How to tell what is actually wrong with your mix
A mix that feels unfinished can have a specific measurable imbalance: crowded low mids, missing low end, restricted dynamics, or unstable width. Mix Check reads your bounce against a rough, hand-drawn guide for the style you pick and names one band to investigate first. It runs entirely in your browser, so the file never leaves your machine.
The file is decoded and analyzed locally; it is not uploaded.
01 · Understand
How it works
The tool decodes your file in the browser and measures octave-band energy, spectral centroid (overall brightness), crest factor (how much dynamic range survived), and stereo width and correlation.
Those numbers are read against a style guide for the genre you pick. Each guide is a rough, hand-drawn sketch of how the style is often described, not measured from reference records, so the tool gives no score or grade. Well-mixed records can sit far from it.
The output is a per-band chart of where your mix sits relative to the guide and one named band to listen to first. A band you can name is a question you can test, which is the point.
Nothing is uploaded. The analysis is Web Audio running locally, which is why it is instant and why your unreleased track stays on your machine.
Method note
Web Audio measures octave-band balance, crest factor, width, and correlation, then shows where those signals sit against a rough, hand-drawn guide for the selected style. Brightness is measured and shown but not compared, because the guides' brightness windows are not measured.
02 · Run it
How to use it
- 1
Bounce a full-length mix, not a loop
Spectral averages need representative material. An eight-bar loop or a single stem will say more about that excerpt than the mix. Export the full song in a format your browser can decode.
- 2
Pick the style you are actually aiming at
The guide changes with the style, so pick the one you are making. Remember it is a hand-drawn sketch: a gap from it is a place to listen, not a verdict on your crest factor or your low end.
- 3
Read the one named band first
Go to the band the tool says sits furthest from the guide and ask which instrument owns that band in your arrangement, and whether the difference is one you meant.
- 4
Fix at the source, then re-check
Reach for the fader and the arrangement before the EQ. Then run the same bounce again and watch whether the deviation actually closed. The before-and-after is the part that teaches you something.
- 5
Do not chase the guide
The guide is a rough sketch, not a target, and plenty of great records sit far from it. Use it to find accidents, not to sand off intent.
03 · Read the result
What your Mix Check result means
How to use this result
The style guides are hand-drawn, not measured from reference records, so nothing here is a grade or release approval. Audition the named band, compare with trusted records, and check translation before changing the mix.
No band stands outNothing in the tonal shape sits far from the guide. That is not release approval; reference by ear and check translation on other playback systems.
One band just outside the guideA band sits a little outside the guide's hand-set tolerance. Audition it, identify which sources own it, and change something only if you can hear a reason to.
One band well outside the guideA band sits well away from the guide. Start there, compare against a trusted reference record, and change one source or balance decision at a time.
Several bands far from the guideCheck the style choice and your monitoring first. The guide is a rough sketch, so a whole-spectrum difference often means the sketch fits your record poorly, not that the mix is wrong.
04 · Questions
Questions
Do you upload or keep my music?
No. The analysis runs in your browser with the Web Audio API. The file is never sent anywhere, and there is nothing to store or delete.
Is a band far from the guide bad?
Not by itself. The guides are hand-drawn, not measured from reference records, and well-mixed records can sit far from them. A gap is a prompt to listen to that band and decide whether the difference is intentional.
What file format should I use?
Use a full-song WAV when practical. Other browser-supported formats may work, but codec support varies by device; if a file will not decode, export a WAV and try again.
Why does the same mix read differently against another style?
Because each style has its own guide. The same bounce will sit closer to some sketches than others; that tells you about the sketches as much as the mix.
Does this replace referencing against commercial tracks?
No. It is faster at pointing to a band worth listening to, but listening against records you love is still the best training. Use both.
Keep practising
Keep the question, then audition the fix
Open Studio Mix to investigate the named band and compare the mix with a trusted reference.
Free account, no card. Tool results stay in your browser unless you choose to save them.
Stay inside Main Act Studio
Choose the next guide
Studio Mix · Listening benchmark
How the Golden Ears frequency test works
Best for: Finding which neighboring frequency regions still blur together
Read this guideLive Sound · Listening benchmark
How to identify microphone feedback by ear
Best for: Practicing fast region identification before a real soundcheck
Read this guide