Live — upload now

Your mix has problems.
Let's find them.

Carmina AI analyses your mix like a 20-year engineer would. Not generic tips — specific measurements, specific problems, specific fixes.

Upload your mix — it's free See how it works ↓

WAV · MP3 · FLAC · AIFF · Up to 150MB · No account required

app.carmina.audio/mixtus/analysis

my_track_final_v3.wav

44.1kHz · 24bit · 4:32 · Analysed in 48s

6.3

/10 mix score

Analysis — 3 of 11 insights

⚠️

Low-end masking on kick & bass

Your kick and bass occupy the same 60–90Hz band at similar levels. Cut the bass at 80Hz by 3–4dB or high-pass the kick body above 50Hz. Check the result in mono at low volume.

📊

Integrated loudness: −10.2 LUFS

You're 3.2 LUFS above Spotify's −14 LUFS target. This will be turned down on streaming. Either reduce your limiter ceiling or accept the gain reduction.

🎛️

Vocals are slightly narrow in stereo

Lead vocal stereo width is 0.12 (0 = mono, 1 = full wide). Some subtle doubling or a short pre-delay on a stereo aux would open the image without pushing the vocal off-centre.

🔒

Dynamic range & transient control

Detailed analysis and specific fix recommendation available on Pro.

Pro
🔒

High-frequency air and presence

Detailed analysis and specific fix recommendation available on Pro.

Pro
🔒

Mid-range clarity and honk

Detailed analysis and specific fix recommendation available on Pro.

Pro
🔒

Stereo width consistency across stems

Detailed analysis and specific fix recommendation available on Pro.

Pro
🔒

True peak clipping on 3 stems

Detailed analysis and specific fix recommendation available on Pro.

Pro
🔒

Reverb tail build-up in the low-mids

Detailed analysis and specific fix recommendation available on Pro.

Pro
🔒

Drum bus compression artefacts

Detailed analysis and specific fix recommendation available on Pro.

Pro
🔒

Final chain frequency balance

Detailed analysis and specific fix recommendation available on Pro.

Pro
Unlock all 11 insights — from €7.99/month →

← sample output · your results will differ based on your mix

Under the hood

From upload to feedback in under 60 seconds.

01

Upload

Drop your mix — WAV, MP3, FLAC or AIFF, up to 150MB. No account required for a free analysis. File is processed and deleted within the hour.

drag & drop or click

02

GPU stem separation

Meta's Demucs model splits your mix into 6 stems: vocals, drums, bass, guitar, piano, other. Each stem is then measured independently for LUFS, dynamics, frequency content and stereo width.

Demucs · librosa · Modal GPU

03

Engineer feedback

Carmina AI reads the measurements and generates actionable feedback as a 20-year mixing engineer would write it. No filler, no generic advice — band-specific EQ decisions, compression observations, stereo field notes.

Carmina AI · Supabase · self-hosted

What gets measured

11 dimensions. Per stem.

Integrated LUFS

Streaming platform compliance

True peak

Clipping & inter-sample peaks

Dynamic range

Compression & punch

Stereo width

Per-stem imaging

Frequency masking

Kick/bass, mid-range conflicts

Spectral balance

Low/mid/high distribution

Transient response

Attack & release character

Reverb decay

Tail build-up in mids

Mono compatibility

Phase & width collapse

Noise floor

Hiss, hum, background noise

Crest factor

Headroom & limiting artifacts

Mix score

Composite quality index

Pricing

Start free. Go deeper when you need it.

Most tools give you a LUFS number and call it feedback. Mixtus gives you a diagnosis.

Feature

Free

€0

forever
Most popular

Pro

€7.99

/month

Studio

€14.99

/month
Uploads
Analyses per month 3 Unlimited Unlimited
Files per upload 1 1 Up to 10
Max file size 50 MB 150 MB 150 MB
Formats WAV MP3 FLAC AIFF WAV MP3 FLAC AIFF WAV MP3 FLAC AIFF
Account required No Yes Yes
Analysis
Mix type toggle (mix / mastered) ✓ ✓ ✓
Insights per analysis 3 11 11
LUFS + true peak ✓ ✓ ✓
Overall mix score ✓ ✓ ✓
Stereo width & mono check Basic Full Full
Frequency masking analysis — ✓ ✓
Per-stem level balance — ✓ ✓
Vocal intelligibility score — ✓ ✓
Reference track compare — ✓ ✓
Reports & workflow
PDF export — ✓ White-label
Email delivery — ✓ ✓
Analysis history — 30 days Unlimited
Client sharing links — — ✓
API access — — 100 calls/mo
Support
Processing queue Standard Priority Priority
Support — Email Priority email
Try free → Start Pro → Start Studio →

All plans: files deleted within 1 hour · no audio stored · cancel anytime

FAQ

Questions

Does Mixtus replace a human mixing engineer? +
No — and it doesn't try to. Mixtus gives you the measurement-based observations that an experienced engineer would make in the first 5 minutes of listening. What you do with that information is still your creative decision.
Is my audio file kept after analysis? +
No. Files are processed on an isolated GPU container and deleted within one hour of upload. We don't store your music, sample it or use it for training. Privacy policy applies.
What formats does Mixtus accept? +
WAV, MP3, FLAC and AIFF. Up to 150MB per file. For best results, upload the highest quality version of your mix — 24-bit WAV before any streaming conversion.
Is this really built by a real audio engineer? +
Yes. Carmina.audio is a solo project by a French audio professional with 20 years of live and studio experience. The AI feedback model is built on that experience, not scraped from forums.
How is Mixtus different from LANDR or iZotope? +
LANDR auto-masters your track. iZotope helps you process it. Mixtus only analyses and advises — it never touches your audio. The feedback is language-based and specific to your track, not algorithmic presets.
Can I cancel at any time? +
Yes. No contracts, no minimum term. Cancel from your account page and your subscription stops at the end of the billing period.

Ready to hear the truth
about your mix?

Free. No account required. Results in under 60 seconds.

Upload your mix now →