MusicUnderstanding is a machine learning framework for analyzing the musical properties of audio, including instrument activity, loudness, and other acoustic characteristics. You drive analysis through a MusicUnderstandingSession actor, selecting the properties to examine with AnalysisType, and receive results as dedicated value types such as InstrumentActivityResult, LoudnessResult, KeyResult, PaceResult, RhythmResult, and StructureResult. Failures during analysis surface as MusicUnderstandingError. The framework is available on iOS, macOS, tvOS, watchOS, and visionOS.
Essentials 2
The session actor and the configuration value that drive music analysis.
- AcMusicUnderstandingSessioniOS 27.0+An object that performs music analysis on an audio source and provides the results.
- StAnalysisTypeiOS 27.0+A type that specifies which musical properties an analysis session examines.
Analysis Results 6
Value types that carry the musical properties detected during analysis.
- StInstrumentActivityResultiOS 27.0+A structure containing the activity levels and detected time ranges for instruments in a song.
- StLoudnessResultiOS 27.0+## Discussion
- StKeyResultiOS 27.0+A value describing the musical key detected over a time range.
- StPaceResultiOS 27.0+A structure that contains pace analysis results for a song.
- StRhythmResultiOS 27.0+A structure that describes the rhythm of analyzed audio.
- StStructureResultiOS 27.0+A structure that provides a song's structure boundary information including sections, segments, and phrases.
Errors 1
The error type that reports failures encountered during analysis.
- EnMusicUnderstandingErroriOS 27.0+An error that occurs during a music understanding session.