Audio to score transcription

Platform modules

Built on MusicXML and Score JSON

Each feature page maps back to the same score project model rather than a separate one-off converter.

Module

Audio upload

Connected to the score project workflow.

Module

Basic Pitch worker

Connected to the score project workflow.

Module

MIDI candidate

Connected to the score project workflow.

Module

Score JSON correction

Connected to the score project workflow.

Real product example

One score, structured input and reusable output

Captured from the candidate and revision workspace used after transcription. Captured 2026-07-15.

Score project workspace used to review an audio transcription candidate

Input

Permitted mono WAV or MP3 melody recordingDownload input case

Output

Basic Pitch MIDI candidate and first-pass editable Score JSONDownload output case

Polyphony, reverb and noisy accompaniment can reduce accuracy and require correction.

Workflow

How this path works

The public page leads into the authenticated app workflow when users are ready to process a score.

Upload an audio source

The app stores the source audio asset and queues an audio transcription job tied to a score project.

Generate a MIDI candidate

When configured, the worker calls Basic Pitch, stores the generated MIDI file, and records diagnostics in the score jobs panel.

Create an editable score draft

The platform attempts a first-pass MIDI-to-Score JSON revision so the result can be previewed, corrected, converted to Jianpu, played, transposed, and exported.

Product detail

What to expect

Candidate-first transcription

Audio transcription is treated as a draft import path, not a guaranteed final score. Polyphonic music, accompaniment, reverb, and noisy recordings usually need correction.

Same project workflow after import

Once a candidate revision exists, it uses the same Score JSON and MusicXML model as OMR, MIDI, Jianpu, and MusicXML imports.

FAQ

Questions about Audio to score transcription

Practical limits and expected workflow for this score tool.

What input does Audio to score transcription accept?

The app stores the source audio asset and queues an audio transcription job tied to a score project.

Can I edit the result after conversion?

Yes. Successful imports become Score JSON revisions that can be corrected, transposed, played, and exported from the score workspace.

What should I check before relying on the result?

Use this for permitted audio sources and expect human review. The site should not promise perfect automatic transcription from commercial recordings.

Related workflows

Continue with the same score project

These pages use the same MusicXML and Score JSON source instead of sending the score through disconnected converters.

Online score editor and correction

Correct imported scores through property panels for notes, rests, key/time signatures, lyrics, chord symbols, dynamics, wedges, articulations, and barlines.

MusicXML converter and MIDI score tools

Use MusicXML as the interchange format and Score JSON as the editable project model, with MIDI import/export and portable Score JSON snapshots.

Score to audio practice generator

Turn structured score projects into browser playback and queued MIDI, WAV, or MP3 practice exports with tempo, loops, count-in, metronome, and per-part controls.