Options
All
  • Public
  • Public/Protected
  • All
Menu

Interface MelodyConverterArgs

Converts between a monophonic, quantized NoteSequence containing a melody and the Tensor objects used by MusicVAE.

Melodies are represented as a sequence of categorical variables, representing one of three possible events:

  • A non-event, i.e. holding a note or resting. (0)
  • A note off. (1)
  • A note on with a specific pitch. (> 1)

The Tensor output by toTensor is a one-hot encoding of the sequence of labels extracted from the NoteSequence.

The expected Tensor in toNoteSequence is a one-hot encoding of melody sequence labels like those returned by toTensor.

param

The length of each sequence.

param

The minimum pitch to model. Those above this value will cause an error to be thrown.

param

The maximum pitch to model. Those above this value will cause an error to be thrown.

param

(default: true) If false, an error will be raised when notes start at the same step. If true, the highest pitched note is used and others are ignored.

param

(Optional) The number of conductor segments, if applicable.

Hierarchy

Index

Properties

Optional ignorePolyphony

ignorePolyphony: boolean

maxPitch

maxPitch: number

minPitch

minPitch: number

Optional numSegments

numSegments: number

Optional numSteps

numSteps: number

Generated using TypeDoc