AudioInstanceQ

AudioInstanceQ[audio,obj]

gives True if audio sounds to be an instance of the object obj, and gives False otherwise.

AudioInstanceQ[audio,obj,cat]

assumes that audio is the sound of something in the category cat.

Details and Options

  • AudioInstanceQ tries to identify whether an audio signal is an instance of a given category.
  • Possible forms for both object obj and category cat are:
  • "class"named sound class, as used in "Sound" entities
    Entity[]any appropriate entity
    instance1|instance2|any of the instancei
  • The following options can be given:
  • AcceptanceThresholdAutomaticminimum probability to consider acceptable
    MaskingAllinterval of interest
    PerformanceGoalAutomaticwhat to optimize in the identification
    RecognitionPrior0.5prior probability for a True result
    TargetDevice"CPU"the target device on which to compute
  • AudioInstanceQ uses machine learning. Its methods, training sets and biases included therein may change and yield varied results in different versions of the Wolfram Language.
  • AudioInstanceQ may download resources that will be stored in your local object store at $LocalBase, and can be listed using LocalObjects[] and removed using ResourceRemove.

Examples

open allclose all

Basic Examples  (2)

Test whether an audio signal is an instance of a particular object:

Test whether an audio signal is an instance of either of the given categories:

Scope  (1)

By default, no assumptions are made about the sound category of the audio signal:

Assume that the audio signal belongs to a specific sound category:

Assume that the audio signal belongs to a specific set of sound categories:

Options  (3)

AcceptanceThreshold  (1)

By default, an automatic acceptance threshold is used:

Specify the acceptance threshold:

Masking  (1)

By default, the entire audio signal is checked for the specified object:

Restrict the test to a specific interval:

RecognitionPrior  (1)

The default RecognitionPrior is 0.5:

Assign a very low prior probability to the object to be tested for:

Applications  (2)

Find all recordings in the ExampleData collection that are recognized as a musical instrument:

Find other sounds that are similar to a recording:

Compute an object identification using AudioIdentify:

Find sounds with the same identification in the ExampleData collection:

Compute a list of more general objects using the broader sound property of Entity objects:

Find sounds within broader categories:

Compute a larger list of related sounds using AudioIdentify:

Find similar sounds among the ExampleData collection:

Properties & Relations  (2)

AudioInstanceQ and AudioIdentify offer complementary functionality:

Compute the most probable object identification:

Pass the identification to AudioInstanceQ:

Get the probability associated with the identification:

The minimum acceptable probability corresponds to the confidence in the original identification:

Specifying a minimum acceptable probability for the tested-for object will have no effect if the tested-for object does not belong to one of the assumed categories, if specified.

Test whether a recording of an acoustic guitar is an animal:

Specify a very low acceptable threshold:

Test whether the recording is a guitar, assuming that it belongs to the category of animal sounds:

Specifying a very low acceptance threshold will not yield a match since a "guitar" sound is not in the "animal" category:

Introduced in 2020
 (12.1)