Original DAW bounces
A direct render of the target instrument preserves attacks, low-level notes and harmonics better than a lossy distribution copy.
Use a lossless WAV when you want the strongest input this model can receive. Your file stays on this device from decoding through MIDI export.
For a DAW export, choose a dry solo track at 44.1 kHz and 16-bit or better—not the compressed master bus.
6.5-second synthetic melody · a workflow demo, not an accuracy benchmark.
The selected file is {size}. MIDIFLOW does not upload it or impose a server limit, but very large files can run out of memory, especially on a phone.
Compare the original with a simple synth preview of the detected notes. The preview does not reproduce the original sound or pitch bends.
Too many or missing notes? — increase it to keep fewer, more confident notes; decrease it to include quieter notes. Then compare the previews again.
Free foreverNo signupNo watermarkAudio never uploaded
Best input
A standard PCM WAV stores the waveform without perceptual encoding. It cannot guarantee a perfect transcription, but it avoids throwing away musical detail before analysis begins.
A direct render of the target instrument preserves attacks, low-level notes and harmonics better than a lossy distribution copy.
Portable recorders and digital instruments can provide uncompressed captures with stable timing and enough bandwidth for note estimation.
Because WAV is lossless, trimming silence or normalizing conservatively in an editor does not require another perceptual encode.
What can go wrong
WAV describes a container, not a guaranteed recording standard. An 8 kHz or 16 kHz WAV has already discarded upper-frequency content, while an 8-bit recording has coarse level resolution and audible quantization noise.
Room noise, preamp hiss and clipping are preserved faithfully too. Lossless storage cannot repair a microphone placed too far away, a distorted input stage or background sounds that mask a quiet note.
A WAV made by decoding an MP3 remains limited by the MP3 source. File size increases, but the missing information and compression artifacts stay in the samples.
Choose an MP3, WAV, M4A, FLAC or OGG recording, or try the sample. Start with a short solo part to check the result.
The converter estimates notes on this device. Your recording is not uploaded and no account is needed.
Compare the original with Play MIDI notes, apply a different note threshold if needed, then save the .mid file for your music software.
Transcription is powered by Spotify's open-source Basic Pitch model and runs in your browser.
Improve accuracy
Export the musical part before mastering. The aim is clear note boundaries and an honest dynamic range, not a loud finished mix.
A 44.1 or 48 kHz PCM export with at least 16-bit depth is a sensible source. The converter will then make its own required 22.05 kHz mono analysis copy.
Export the instrument channel you want, not the stereo master. Leave drums, vocals and unrelated accompaniment out of that bounce.
Avoid bus limiting, broad compression and long reverb. Those processes flatten onsets or extend note tails in ways that make MIDI boundaries less clear.
The download is a standard MIDI file containing note timing, pitch, velocity and eligible pitch-bend events. It does not contain the original instrument sound, so assign any software instrument after importing it.
Create or choose a MIDI track, then drag the .mid file into an empty clip slot or the Arrangement. Load an instrument on that track and edit notes in the MIDI Note Editor.
Drag the file into the Tracks area and choose a Software Instrument track. Open the Piano Roll to correct timing, note lengths or velocity before arranging.
Drag the file into the Channel Rack or use File › Import › MIDI file. Send the imported notes to a chosen instrument, then clean them in the Piano roll.
An original PCM WAV is the better technical source because it avoids lossy encoding. A noisy full-mix WAV can still perform worse than a clean solo MP3, so content remains important.
Use 44.1 or 48 kHz when available. MIDIFLOW keeps the original file for playback and separately resamples a mono analysis buffer to the model's required 22.05 kHz.
Not necessarily. Twenty-four-bit audio offers more recording headroom, but a clean 16-bit export already has enough resolution for this task. Arrangement and noise usually matter more.
It may decode, but an 8 kHz sample rate only represents frequencies below roughly 4 kHz. That limited bandwidth can weaken the harmonic cues used for pitch estimation.
Gentle peak normalization can help a very quiet recording, but avoid clipping or heavy loudness processing. The model benefits more from a clean signal than a maximally loud one.
Uncompressed files use more memory and long recordings require more model windows. The work happens on your device, so available RAM and processor speed determine the wait.