MP4 to MIDI
Drop an MP4 or another media file here, or choose one
MP4, MOV, WebM and MKV video, and WAV, MP3, OGG, FLAC and AAC audio. Up to 2 GB and 15 minutes.
The notes found in the section you pick appear here, with their range and length, a MIDI file to download and a piano roll to edit them in.
What MP4 to MIDI does with a video's audio track
Pick a video file and the page decodes the sound inside it, then transcribes the pitches it hears into a Standard MIDI File. The file is read and transcribed in the browser tab, so nothing is uploaded and nothing is billed.
Reads the audio track of an MP4
MP4, MOV, WebM and MKV files carry their sound as a separate track. The page reads that track alone from the file, so a long or high-resolution video takes no longer to open than its soundtrack.
Writes what it hears as MIDI notes
The basic-pitch model behind this page detects note onsets, pitches and durations, and writes each one into a downloadable MIDI file.
Keeps the file on your device
Decoding and transcription both run in the tab, with the model fetched from this site, so the video never leaves your machine and no credits are drawn.
How it works
Three moves from a video file to a MIDI file
Choose a media file
Pick an MP4, another video format or an audio file. The size is checked before anything is read.
Pick the section to transcribe
The audio track is decoded at 22,050 Hz and mixed down to one channel; the file plays in the page, and the start and end sliders choose the part that goes to the transcription model.
Download the MIDI
When transcription finishes, the page reports the notes it found, their lowest and highest pitch and the section length, draws them as a piano roll and offers a .mid download.
The limits behind a video-to-MIDI conversion
The largest file the page opens. Only its audio track is read; a file in a format the page does not recognize is read whole, up to 90 MB.
The longest audio track this page transcribes. A longer track is refused from the file's header when the header states the length, and otherwise as soon as decoding reaches the limit.
How many tracks are read per file: its primary audio track. The video and any other tracks are skipped.
The sample rate the audio track is decoded at before it reaches the transcription model.
The result is a Standard MIDI File carrying each note's onset, pitch and duration.
The transcription runs in the browser and submits no generation job.
Who turns video audio into MIDI
Songwriters lifting a melody from a clip
Capture a hummed idea from a phone video and open the notes in a DAW instead of replaying the clip by ear.
Producers rebuilding a part they cannot notate
Turn a recorded instrument line into editable notes, then change the sound without re-recording the take.
Teachers preparing a score from a recording
Transcribe a student's performance or a reference recording and bring the notes into notation software.
MP4 to MIDI FAQ
Take the transcription further
Download the .mid file, or open the audio to MIDI workbench when you want to adjust the detection thresholds before exporting.