sample peaks. Those are inter-sample peaks; if they exceed the playback chain's available headroom, they can clip and distort.

Apple's document explains the playback end of this. Oversampling upsamples the digital data at four times the original sample rate on the way to analog, and if the original data is at 0 dBFS, that oversampling can produce clipping. If the original was already clipped, oversampling can make it worse.

Spotify makes the same point from the encoding end. Its guidance is to keep true peak below -1 dB TP because that suits lossy formats, and below -2 dB TP for masters louder than -14 LUFS, to avoid extra distortion.

Use a true peak limiter with a ceiling of -1 dB or lower. True peak limiters oversample the signal and catch the peaks that would otherwise land between samples.

Format Specifications

What to Deliver to Your Distributor

Your distributor's spec is the one that binds you. Where the platforms publish their own preferences, they agree on more than the loudness discussion suggests.

Bit depth. Apple asks providers to use only 24-bit sources, and warns against bit-padding or recapturing 16-bit files in 24-bit. YouTube recommends 24-bit for music video audio and accepts 16-bit.

Sample rate. Apple asks providers to deliver the highest native sample rate available, and downsamples to its own production rate with its own sample rate conversion. YouTube recommends 48 kHz for music videos and accepts higher rates. Spotify's lossless playback runs up to 24-bit/44.1 kHz FLAC.

File format. WAV or FLAC. YouTube's specs name FLAC or linear PCM as preferred for sound recordings, and its guidance accepts 320 kbps MP3 when a lossless file is not available.

Channels. Stereo for a standard release. Spatial audio deliverables run through separate workflows with their own requirements.

The old advice to bounce everything to 16-bit/44.1 kHz does not match what the platforms ask for. Send the resolution you have.

What Platforms Do to Your Audio

Distributors deliver your file to streaming platforms, which encode it for different playback situations:

• Spotify: approximately 24, 96, 160 or 320 kbit/s depending on plan and setting, plus lossless playback up to 24-bit/44.1 kHz FLAC. The web player runs AAC.

• Apple Music: 256 kbps AAC encodes, plus Apple Lossless for the lossless tier.

• YouTube: uploaded audio is transcoded into a range of delivery formats.

Other services publish their own quality tiers. Read the service's help pages rather than a third-party summary, which is the lesson of the loudness numbers above. What you control is delivering a clean file that encodes well. If mastering is one deadline among several, Apollo, Orphiq's AI music strategist, generates a release timeline working backward from your release date and readjusts it when dates move.

Genre and Material

The -14 LUFS working target is a starting point, not a rule. Genre practice varies, and the per-genre loudness figures that get quoted around are as unsourced as the platform table they usually sit next to.

Electronic, hip-hop and pop masters tend to run louder than -14 LUFS. Spotify says it applies negative gain to louder masters during normalized playback. That does not cover every listening path: Spotify says its web player and third-party devices such as TVs or speakers do not use normalization, and Premium listeners can turn it off. What you are choosing is how much of the distance between quiet and loud you are willing to trade for density.

Rock and metal sit across a wide band depending on how dense the production is. Drum transients are the thing over-limiting kills first, and they are usually what the genre is carrying.

Wide-dynamic-range material needs separate treatment. The AES document gives classical music as an example and discusses lower distribution loudness to avoid peak limiting. Its table lists news and talk and drama at -18 LUFS, while mixed-format and sports streams are listed at -17 LUFS.

Common Mastering Mistakes for Streaming

Over-Limiting for Loudness

Pushing a limiter hard to hit -8 LUFS made sense in 2008. It does not make sense against playback normalization. You trade impact for loudness that gets taken back at the playback end, and the song ends up sounding smaller than a cleaner master at the same level.

Ignoring True Peak

A master that peaks at -0.1 dB has nothing left for the encoder or the converter. Set a true peak ceiling of -1 dB or lower and check it with a meter that measures true peak, not sample peak.

Mastering from Lossy Sources

If you master from an MP3 or a low-quality bounce, encoding artifacts compound when the platform re-encodes. Master from the highest quality source you have, and do not upsample to fake resolution you never captured.

One Master for Everything

Some contexts benefit from different masters. A vinyl cut has different requirements than a streaming master, and a DJ promo might want a louder, denser version. Consider whether your release needs more than one.

Testing Your Master

Before finalizing your master for distribution:

1. Check loudness with a proper meter. Verify integrated LUFS and true peak. Do not trust your DAW's default meters unless you know they measure LUFS.

2. Listen at low volume. Does the song still have impact and clarity when played quietly? Over-compressed masters lose definition at low volumes.

3. Compare to reference tracks. Pick commercially released songs in your genre that sound good on streaming platforms. How does your master compare in perceived loudness, punch, and clarity?

4. Test on multiple systems. Phone speakers, earbuds, car stereo, studio monitors. Your master should translate across all of them.

For the full distribution workflow including mastering timing, see the How to Release Your Music: Distribution Guide. For what mastering does in the first place, see the Mastering Guide: What It Does and When You Need It.

Frequently Asked Questions

Should I master differently for Spotify vs. Apple Music?

Spotify publishes a target of -14 LUFS with true peak below -1 dB TP. Apple publishes no LUFS figure, so there is no second number to master toward. One master covers both.

Where does the "-16 LUFS for Apple Music" figure come from?

Not from Apple. Apple's own Apple Digital Masters document describes Sound Check and gives no LUFS figure anywhere in it. The number circulates widely in mastering advice without a primary source behind it.

Will Spotify turn up quiet masters?

Spotify's artist documentation says positive gain may be applied to softer masters toward -14 dB LUFS, while preserving 1 dB of headroom for lossy encodings; its example raises a -20 LUFS track only to -16 LUFS because of true-peak headroom. Premium listeners can switch normalization off, so a quiet master may remain quiet for them.

Does sample rate matter for streaming?

Apple asks for the highest native sample rate you have and 24-bit sources. YouTube recommends 48 kHz and 24-bit for music videos. Deliver your native resolution rather than converting down first.

Read Next

• Mastering Guide: What It Does and When You Need It

• How to Plan a Music Release: Step-by-Step Checklist

• How to Release Your Music: Distribution Guide

• ISRC Codes Explained: What Artists Need to Know

Prepare Your Release:

Orphiq's release planning tools help you build a release timeline that works backward from your release date, so mastering, distribution, and marketing sit on one plan.