Mix analysis is the process of evaluating an audio mix across technical dimensions like loudness, frequency balance, stereo width, and dynamic range to identify problems before mastering. It tells you objectively what your ears might miss after long sessions — like a reference engineer giving you notes. MixDiagnose automates this with AI so you get a full diagnostic report in seconds. You can try it free on the home page.
MixDiagnose runs your uploaded audio through DSP algorithms that measure integrated LUFS, true-peak dBTP, frequency-band balance, stereo correlation, and crest factor, then compares each metric against universal mix targets. The AI grades severity (green / yellow / red) and generates plain-English recommendations. It also produces a spectrogram and frequency plot so you can see the problems visually. No genre selection needed — the thresholds are flat and universal.
MixDiagnose checks loudness (integrated LUFS and true peaks), low-mid frequency balance (detecting mud and muddiness), high-end clarity, stereo width and phase correlation, dynamic range / crest factor, and overall tonal balance. Each issue is graded by severity so you know what to fix first. The report is the same complete diagnostic on every tier — free users get it too.
Yes — when the engine detects a measurable problem (e.g. bass 5 dB too hot or LUFS too low), it can apply a corrective EQ filter to your uploaded audio and give you a downloadable processed file. This is called an "auto-fix." Free tier gets 1 auto-fix per month; Producer and Pro plans get unlimited auto-fixes. See pricing for plan details.
A typical track is analyzed in under 30 seconds depending on file length and format. Non-WAV files (MP3, FLAC, M4A, AAC) are converted via ffmpeg first, which adds a few seconds. Pro plan subscribers get priority queue processing for faster turnaround during peak load. You'll see results as soon as the spectrogram and diagnostic report finish rendering.
LUFS (Loudness Units Full Scale) is a perceptual loudness standard that measures how loud audio sounds to the human ear, not just electrical level. It's the unit used by Spotify, YouTube, Apple Music, and broadcast standards (EBU R128) to normalize playback. LUFS are more accurate than peak dB because they account for sustained energy and frequency weighting. Check your track's value with our free LUFS checker.
Spotify normalizes to −14 LUFS integrated, so masters near that level avoid being turned down by the platform. Going louder (e.g. −8 LUFS) doesn't make your track sound louder on Spotify — it gets attenuated back down and you lose dynamic range and punch. Aim for −14 LUFS with peaks below −1 dBTP to keep dynamics intact. Measure yours with the LUFS checker before uploading.
YouTube normalizes to −14 LUFS integrated, the same target as Spotify, so tracks around that level avoid being turned down. Unlike Spotify, YouTube's replay gain is slightly less aggressive, but over-compressing still loses punch once it's attenuated. Keep true peaks below −1 dBTP and let the platform handle loudness normalization. Verify your integrated value with the free LUFS checker.
dB (decibels) measures electrical signal level — usually peak dBFS, which only captures instantaneous maximums and ignores how loud something actually sounds. LUFS measures perceived loudness over time using K-weighting, so two tracks at the same peak dB can have very different LUFS values. Streaming platforms normalize by LUFS, not peak dB, which is why LUFS is the number that matters for distribution. Run both checks in a MixDiagnose analysis report.
Yes — if your master is louder than −14 LUFS integrated, Spotify attenuates it down to that target during playback normalization. That's why hyper-compressed masters actually lose punch on the platform: the gain reduction eats your dynamics. If you master to −14 LUFS with healthy dynamic range, Spotify leaves it alone and your track keeps its full energy. Check your integrated LUFS with the LUFS checker before release.
Mud usually lives in the 200–500 Hz range, where bass, guitars, and low-mid keys pile up and mask clarity. Fix it by high-passing non-bass elements around 100–150 Hz and cutting 2–4 dB in the 250–400 Hz zone on instruments that don't need body there. Make cuts on individual tracks, not the master, so you don't hollow out the overall mix. MixDiagnose flags excessive low-mid energy in its frequency balance report — run an analysis to see exactly where your mud is.
Phase issues happen when two signals covering the same source are slightly time-misaligned — most commonly multi-mic drum setups, DI plus amp signals, or stereo widening plugins that delay one side. The result is comb filtering: frequencies cancel and the sound gets thin or hollow. MixDiagnose measures stereo correlation and flags phase problems in the spectral analysis; a correlation reading near zero or negative indicates a phase mismatch worth fixing.
A mix is ready for mastering when it's balanced across the frequency spectrum, has no obvious mud or harshness, leaves headroom (peaks below −6 dBFS), and still has dynamic range the mastering engineer can work with. If your LUFS is already pinned at −8 with zero dynamics, mastering has nothing to shape. Run a MixDiagnose analysis — if every category reads green or yellow, you're in good shape to send it off.
Dynamic range is the difference between the loudest and quietest parts of your track, usually expressed as crest factor (peak minus RMS) or as the gap between true peak and integrated LUFS. It matters because streaming platforms normalize loudness, so a track with more dynamics sounds punchier and more alive after normalization than a hyper-compressed one. Crush it and you lose transient impact; keep it healthy and your mix breathes. Measure yours with the dynamic range calculator.
Keep low frequencies (below ~120 Hz) mono — wide bass causes phase cancellation on playback systems with summing issues and makes the low end weak. Mids and highs can spread wide for space and width, but maintain healthy stereo correlation (above 0) so the mix doesn't collapse in mono. A good target is a correlated, coherent stereo field with controlled width, not a maximally wide image. MixDiagnose reports stereo width and correlation in the spectral analyzer output.
For vocals start around 80–100 Hz, guitars 100–150 Hz, synths and keys 120–180 Hz depending on how much low body you want, and cymbals/percussion 200–300 Hz or higher. The goal is removing low-frequency rumble and unnecessary energy that clutters the bass region without thinning the instrument's useful body. Solo each track with the filter engaged and sweep until you hear it thin out, then back off slightly. MixDiagnose's frequency balance report shows whether your low end is overloaded and where cuts would help — run an analysis to find out.
A/B reference matching means level-matching your mix to a professionally released track in the same genre and comparing tonal balance, stereo width, and loudness side by side. Always match loudness first (within ±0.5 dB) so you're comparing EQ and dynamics, not volume. Pro plan subscribers can use MixDiagnose's reference track matching feature to compare your mix against a reference automatically. See Pro plan features for details.
Yes — the Free tier gives you 3 analyses and 1 auto-fix per month at no cost, with no credit card required. You get the full diagnostic report, spectrogram, and frequency plot on free analyses. Upgrade to Producer or Pro when you need more volume or features like batch processing and reference matching. See pricing for the full plan comparison.
No. MixDiagnose runs entirely in your browser — there's nothing to download or install. Upload your audio file from any device and the analysis runs in the cloud, returning a full report in seconds. It works on desktop, tablet, and mobile browsers alike. Just visit mixdiagnose.com and start analyzing.
You can upload WAV, MP3, FLAC, M4A, and AAC files. Non-WAV formats are automatically converted via ffmpeg on upload so the analysis runs on a consistent waveform. Export format depends on your tier: Free exports 16-bit WAV, Producer adds HD WAV / MP3 / FLAC at 24-bit, and Pro unlocks 32-bit float export. See pricing for the format matrix per plan.
Yes. Uploaded files are processed on upload and are not stored permanently — they're identified by UUID, not tied to personal identity, and the free tier doesn't even require an account. Processed audio for auto-fixes is available for download and then cleared. Read the full Privacy Policy for specifics on data handling and retention.
Yes — MixDiagnose is fully web-based and works in mobile browsers on iOS and Android. The upload, analysis, and report views are all responsive, so you can check a mix from your phone between sessions. No app install needed. Just open mixdiagnose.com in your mobile browser.
Free gives you 3 analyses and 1 auto-fix per month with 16-bit WAV export and the complete diagnostic report. Pro ($49/mo annually) unlocks unlimited analyses and auto-fixes, 32-bit float export, reference track matching, priority queue processing, API access, and stem separation (coming soon). The middle Producer tier ($19/mo annually) adds 50 analyses, unlimited auto-fixes, batch processing, and genre profiles. Compare all three on the pricing page.