SoundPure transforms AI-generated music at the acoustic level — reshaping spectral fingerprints, adding organic texture and natural imperfection — so your tracks carry the same depth and character as a real studio recording.
60%+
Human-score improvement
6
Specialized tools
100%
Browser-based, no installs
Industry
Detection standards verified
Acoustic Intelligence
Built for how modern platforms listen.
Today's streaming platforms and DSPs use sophisticated audio analysis tools — like ACRCloud — to evaluate the acoustic characteristics of every uploaded track. They examine spectral fingerprints, envelope consistency, harmonic patterns and dozens of other acoustic layers.
SoundPure's multi-layer DSP pipeline works on each of these dimensions, consistently achieving a 60%+ improvement in human-likeness score — giving your music the acoustic depth and natural imperfection that sets it apart.
Human-likeness score · Before vs After SoundPure
Before
After
Spectral
—
—
Harmonic
—
—
Envelope
—
—
Overall score
—
—
■ Original AI track ·
■ After SoundPure processing ·
Higher = more human-like acoustic profile
🛡
Anti-Detection DSP
Targets 12 ACRCloud classifiers simultaneously — spectral fingerprint, envelope, groove lock, CLEWS layer and more.
🫁
Organic Layer
Synthesizes real acoustic elements — breaths, room tone, vinyl crackle, finger noise — layered over your track at sub-perceptible levels.
🌿
Structural Humanizer
Breaks mathematical repetition in timing, pitch and harmonic patterns using irrational frequency LFOs that mimic human imperfection.
✂️
Smart Finishing
Automatic tail trim, fade-in/out and LUFS normalization. Every export is ready to upload immediately.
AI Mastering · Professional Reference Matching
Studio-grade mastering, in seconds.
Upload your track and a reference song. SoundPure's AI analyzes the reference's loudness, EQ curve, stereo width and dynamic range — then applies the same sonic signature to your mix with surgical precision, delivering results that rival professional mastering studios.
Reference-matched loudness targeting (LUFS)
Spectral EQ curve transfer from any reference
Stereo image & dynamic range matching
Lossless 16-bit WAV output, no quality degradation
No plugins, no DAW — works entirely in your browser
Input track
↓ AI Reference Analysis ↓
Mastered output
−14
LUFS target
16-bit
WAV export
< 90s
Processing
How it works
Four steps to a track that sounds — and passes — as human
1
Upload
Drop any MP3, WAV, FLAC or OGG file directly in the browser. No installs, no limits on track length.
2
AI Process
The anti-detection pipeline rewrites the spectral fingerprint, adds organic acoustic layers and breaks every mathematical pattern ACRCloud looks for.
3
AI Mastering
Upload a reference track. Our AI matches its loudness, EQ curve and stereo width — giving your music a professional, label-ready master in seconds.
4
Distribute
Upload confidently to Spotify, TikTok, YouTube and every DSP — your track now reads as human to every detector and sounds like a pro release.
🔬
Transparent by design
Don't take our word for it — test it yourself.
There are several free tools online that use industry-standard audio analysis — the same technology platforms rely on — to evaluate how human a track sounds. We encourage you to run your original file through one of them, then process it with SoundPure and run it again. The difference in score speaks for itself.
A few you can try: TheGhostProduction, AHA Music or any similar free service. No account needed, no strings attached.
Start free. No credit card required.
Every new account gets 2 free uses on each tool. See what SoundPure can do for your music before committing to a plan.
⚡ AI Process
Humanize your track in one click
Upload any track and the default pipeline takes care of everything: anti-detection, organic layering, structural humanization and final finishing. The result sounds natural and passes industry-standard acoustic analysis tools.
Want to go deeper? Each parameter below is independently adjustable — feel free to experiment with the sliders and optional add-ons to find the sweet spot for your track.
♪
Upload your track
MP3 · WAV · FLAC · OGG · M4A — drag here or click per scegliere
FILE
—
—
Original waveform
Mastering & Cleanup
Noise Reduction
Removes hiss, hum and background noise typical of AI tracks.
55%
De-Harshness EQ
Attenuates the "plastic" 2–8 kHz frequencies typical of Suno/Udio output.
60%
Low-End Punch
Boosts low end and rebuilds the sub presence of professional mastering.
45%
Stereo Width
Expands the stereo field for a natural spatial perception.
50%
Compressor
Evens out dynamics while preserving natural punch and breath.
50%
Limiter / LUFS
Brings to target level without clipping (–14 LUFS = streaming standard).
-14 L
Anti-Detection (AI Fingerprint Masking)
Spectral Randomizer
Fakeprint Destruction (Afchar et al., ISMIR 2025 — Best Paper).
3 layers in cascade: A) Notch IIR Q=180 on 7 peak frequencies (sr/4…sr/32) · B) Spectral Envelope Randomization on 6 bands 1–8 kHz with irrational LFOs — invalidates the spectral average used to extract the fakeprint · C) Stochastic Frame Perturbation on 256ms windows → long-term average collapses to noise.
65%
Micro-Timing Humanizer
Slowly modulates the envelope (0.8–3.7 Hz) simulating human micro-timing variations. Targets the "Rhythmic Quantization" layer without audio artifacts.
70%
Pitch Micro-Drift
Adds a very slow intonation micro-drift (±8 cents over 60s) — real instruments always have a slight temperature-induced drift.
55%
Harmonic Injection
Injects low-amplitude even and odd harmonics simulating natural analog saturation — absent in pure AI signals.
40%
Noise Floor Naturalize
Organic pink noise at –88 dBFS + Texture Loop Breaker: breaks the mathematical repetition of the noise floor (CLEWS layer) with irrational periodic gain.
50%
Transient Reshape
Targets the "Envelope Consistency (Groove)" layer: introduces LFOs at 3 irrational frequencies that break the mathematical lock between stems.
45%
🫁 Organic Layer
The most powerful layer against the Macro Structure classifier.
Synthesizes 4 real acoustic elements layered over the track via DSP:
· Human breaths (0.22–0.30 Hz, envelope asimmetrica fisiologica)
· Room tone (early reflections at 17/31/59ms + pink noise tail)
· Vinyl/tape crackle (Poisson distribution, anti-CLEWS)
· Finger/body noise (irregular transients 3–18ms)
60%
🎵 Pitch Shift
Real transposition of ±0–3 semitones via double resampling. Shifts all frequencies, destroying ACRCloud's spectral fingerprint. Completamente inaudible sotto 1 semitono.
40%
⏱ Time Stretch
Time stretching 0–4% without changing pitch. Degrades SpectTTTra which uses spectro-temporal tokenization: tokens no longer match known Suno patterns.
35%
🗣 Speech Masking
BMAT/UPF 2026 research: adding speech at –20 dB degrades CNN detectors from 99% to <60% F1. Synthesizes vocal formants (F1/F2/F3) at subliminal level (~–22 dBFS).
45%
⚙️ Optional add-ons
Default pipeline: Anti-Detection → Organic → Structure → Finishing
Enable the options below to add extra processing steps.
Processing happens entirely in your browser — no files are uploaded online
Anti-detection: fakeprint destruction + timing humanizer + pitch drift + harmonic injection + noise floor + transient reshape + organic layer + human samples
Analysis
EQ / Mastering
Spectral mask
Humanize
Limiting
Export
✓ Complete
Mastering completato
—
LUFS
—
Dynamic Range
—
True Peak
—
Gain applicato
Original
0:00 / 0:00
Processed
0:00 / 0:00
Seleziona formato di export
WAV/WAV24: encoding nativo puro · MP3: via lamejs (320kbps) · OGG: via MediaRecorder
Nota: il FLAC puro richiede WebAssembly — WAV 24-bit è equivalente lossless
🎵
Upload track to analyze
MP3 · WAV · FLAC · OGG · M4A — drag here or click
FILE
—
—
Waveform
Analyzing...
🥁 BPM
—
—
🎹 Key
—
—
—
Additional metrics
Chromagram — energy per note
CC# · D · D# · E · F · F# · G · G# · A · A# · B
✨ AI Mastering
Upload a target track and a reference track.
The AI will analyze the reference and apply the same characteristics to your track. Tip: you can also use the original version of your track as reference — useful after AI Process to recover some of the original tonal balance and dynamics.
Server: — checking...
Upload a watermark (a short audio clip — jingle, voice tag, tone, 1–15s recommended) and a track to protect. The watermark will be mixed into the track at the position and volume you choose, then exported as a new file.
Start after
Delay in seconds before the first watermark plays. Use this to keep the intro clean.
Repeat every
How often the watermark repeats (in seconds). Disable to play it only once.
Watermark volume
Level of the watermark relative to the main track. –6 dB is a good default — audible but not intrusive.
Duck main track
Lowers the volume of the music while the watermark is playing, so the watermark is always clearly heard.
Watermark fade
Smooth fade in/out applied to the watermark to avoid clicks or abrupt transitions.
Crop output
Export only a portion of the track — useful for sending a short preview. Set Duration to 0 to export the full track.
Track fade in/out
Applies a smooth fade at the beginning and/or end of the exported track.
DistroKid/TuneCore: 0% · CD Baby: 9% · Major label: 50–85%
Your share (%)
If you co-own the rights, enter your percentage
Gross Earnings / month
$0.00
before fees
Your Net / month
$0.00
after fees and split
Net / year
$0.00
annual projection
Stream milestones → estimated net earnings
📖 How Spotify payments work (2026)
Spotify does not publish a fixed per-stream rate. It operates a pro-rata royalty pool: collecting revenue from subscriptions and advertising, retaining its share (~30%), and distributing the rest to rights holders proportional to their share of total streams on the platform.
In practice, the rate settles between $0.003 and $0.005 per stream, with an average of around $0.004.
🌍 Listener geography
USA, UK, Germany, Australia and Scandinavia generate the highest royalties. India, Brazil and Indonesia significantly less.
💎 Subscription tier
Premium streams pay 20–40% more than the free (ad-supported) tier. The Superfan tier (rolling out 2025–26) may add direct payments.
📦 Distributor
DistroKid, TuneCore, Amuse: 0% fee. CD Baby: 9%. Major label: typically 50–85% retained, leaving the artist 15–25 cents per gross dollar.
🎯 1,000 stream threshold
Dal novembre 2023 Spotify richiede almeno 1.000 stream annui prima chand a traccia entri nel royalty pool. Sotto this soglia gli stream non generano pagamenti.
📊 Reference rates by platform (2026)
Platform
Min rate
Max rate
Est. average
💚 Spotify
$0.003
$0.005
$0.004
🍎 Apple Music
$0.007
$0.010
$0.008
🎵 Tidal
$0.010
$0.014
$0.012
▶️ YouTube Music
$0.001
$0.003
$0.002
🎧 Amazon Music
$0.004
$0.007
$0.005
Rates vary significantly based on listener country, subscription tier and distribution agreement. The values above are estimates based on publicly available data and industry reports updated to 2026.
⚠️ Disclaimer
This calculator provides indicative estimates for informational and educational purposes only. Actual earnings depend on many factors including: listener country, subscription tier (Premium vs Free), playlist placement, specific distributor agreement, and Spotify royalty pool variations.
The rates shown reflect estimated industry averages for 2026 and may vary significantly. Spotify does not publish official fixed per-stream rates. For precise data on your earnings, consult your distributor or Spotify for Artists.
We accept no liability for financial decisions made based on these estimates.
Simple, transparent pricing
Start for free. Upgrade when you need more.
MonthlyAnnual Save 2 months
Base
€6.99/mo
10 credits/mo — AI Process
10 credits/mo — AI Mastering
10 credits/mo — Watermark
Unlimited BPM & Key analysis
Unlimited Royalty Calculator
WAV, MP3 & OGG export
Browser-based — no installs
Most popular
Pro
€12.99/mo
50 credits/mo — AI Process
50 credits/mo — AI Mastering
50 credits/mo — Watermark
Unlimited BPM & Key analysis
Unlimited Royalty Calculator
WAV, MP3 & OGG export
Priority processing
Browser-based — no installs
Infinite
€19.99/mo
Unlimited — AI Process
Unlimited — AI Mastering
Unlimited — Watermark
Unlimited BPM & Key analysis
Unlimited Royalty Calculator
WAV, MP3 & OGG export
Priority processing
Browser-based — no installs
All plans include a free trial with 2 uses per tool · No credit card required to sign up
Prices in EUR · Cancel anytime · Secure payments via Stripe
Account
Personal details
Update your name, phone and preferences.
Marketing notifications
Receive updates, tips and promotions via email
Subscription
Manage your current plan and billing.
Plan—
Billing—
Status—
Amount—
Next billing date—
⚠️ If you cancel, your plan stays active until the end of the current billing period. After that, no more credits will be added and your plan will revert to Free.
Contact us
Send us a message and we'll get back to you as soon as possible.