For audio engineers
A fast first pass you can measure
Euphona does the fast, consistent first pass and proves it with numbers you can check against your own meters. Every call it makes is yours to overrule.
What gets in the way
Automatic mastering is sold as a replacement for you
It is not one, and the marketing that says otherwise is why the category is treated with suspicion. The useful job is narrower and more valuable: a fast, consistent correction pass that a person supervises.
Nobody publishes their metering
A loudness figure with no standard, gating or window behind it is decoration. When two tools disagree about a track's LUFS, one of them is measuring something else and neither will say which.
Results arrive without proof
A file that is louder than before proves nothing. A file whose loudness, true peak and range are measured before and after, to a stated standard, is a result you can check, defend and deliver.
The workflow
How it goes
- 01
Measurement you can reproduce
Integrated loudness to ITU-R BS.1770-5 with the standard's gating, loudness range per EBU Tech 3342, true peak at 4× oversampling. Stated on every result, so a disagreement with your own meter has an answer.
- 02
Decisions from your material
Every move is derived from a measurement of this source, from a reference you supplied, or from a rendition you chose by ear. There is no preset table and no genre guess, and the report carries the measured result.
- 03
Overrides that stick
Set the loudness, set the ceiling, supply a reference. Nothing recomputes your override on the next run.
- 04
Failure that is stated
Jobs that fail for our reasons are not charged, an identical re-run returns the earlier result instead of charging twice, and the fault codes are part of the published API contract, not a support-ticket outcome.
The suite
The tools you would use
AI Mastering
Loud, clear and ready for every streaming service, with the numbers to prove it.
Audio Restoration
Rescue a recording: lose the noise, echo and hum, or keep only the voice.
AI Mixing
A balanced first mix from your stems, in minutes.
Spatial Audio
Immersive headphone and surround mixes from your stems.
Questions
What people ask before starting
- What standard is the loudness measurement?
- ITU-R BS.1770 for integrated loudness and true peak, with the peak 4× oversampled, and loudness range as EBU Tech 3342 describes it. The test suite checks the meter against the standard's own calibration point and against libebur128, an independent open-source meter, on real recordings, so the claim is a test result rather than a sentence.
- What does this do that a chain of plugins does not?
- It derives the settings from a measurement of your material instead of settings tuned on someone else's. You hear it most in what it leaves alone: it will not compress a track that already has the dynamics it needs, because it measured that it does.
- Can I get the analysis without the processing?
- Yes. Every upload is measured free before any job runs: integrated loudness, loudness range, true peak, peak-to-loudness ratio, clipped runs, noise floor and stereo correlation, and the measurement is never metered.
- Can I override what it chose?
- Everything the chain works from: the target, the ceiling and the reference, and the report measures what came out. On a vocal, even the strongest correction stays natural.
- Where does my audio go, and who can hear it?
- To our own servers in the European Union, which store it, and while one of our AI models works on it, to GPU computers we rent in the European Economic Area, which delete their copy when the work is done. Staff cannot listen to it by default, every internal read is recorded, and audio moves only through signed URLs. It is never used to train anything without your separate, explicit agreement.
Next