Help & About

This app measures the real tuning reference of any audio file you give it — the current A4 frequency — then retunes the whole recording so A4 lands on exactly 444.000 Hz or 432.000 Hz, whichever you choose in Settings. Every note moves by the same ratio, nothing is time-stretched, and all processing happens locally in your browser. Your audio is never uploaded to a server.

How to use it

  1. Drop in — or tap to choose — an .mp3, .wav, .aac, .m4a or .flac file. Select several files at once to use batch mode (up to 30 files or 1 GB, whichever comes first).
  2. Review the detected A4, its cents offset from the target, and the file details.
  3. Choose your tolerance, then press Retune. The target is 444 Hz by default; switch to 432 Hz in Settings or from either landing page.
  4. Check the verification result — the measured A4 of the converted file, whether it passed, and how many passes it took. Press “Re-verify tuning” any time to re-measure.
  5. Compare the original and retuned audio in the A/B player, then download. The download is always the same file type you uploaded — a WAV in, a WAV out; an MP3 in, an MP3 out. Batch results download individually or as a single ZIP.

How detection works

The track is sampled across many analysis windows and transformed with an FFT. Each strong spectral peak is refined by interpolation, then converted into a cents deviation from the equal-tempered grid. Those deviations are combined with a shift-equivariant circular weighted mean — so the estimate does not drift toward 440 — giving the track's true tuning reference, reported to three decimal places.

How retuning works

The exact ratio target ÷ detected-A4 is applied to the entire recording, so every partial moves together and no pitch-shifting artifacts are introduced. WAV and FLAC are retuned losslessly by rescaling the sample rate; MP3, AAC and M4A are re-encoded, because those formats only permit fixed sample rates. Because the shift is uniform, track length changes by that same ratio — under one percent either way.

Verification & accuracy

The encoded output file is decoded and re-analyzed with the same detector — not just the in-memory buffer — so what you hear is what was measured. A result passes when the verified A4 lands inside the selected tolerance band: ±0.4 Hz (reported as 444), ±0.04 Hz (444.0) or ±0.004 Hz (444.00), and the same for 432. If it misses, the app re-converts from the freshly measured value and tries again, keeping the closest result it has found. Iterations stop early if the results start cycling or stop improving, so no time is wasted, and always stop at your maximum iteration count.

Global settings

Settings are saved on this device and apply to every conversion:

  • Tuning — target A4 of 444 Hz (default) or 432 Hz.
  • Tolerance — ±0.4, ±0.04 (default) or ±0.004 Hz. Verified results are reported to 0, 1 or 2 decimals to match.
  • Max iterations — how many retune passes a track may take before it stops. Whole numbers from 1 to 44; the default is 8, which is ample since most tracks converge in one or two passes.
  • Audio processing — high precision (default) uses more analysis windows and a larger FFT; fast render is quicker and lighter on CPU.
  • Appearance — light (default), dark or follow your system.

Output format always matches the source file, so there is nothing to configure there. Batch files are always measured and corrected individually, which is the most consistent method.

The top menu bar

  • Back — steps back through the conversion flow, or returns to the converter.
  • Refresh — clears the current file and result so you can start over.
  • Settings — the global preferences described above.
  • Help / About — this page.