// methodology

Flight Score: the 9-component audio quality score, explained

Last updated: May 9, 2026 · Authored by , Founder of PreDrop and co-host of Track Trip.

Flight Score is a two-tier audio quality score that grades a track on nine measurable components before it leaves the studio. The General Score covers release-readiness — loudness, file format, sample rate, metadata completeness, and copyright clearance. The Technical Score covers the audio engineering itself — true peak, loudness range, crest factor, and stereo correlation. Each component is measured against an objective threshold drawn from streaming-platform specs (Spotify, Apple Music, Beatport) and standard broadcast-engineering practice (EBU R128, ITU-R BS.1770). Scores are calculated client-side from the raw audio, weighted, and combined into a single number from 0 to 100. The result is a repeatable, transparent grade that tells an artist — and a label — whether a track is ready to ship.

What is the Flight Score?

Flight Score answers a question every independent producer asks before sending a demo: is this actually ready? Most artists have no objective way to check. They listen on a two-hundred-dollar pair of headphones, send the track to a label, and then wonder whether the rejection was about the music or the mix. Flight Score makes the mix part objective. It runs the same measurements a mastering engineer would run — LUFS, true peak, loudness range, crest factor, stereo correlation — and compares them to the targets that streaming services and electronic-music labels actually use. If a track scores 90, it's in the shape it needs to be in. If it scores 60, the score will tell you exactly which of the nine components dragged it down and what to do about it.

The methodology is open. Every threshold on this page is the threshold the algorithm uses. There is no proprietary "AI judgment" layer rating your track on vibes — every point awarded or deducted comes from a measured number compared to a published spec.

How is Flight Score calculated?

Flight Score has two tiers. The General Score (5 components, weighted) measures whether a track is structurally ready for distribution: is it loud enough, in a good file format, at a proper sample rate, with complete metadata, and clear of copyright matches. The Technical Score (4 components, weighted) measures the engineering quality of the audio itself: true peak ceiling, dynamic range, transient preservation, and mono compatibility. Each component is graded on a 0–100 scale based on how close the measurement is to the ideal range. Components are then weighted, summed, and averaged. A track's overall Flight Score is the average of its General and Technical scores. If a track is mono, the stereo component is excluded as N/A and the remaining components are re-weighted accordingly.

All measurements run in the artist's browser. The audio file never leaves your device for analysis. Decoding, BPM detection, key detection (with Camelot notation), and the full suite of LUFS / true-peak / LRA / crest-factor / stereo calculations all happen locally via Web Audio API and a Wasm-compiled audio toolkit. The result is stored alongside the track in PreDrop so you, your collaborators, or a label can see the grade without re-running the analysis.

The General Score: 5 components for release-readiness

The General Score is everything that has to be right before the track is even considered for distribution. It is calculated as a weighted average of the five components below. A track that fails this tier won't survive a streaming-service ingestion process even if the audio inside is perfect.

1. Loudness (LUFS) — weight 30

target: −14 LUFS

Integrated loudness measured to the EBU R128 / ITU-R BS.1770 standard. Streaming services normalize playback to approximately −14 LUFS, so a track that hits exactly that target is delivered at full intended loudness. Flight Score awards 100 points at −14 LUFS and deducts 10 points per dB of deviation in either direction. A track at −10 LUFS (significantly hot) scores 60. A track at −18 LUFS (significantly quiet) also scores 60. The penalty is symmetric because both extremes hurt your music — too loud and Spotify turns you down, costing you the punch you fought for; too quiet and you sound smaller than your neighbours in any playlist.

2. Metadata — weight 25

target: 7/7 fields

Seven fields are checked: title, artist, BPM, key, genre, artwork, and ISRC. Title and artist are weighted 2× each because they're non-negotiable — every other field fails gracefully but a track with no title is unusable. Complete metadata isn't just bookkeeping; it's what lets DJs find your track in their library, lets Beatport route your release into the right genre filter, and lets streaming algorithms cluster your work with similar artists for "Discover Weekly"-style recommendations. Score is the percentage of weighted fields filled in.

3. Format — weight 20

target: lossless (WAV / FLAC / AIFF)

Lossless formats (WAV, FLAC, AIFF) score 100. Compressed formats are scored by bitrate: 320 kbps MP3 scores 85, 256 kbps scores 70, 192 kbps scores 50 (with a warning), 128 kbps scores 30, and below that the score collapses. Most electronic-music labels and DJ pools require lossless uploads — pitching a 192 kbps MP3 is the fastest way to end a conversation. PreDrop is built around lossless hosting precisely because this is a non-negotiable in the workflows we built it for.

4. Sample Rate — weight 15

target: 44,100 Hz or higher

44.1 kHz is the CD-quality baseline and the floor for any streaming or distribution platform. Tracks at 44.1 kHz or higher (48 kHz, 96 kHz) score 100. Anything below CD quality earns a proportional penalty and a flag advising re-export from your DAW. This is rare in modern workflows but does occasionally appear in old bounce-down sessions.

5. Copyright Status — weight 10

target: clean or owned

PreDrop runs every Pro-tier upload through a copyright scan (powered by ACRCloud) that checks the audio fingerprint against a database of commercially released tracks. A clean result — no matches — scores 100. A track you've marked as your own (uploaded under your verified artist identity) also scores 100. A "risky" result (potential match worth reviewing) scores 40. A confirmed match scores 0 — distributing it would invite a takedown. This is the only component that can be N/A; if a track hasn't been scanned (lower tiers), the component is excluded and the remaining four are re-weighted.

The Technical Score: 4 components for audio quality

The Technical Score is what an experienced mastering engineer would check before signing off on a release. It measures the engineering decisions baked into the audio: how the limiter was set, how much dynamic range survived, whether the mix translates to mono. Available on Starter and Pro plans.

6. True Peak (dBTP) — weight 30

target: ≤ −1 dBTP

Inter-sample true peak measured via 4× oversampling. Sample peak (the value you see on a normal meter) lies — once your audio is decoded for playback, the actual analog waveform can overshoot the highest sample by up to 3 dB. Streaming services use lossy codecs (AAC, Vorbis) that can introduce additional inter-sample peaks. The industry-standard ceiling is −1 dBTP; Apple Music recommends −2 dBTP for AAC encoding. Flight Score scores 100 at ≤ −2 dBTP, drops linearly to 80 at −1 dBTP, then to 40 at 0 dBTP, then to 0 once you're actively clipping. A track at +0.3 dBTP isn't just "technically loud" — it's going to introduce audible distortion on playback.

7. Loudness Range (LRA) — weight 25

target: 7–20 LU

Loudness Range, measured in LU (loudness units) per EBU R128, is the difference between the quietest and loudest passages of a track. It tells you how much dynamic variation survived mastering. Healthy electronic music sits between 7 and 20 LU — enough variation to give intros, breakdowns, and drops their identity, but not so much that a club system has to ride the gain. Below 4 LU you're squashed flat (most modern over-mastered tracks); above 25 LU you're inconsistent enough that quiet sections may disappear on phone speakers or in a noisy bar. Score drops in both directions outside the healthy range.

8. Crest Factor — weight 20

target: 8–20 dB

Crest factor is the peak-to-RMS ratio in dB. Translation: how much room your transients have to breathe. A kick drum with 12 dB of crest factor punches; the same kick smashed to 4 dB of crest factor goes "thud." Modern loudness-war mastering routinely crushes crest factor below 5 dB — the track gets louder on a meter but loses the impact that made it sound exciting on a quieter reference. Flight Score awards 100 at 8–20 dB and drops proportionally as you compress harder.

9. Stereo Correlation — weight 25 (or N/A for mono)

target: 0.5 to 0.95

Correlation between the left and right channels, on a scale from −1 (out of phase) to +1 (identical, i.e. mono). The healthy zone is 0.5–0.95: stereo enough to feel wide, correlated enough that the mix collapses gracefully to mono. Above 0.95 you're effectively mono — fine for a bass mix, weak for a finished record. Below 0.2 you have a very wide image that may lose its low end on club subs, which sum to mono. Below 0 you're actively out of phase: bass and vocals will partially cancel on any mono playback (clubs, phone speakers, AM radio). For mono tracks the component is excluded entirely.

What's a good Flight Score?

Across the 500+ tracks analyzed on PreDrop, the average Flight Score is 92. Here is how to read your own number:

  • 90–100: Release-ready. Send it to a label, submit it to Beatport, push it to your distributor. Any further changes are taste, not technical deficits.
  • 75–89: Strong. One or two components flagged, usually fixable in a single bounce from your DAW. Expand the breakdown to see which.
  • 60–74: Needs another pass. Almost always a loudness or true-peak issue from aggressive mastering. Pull the limiter back 2 dB and re-run.
  • 40–59: Significant technical issues. Usually a combination of squashed dynamics (low LRA + low crest factor) and clipping. Don't pitch this; it will be rejected before anyone listens to the music.
  • Below 40: Re-master or re-export. This is where bad bounces, wrong sample rates, or accidentally compressed source files surface.

How does Flight Score compare to LANDR or a mastering engineer?

Flight Score is a diagnostic, not a fix. LANDR and CloudBounce will master your track — apply EQ, compression, limiting, and stereo widening based on a learned model — and hand back a finished file. Flight Score does the opposite: it tells you what is and isn't working with the master you already made, so you can decide whether to re-master, send it to a human engineer, or ship it. The two workflows are complementary. Run a track through LANDR, then through Flight Score, and you'll see exactly which components the automated master nailed and which need a human touch.

Compared to a human mastering engineer's technical sheet, Flight Score uses the same measurements (LUFS, dBTP, LRA, crest factor, stereo correlation) and the same broadcast-engineering standards (EBU R128, ITU-R BS.1770). The difference is volume: an engineer charges $50–$300 per track and turns a master around in 24–72 hours. Flight Score runs in your browser in under 30 seconds. It won't replace the engineer's ear — but it will keep you from sending them a track that's clearly broken before they even hear it.

Why does Flight Score matter when pitching to record labels?

Track Trip — the trance podcast I co-host — receives demos every week from independent producers hoping to get featured. The most common reason a demo gets passed over isn't the music. It's a technical problem the artist didn't know about: a bounce that's 6 dB too quiet next to the track before it, a clipping master that distorts on monitors, a stereo image so wide the bass disappears in the club PA. By the time a label hears any of that, the demo is already in the "maybe later" folder. Flight Score exists so an independent artist can catch those problems before the demo goes out — turning a rejection-on-technicals back into an honest conversation about whether the music is right for the label.

When you pitch a track via PreDrop, the recipient label sees the Flight Score next to the audio. A 92 tells them you've done the technical work. A 64 tells them they probably won't need to write back about the master.

How do I improve my Flight Score?

Flight Score won't just give you a number — every component that scored below 80 comes with a specific piece of actionable advice tied to the measurement. A few of the most common patterns:

  • Loudness flagged as too hot (> −12 LUFS): pull your final limiter threshold up 2–3 dB. You'll lose perceived loudness in your DAW but gain it back through Spotify's normalization, plus regain the punch you lost.
  • True peak flagged (> −1 dBTP): set your limiter ceiling to −1 dBTP (or −2 dBTP if you're distributing AAC-encoded). True-peak limiting is built into FabFilter Pro-L 2, Ozone Maximizer, and Pro Tools' native limiter — turn it on.
  • Crest factor low (< 5 dB): you're over-limiting. Reduce your limiter's gain reduction, or — better — fix the problem upstream by easing off the bus compressor on your drum buss.
  • Stereo correlation negative: phase issues. Check your hi-hat, pad, and reverb stereo plugins. A vintage stereo widener applied aggressively to a low-frequency element is the most common culprit.
  • Metadata incomplete: fill in BPM, key, genre, and artwork before you ship. PreDrop detects BPM and key automatically; you only have to type the title, artist, and genre.

What's next

Upload a track to PreDrop and you'll see your General Score immediately. The Technical Score, Spectral Balance, and Loudness-over-time breakdown are available on Starter and Pro plans. The methodology on this page is the methodology in the code — no hidden adjustments, no proprietary "quality AI." Every threshold, weight, and formula is the one running on your audio.