Module 11 · Mastering Track
Mastering for Streaming & Delivery
Streaming services normalize loudness. The "loudness wars" of the 2000s are over — turning your master up no longer makes it stand out. This module shows you the LUFS targets per platform, how to use a true-peak limiter, and how to deliver files that translate everywhere from Spotify to a CD to a YouTube upload.
For 30 years, "louder is better" was the rule. Engineers competed to crush masters into brickwall limiters, sacrificing dynamics for perceived volume so their record would stand out next to the previous track on the radio. That era is over.
Today, every major streaming service measures the loudness of every uploaded track and turns it down to a standard target before playing it back. Spotify normalizes to -14 LUFS. Apple Music to -16 LUFS. YouTube to -14 LUFS. If you upload a track that's louder than the platform's target, the platform turns it down. If you upload a track that's quieter, the platform leaves it alone (or sometimes pulls it up a hair). The result: a heavily-limited "-7 LUFS" master and a dynamic "-12 LUFS" master will play back at the same loudness on Spotify — but the dynamic master will sound better, because it has room to breathe.
Four ideas first about loudness in the streaming era, then the platform target table, then how to set your limiter and deliver files that work everywhere.
First, the words
Four concepts about loudness, peaks, and platform behavior that change how you master in 2026.
Concept 1
LUFS measures perceived loudness
LUFS (Loudness Units Full Scale) is a meter that estimates how loud a track sounds to a human, not just how high its peaks are. Streaming services use LUFS to set normalized playback level.
Think of dBFS as a track's height (how tall the peak spike is) and LUFS as its weight (how heavy the whole track feels). Two tracks can be the same height and weigh very differently.
A traditional peak meter shows the maximum sample value — useful for checking digital clipping but useless for comparing perceived volume. A LUFS meter weights frequency response (matching how human ears respond) and integrates over time, giving a single number that correlates with how loud the track seems. Three sub-readings matter: Integrated LUFS (average over the whole song — this is what platforms target), Short-Term LUFS (3-second sliding average — useful for checking the loudest section), and Momentary LUFS (400ms window — useful for catching transient spikes). The integrated reading is your delivery target.
Concept 2
True peak is the only ceiling that matters
A "0 dBFS" sample peak still distorts on real playback systems. True-peak (dBTP) measurement catches inter-sample peaks. Set your limiter ceiling to -1 dBTP, not 0 dBFS.
Think of sample peaks as marker dots on a sketch and true peaks as the smooth curve you drew. The curve always goes higher than any single dot.
Digital audio is a series of sample points. When that signal gets reconstructed by a DAC (the chip that converts your file to actual sound) or compressed to MP3/AAC for streaming, the smooth analog curve between samples can rise above any individual sample value — sometimes by 1–3 dB. So a master with samples at 0 dBFS can clip on Spotify, in your car, on AirPods. True-peak limiters use oversampling to catch these inter-sample peaks before they happen. The industry standard ceiling is -1 dBTP — leaves enough headroom for any platform's lossy encoding without losing measurable loudness. Some engineers go to -2 dBTP for extra safety on heavily lossy formats. Never deliver a master at 0 dBFS unless you want it to crackle on streaming.
Concept 3
Streaming services turn loud tracks down
Every major streaming service normalizes loudness. Pushing your master past -10 LUFS doesn't make it louder on Spotify — it just makes it more squashed at the same playback level.
Think of it like a restaurant where everyone's drinks are topped up to the same line — pouring yours overflowing doesn't get you more, it just gets it scooped back out and you lose the foam.
Spotify, Apple Music, YouTube, Tidal, Amazon Music, and SoundCloud all measure your track's integrated LUFS and adjust playback to a platform target. The implications are huge: the louder you master, the more dynamics you sacrifice, and you get nothing back at the streaming end because the platform turns you down anyway. Master to your platform's target loudness — or just slightly under — and you keep the dynamics that make the song feel exciting. Master 6 dB hotter than the target and the platform turns you down 6 dB, leaving you with the same loudness but flatter, more fatiguing, less dynamic music. The smart move in 2026 is to master quieter than you used to and trust the platform.
Concept 4
Reference, then match
Pick 2–3 commercial tracks in your genre. Loudness-match them to your master using a LUFS meter. Now A/B against them in real-time to check tonal balance, low-end, and stereo image.
Think of a tuning fork — it doesn't tell you what note to play, it tells you whether you're in tune with the orchestra.
Without references, you're mastering in a vacuum. Reference matching is the single most powerful sanity check available — comparable in impact to having a $40,000 mastering room. Pick references in your genre that are well-mastered (this is half the skill — choose wisely; a Lana Del Rey master and a Skrillex master demand different things). Loudness-match them to your master using a meter so loudness alone doesn't bias the comparison. Then A/B: does your low-end land in the same place? Is your top-end the same brightness? Is your stereo image as wide? Listen for tonal differences, not loudness differences. Free tools that handle this: Youlean Loudness Meter (free), MeterPlugs Perception (free), iZotope Tonal Balance Control (paid). Reference matching is the closest thing to a guaranteed mastering improvement.
Below is the platform target table. These are the integrated LUFS targets and the maximum true-peak ceilings each major service uses. Master to these numbers and your track plays at full intended loudness everywhere — no surprises.
The "one master fits all" question
Most independent artists make a single master and upload it everywhere. This is fine. Master to -14 LUFS, -1 dBTP and you're aligned with Spotify, YouTube, Tidal, Amazon (within 1 dB), SoundCloud, and you'll be 2 dB hotter than Apple's target which is barely audible. If you're delivering for vinyl, club, or Bandcamp lossless, make a separate louder "non-streaming master" — but for the digital streaming ecosystem, one master at -14 LUFS handles it.
A well-set true-peak limiter is mostly invisible — it works on the loudest 1–3 dB of peaks, not on the body of the signal. If you can hear the limiter pumping, you're pushing it too hard.
Try this · 45 minutes
Master to -14 LUFS for streaming
Open the prepped mix bounce from Module 10's exercise. Open a new mastering session, drop the WAV onto a stereo track, and walk through this exact sequence.
- Install a free LUFS meter if you don't have one. Best free option: Youlean Loudness Meter 2. Drop it on the master output (post-fader, end of chain).
- Reference 2–3 commercial tracks in your genre. Drag them into the same project on a separate track. Use a gain plugin to match their LUFS reading to your unmastered mix's LUFS reading. Now you can A/B without loudness biasing the comparison.
- Build the mastering chain from Module 10's visual: subtle EQ, gentle bus comp, optional multiband, saturation, optional stereo image, true-peak limiter. Use stock plugins.
- Set the limiter ceiling to -1 dBTP. In Logic's Adaptive Limiter, this is the "Out Ceiling" knob. In Ableton's Limiter, switch on "True Peak" mode and set ceiling. In FabFilter Pro-L 2, oversampling on, true-peak detection on, output -1 dBTP.
- Adjust the limiter input gain until the integrated LUFS reading on Youlean lands at -14. Watch the gain reduction meter — it should be doing 1–3 dB GR on the loudest sections, not 6+ dB. If it's pulling more than 3 dB, your mix is too dynamic for -14 LUFS — go back and add some bus compression upstream.
- A/B with your references. Loudness-match the references to -14 LUFS too. Listen for tonal differences. Is your low-end as full? Top-end as bright? Adjust the EQ stage on your master ±1 dB at a time until the references and your master sit in the same tonal neighborhood.
- Bounce the master. 24-bit WAV at session sample rate. Apply dither if you'll deliver as 16-bit (CD or some streaming uploads). Most modern streaming uploads accept 24-bit — no dither needed.
- Re-measure the bounce. Drop the bounced WAV into a new project. Run Youlean across it. Confirm integrated LUFS = -14, true peak ≤ -1 dBTP. If anything's off, fix it.
- Listen on three different systems. Studio monitors → headphones → phone speaker. Same impression on all three? Ship it. Big tonal difference? The mix might still need work — back to Module 8.
This is the master you'd upload to streaming platforms. Hold onto the project file — Module 13's certification capstone will ask for a master like this.
The end of the loudness wars
From the late 1980s through the late 2000s, mastering engineers competed to make records louder. Brickwall limiters got more aggressive each year. By 2008, average commercial pop masters were hitting -7 to -8 LUFS — squashed flat with almost no dynamic range. Records sounded loud and exciting on the radio for the first 30 seconds and fatiguing after three minutes. This was the loudness war.
The end came from the streaming services. In 2014, Spotify quietly enabled loudness normalization by default. Within five years, every major service had followed: Apple Music, YouTube, Tidal, Amazon, SoundCloud. The competitive advantage of being loud disappeared overnight — turning your master up just made the platform turn it down.
What replaced it: a slow return to dynamics. Modern pop masters trend toward -10 to -12 LUFS — still loud, but with breath. Modern indie/folk/jazz masters trend toward -14 to -18 LUFS — fully dynamic, taking advantage of the platform's normalization. The 2026 mastering ethos is "master to your music, not to a meter" — let the song decide how loud it wants to be, and trust the platform to handle the rest.
For independent artists this is great news. You can now release a dynamic, well-mixed record without worrying it'll sound "weak" next to commercial releases — when both play through Spotify, both land at -14 LUFS, and the more dynamic record actually sounds better because it has more contrast and less listener fatigue.
File formats & delivery
Master file (what you bounce)
- Format: WAV or AIFF (uncompressed PCM)
- Bit depth: 24-bit (all major streaming platforms accept this)
- Sample rate: 44.1 kHz (CD-aligned) or 48 kHz (video-aligned). Match your session.
- Dither: Apply only if downsampling bit depth (24→16). Most modern streaming = no dither needed.
- File naming: "Artist - Songname (Master) v3 2026-04-30.wav" — version + date so you don't lose the right one.
Distribution upload formats
- Spotify, Apple Music, Tidal, Amazon, YouTube Music: upload the 24-bit WAV via your distributor (DistroKid, TuneCore, CD Baby, AWAL, etc.). Distributors transcode to the platform's preferred format.
- SoundCloud: WAV or AIFF preferred. Will transcode to AAC for streaming, but uploading lossless preserves more quality.
- Bandcamp: upload WAV or FLAC. Bandcamp keeps the lossless original for buyers who want it.
- YouTube (video upload): WAV inside an MP4/MOV container. Don't upload bare audio without a video — YouTube will strip metadata.
- CD master: 16-bit, 44.1 kHz WAV. Apply dither when downsampling. Burn from a DDP image for replication houses; standard WAV files work for short-run duplication.
- Vinyl master: 24-bit, 44.1 or 48 kHz. Slightly different mastering — wider dynamics, mono-below-150 Hz strict, sibilance tamed. Specialized — most engineers send to a dedicated vinyl mastering specialist.
What about MP3?
Don't deliver MP3 to distributors or streaming services. They want the highest-quality lossless source so their own encoders can transcode without compounded loss. MP3 is fine for personal sharing, demos, or your own archive — not for delivery. If you absolutely have to deliver MP3 (some sync libraries still ask), 320 kbps CBR is the standard. Don't go below 256 kbps for anything serious.
Reference matching tools
- Youlean Loudness Meter 2 — free, the industry standard for LUFS measurement. Use this to check your integrated LUFS and true peak.
- iZotope Tonal Balance Control — paid, shows your master's spectral curve against a genre reference. Excellent for catching mid-range buildups and missing top-end.
- MeterPlugs Perception — free, automatically loudness-matches a reference track to your mix for honest A/B.
- Sonarworks SoundID Reference — corrects your monitors and headphones to a flat response. Different category but related: trust your room first, then everything else gets easier.
"The loudness wars are over. Spotify won. The good news is, dynamics are back."— Ian Shepherd, Production Advice