Tool Open source
Count-In is a music-practice application that detects a song's tempo and first downbeat, then adds a one- or two-bar drum count-in in 4/4, 3/4, or 6/8. Count-in patterns can be even, built up in the second half, or imported from a MIDI file; the downbeat and tempo can be corrected on the waveform when detection is wrong.
It can also create backing tracks by using a neural network to separate guitar, vocals, bass, drums, and piano, with independent levels and mute controls. Quiet sections can receive hi-hat beats, and playback can be slowed to 80% while preserving pitch. The application accepts MP3, FLAC, ALAC, and WAV files and exports those formats while preserving tags and cover art.
Count-In works offline after its models are downloaded, uses Core ML on Apple Silicon Macs and DirectML on Windows for separation, and includes a graphical application and command-line interface. It provides unsigned builds for Windows 10/11 x64 and macOS 12 or later on Apple Silicon, and is released under the MIT License.
1 use taken from transcripts — each links to the moment in the video.
A music practice app that detects a song's tempo and first downbeat, then adds a drum count-in before it, with waveform timing correction, count-in patterns importable from MIDI, instrument separation, and slowed playback that preserves pitch.
1 in the library.