Package com.tagtraum.audiokern

Interface SongPropertyAnalyzer<O>

All Known Implementing Classes:

LowLevelFeatureAnalyzer, RelativeSpectralEntropyMean, RMSMean, RMSStandardDeviation, SpectralCentroidMean, SpectralFluctuationPeak, SpectralFluxMean, SpectralNoveltyMean, SpectralNoveltyStandardDeviation, SpectralSpreadMean, SpectralVariabilityMean

public interface SongPropertyAnalyzer<O>

Capable of creating a pipeline that analyzes audio with an AudioSong-property as result.

Implementations might be picked up automatically by applications like beaTunes, if registered as plugins.

Author:

Hendrik Schreiber

Method Summary

Modifier and Type	Method	Description
SignalProcessor<AudioBuffer,O>	createPipeline()	Creates a pipeline that produces a result with the property name in question as id.
default String	getId()	Unique id for this analyzer.
String	getName()	Name of this analyzer in the format NAME 0.0.0, where NAME is the name and 0.0.0 its version.
String	getPropertyName()	Name of the property that shall be set on an AudioSong object.
AudioClip	getRequiredClip(AudioFileFormat audioFileFormat)	Indicates which portion of the audio this analyzer needs to have available to do its job.

Method Detail

getId

default String getId()

Unique id for this analyzer. By default the fully qualified classname is returned. Some scripting languages (e.g. Nashorn Javascript) may require you to explicitly override this method.

Returns:

id

getName

String getName()

Name of this analyzer in the format NAME 0.0.0, where NAME is the name and 0.0.0 its version.

Returns:

name

getRequiredClip

AudioClip getRequiredClip(AudioFileFormat audioFileFormat)

Indicates which portion of the audio this analyzer needs to have available to do its job. E.g. the first 20s.

Parameters:

audioFileFormat - file format

Returns:

required audio clip or null, if we need the whole file

createPipeline

SignalProcessor<AudioBuffer,O> createPipeline()

Creates a pipeline that produces a result with the property name in question as id. E.g. to calculate the key property of an AudioSong, this method should create a pipeline that produces a Key object under the id key.

This method must return a new pipeline every time it is called. You must not cache a once created pipeline and return it again and again.

When this instance is used, the produced pipeline is repeatedly called with audio data. Specifically, the SignalProcessor.process(Object) method is called with AudioBuffers for the desired clip (see getRequiredClip(AudioFileFormat)). The audio format is typically 44.1kHz, stereo, signed 16 bit.

Once all buffers are processed, the system calls SignalProcessor.flush() and then expects SignalProcessor.getOutput() to return the computed value.

Returns:

analysis pipeline

See Also:

SignalProcessor.getId(), SignalProcessor.flush(), SignalProcessor.process(Object), SignalProcessor.getOutput()

getPropertyName

String getPropertyName()

Name of the property that shall be set on an AudioSong object. This could be e.g. key, comments, grouping, ...

Returns:

property name