music21.analysis.discrete

Modular analysis procedures for use alone or applied with music21.analysis.windowed.WindowedAnalysis class.

All procedures should inherit from music21.analysis.discrete.DiscreteAnalysis, or provide a similar interface.

The music21.analysis.discrete.KrumhanslSchmuckler (for algorithmic key detection) and music21.analysis.discrete.Ambitus (for pitch range analysis) provide examples.

DiscreteAnalysis

class music21.analysis.discrete.DiscreteAnalysis(referenceStream=None)

Parent class for analytical methods.

Each analytical method returns a discrete numerical (or other) results as well as a color. Colors can be used in mapping output.

Analytical methods may make use of a referenceStream to configure the processor on initialization.

DiscreteAnalysis methods

DiscreteAnalysis.clearSolutionsFound()

Clear all stored solutions

DiscreteAnalysis.getColorsUsed()

Based on solutions found so far with this processor, return the colors that have been used.

DiscreteAnalysis.getSolution(subStream)

For a given Stream, apply the analysis and return the best solution.

DiscreteAnalysis.getSolutionsUsed()

Based on solutions found so far with this processor, return the solutions that have been used.

DiscreteAnalysis.process(sStream)

Given a Stream, apply the analysis to all components of this Stream. Expected return is a solution (method specific) and a color value.

DiscreteAnalysis.solutionLegend(compress=False)

A list of pairs showing all discrete results and the assigned color. Data should be organized to be passed to music21.graph.GraphColorGridLegend.

If compress is True, the legend will only show values for solutions that have been encountered.

DiscreteAnalysis.solutionToColor(solution)

Given an analysis specific result, return the appropriate color. Must be able to handle None in the case that there is no result.

DiscreteAnalysis.solutionUnitString()

Return a string describing the solution values. Used in Legend formation.

Ambitus

class music21.analysis.discrete.Ambitus(referenceStream: stream.Stream | None = None)

A basic analysis method for measuring register.

>>> ambitusAnalysis = analysis.discrete.Ambitus()
>>> ambitusAnalysis.identifiers[0]
'ambitus'

Ambitus bases

Ambitus methods

Ambitus.getPitchSpan(subStream) tuple[music21.pitch.Pitch, music21.pitch.Pitch] | None

For a given subStream, return a tuple consisting of the two pitches with the minimum and maximum pitch space value.

This public method may be used by other classes.

ignores ChordSymbol objects…

>>> s = corpus.parse('bach/bwv66.6')
>>> p = analysis.discrete.Ambitus()
>>> pitchMin, pitchMax = p.getPitchSpan(s.parts[0].getElementsByClass(stream.Measure)[3])
>>> pitchMin.ps, pitchMax.ps
(66.0, 71.0)
>>> p.getPitchSpan(s.parts[0].getElementsByClass(stream.Measure)[6])
(<music21.pitch.Pitch A4>, <music21.pitch.Pitch C#5>)
>>> s = stream.Stream()
>>> c = chord.Chord(['a2', 'b4', 'c8'])
>>> s.append(c)
>>> p.getPitchSpan(s)
(<music21.pitch.Pitch A2>, <music21.pitch.Pitch C8>)

Returns None if the stream contains no pitches.

>>> s = stream.Stream(note.Rest())
>>> p.getPitchSpan(s) is None
True
Ambitus.getSolution(sStream)

Procedure to only return an Interval object.

>>> s = corpus.parse('bach/bwv66.6')
>>> p = analysis.discrete.Ambitus()
>>> p.getSolution(s)
<music21.interval.Interval m21>
Ambitus.process(sStream)

Given a Stream, return a solution (as an interval) and a color string.

>>> p = analysis.discrete.Ambitus()
>>> s = stream.Stream()
>>> c = chord.Chord(['a2', 'b4', 'c8'])
>>> s.append(c)
>>> p.process(s)
(<music21.interval.Interval m38>, '#665288')
Ambitus.solutionLegend(compress=False)

Return legend data.

>>> s = corpus.parse('bach/bwv66.6')
>>> soprano = s.parts[0]
>>> p = analysis.discrete.Ambitus(soprano)  # provide ref stream
>>> p.solutionLegend()
[['',
  [(0, '#130f19'), (1, '#211a2c'), (2, '#2f263f'),
   (3, '#3e3253'), (4, '#4c3d66'), (5, '#5b4979')]],
 ['',
  [(6, '#69548c'), (7, '#775f9f'), (8, '#866bb2'), (9, '#9476c5'),
   (10, '#a382d9'), (11, '#b18eec'), (12, '#bf99ff')]]]
>>> len(p.solutionLegend())
2
>>> [len(x) for x in p.solutionLegend()]
[2, 2]
>>> [len(y) for y in [x for x in p.solutionLegend()]]
[2, 2]
>>> s = corpus.parse('bach/bwv66.6')
>>> p = analysis.discrete.Ambitus()
>>> p.solutionLegend(compress=True)  # empty if nothing processed
[['', []], ['', []]]
>>> x = p.process(s.parts[0])
>>> [len(y) for y in [x for x in p.solutionLegend(compress=True)]]
[2, 2]
>>> x = p.process(s.parts[1])
>>> [len(y) for y in [x for x in p.solutionLegend(compress=True)]]
[2, 2]
Ambitus.solutionToColor(solution)
>>> p = analysis.discrete.Ambitus()
>>> s = stream.Stream()
>>> c = chord.Chord(['a2', 'b4', 'c8'])
>>> s.append(c)
>>> minPitch, maxPitch = p.getPitchSpan(s)
>>> p.solutionToColor(maxPitch.ps - minPitch.ps).startswith('#')
True
Ambitus.solutionUnitString()

Return a string describing the solution values. Used in Legend formation.

Methods inherited from DiscreteAnalysis:

MelodicIntervalDiversity

class music21.analysis.discrete.MelodicIntervalDiversity(referenceStream=None)

An analysis method to determine the diversity of intervals used in a Stream.

MelodicIntervalDiversity bases

MelodicIntervalDiversity methods

MelodicIntervalDiversity.countMelodicIntervals(sStream, found=None, ignoreDirection=True, ignoreUnison=True)

Find all unique melodic intervals in this Stream.

If found is provided as a dictionary, this dictionary will be used to store Intervals, and counts of Intervals already found will be incremented.

MelodicIntervalDiversity.getSolution(sStream)

Solution is the number of unique intervals.

MelodicIntervalDiversity.process(sStream, ignoreDirection=True)

Find how many unique intervals are used in this Stream

MelodicIntervalDiversity.solutionToColor(solution)

Given an analysis specific result, return the appropriate color. Must be able to handle None in the case that there is no result.

Methods inherited from DiscreteAnalysis:

KeyWeightKeyAnalysis

class music21.analysis.discrete.KeyWeightKeyAnalysis(referenceStream=None)

Base class for all key-weight key analysis subclasses.

KeyWeightKeyAnalysis bases

KeyWeightKeyAnalysis methods

KeyWeightKeyAnalysis.getSolution(sStream)

Return a music21 Key object defining the results of the analysis. Do not call process before calling this method, as this method calls process.

Note that all alternative solutions are returned as Key objects and stored on a list found at Key.alternateInterpretations.

>>> s = corpus.parse('bach/bwv66.6')
>>> p = analysis.discrete.KrumhanslSchmuckler()
>>> p.getSolution(s)  # this seems correct
<music21.key.Key of f# minor>
>>> s = corpus.parse('bach/bwv57.8')
>>> p = analysis.discrete.KrumhanslSchmuckler(s)
>>> p.getSolution(s)
<music21.key.Key of B- major>
KeyWeightKeyAnalysis.getWeights(weightType='major') list[float]

Returns the key weights. To provide different key weights, subclass and override this method. The defaults here are KrumhanslSchmuckler.

>>> a = analysis.discrete.KrumhanslSchmuckler()
>>> len(a.getWeights('major'))
12
>>> len(a.getWeights('minor'))
12
KeyWeightKeyAnalysis.process(sStream, storeAlternatives=False)

Takes in a Stream or sub-Stream and performs analysis on all contents of the Stream. The WindowedAnalysis windowing system can be used to get numerous results by calling this method.

Returns two values, a solution data list and a color string.

The data list contains a key (as a string), a mode (as a string), and a correlation value (degree of certainty)

KeyWeightKeyAnalysis.solutionLegend(compress=False)

Returns a list of lists of possible results for the creation of a legend.

>>> p = analysis.discrete.KrumhanslSchmuckler()
>>> post = p.solutionLegend()
KeyWeightKeyAnalysis.solutionToColor(solution)

Given a two-element tuple of (tonicPitch, modality) return the proper color

>>> p = analysis.discrete.KrumhanslSchmuckler()
>>> solution = (pitch.Pitch('C'), 'major')
>>> p.solutionToColor(solution)
'#ff816b'
KeyWeightKeyAnalysis.solutionUnitString()

Return a string describing the solution values. Used in Legend formation.

Methods inherited from DiscreteAnalysis:

SimpleWeights

class music21.analysis.discrete.SimpleWeights(referenceStream=None)

Implementation of simple weights by Craig Sapp for Krumhansl-Schmuckler key determination algorithm.

Values from from http://extras.humdrum.org/man/keycor/, which describes these weightings as “Performs most consistently with large regions of music, becomes noisier with smaller regions of music.”

SimpleWeights bases

SimpleWeights methods

SimpleWeights.getWeights(weightType='major')

Returns the key weights.

>>> a = analysis.discrete.SimpleWeights()
>>> len(a.getWeights('major'))
12
>>> len(a.getWeights('minor'))
12

Methods inherited from KeyWeightKeyAnalysis:

Methods inherited from DiscreteAnalysis:

AardenEssen

class music21.analysis.discrete.AardenEssen(referenceStream=None)

Implementation of Aarden-Essen weightings for Krumhansl-Schmuckler key determination algorithm.

Values from from http://extras.humdrum.org/man/keycor/, which describes these weightings as “Weak tendency to identify the subdominant key as the tonic.”

(N.B. – we are not sure exactly where the minor weightings come from, and recommend only using these weights for major).

AardenEssen bases

AardenEssen methods

AardenEssen.getWeights(weightType='major')

Returns the key weights.

>>> a = analysis.discrete.AardenEssen()
>>> len(a.getWeights('major'))
12
>>> len(a.getWeights('minor'))
12

Methods inherited from KeyWeightKeyAnalysis:

Methods inherited from DiscreteAnalysis:

BellmanBudge

class music21.analysis.discrete.BellmanBudge(referenceStream=None)

Implementation of Bellman-Budge weightings for Krumhansl-Schmuckler key determination algorithm.

Values from from http://extras.humdrum.org/man/keycor/, which describes these weightings as “No particular tendencies for confusions with neighboring keys.”

BellmanBudge bases

BellmanBudge methods

BellmanBudge.getWeights(weightType='major')

Returns the key weights.

>>> a = analysis.discrete.BellmanBudge()
>>> len(a.getWeights('major'))
12
>>> len(a.getWeights('minor'))
12
>>> a.getWeights('major')
[16.8..., 0.8..., 12.9..., 1.4..., ...]

Methods inherited from KeyWeightKeyAnalysis:

Methods inherited from DiscreteAnalysis:

KrumhanslSchmuckler

class music21.analysis.discrete.KrumhanslSchmuckler(referenceStream=None)

Implementation of Krumhansl-Schmuckler/Kessler weightings for Krumhansl-Schmuckler key determination algorithm.

Values from from http://extras.humdrum.org/man/keycor/, which describes these weightings as “Strong tendency to identify the dominant key as the tonic.”

  • Changed in v6.3: it used to be that these were different from the Kessler profiles, but that was likely a typo. Thus, KrumhanslKessler and KrumhanslSchmuckler are synonyms of each other.

KrumhanslSchmuckler bases

KrumhanslSchmuckler methods

KrumhanslSchmuckler.getWeights(weightType='major')

Returns the key weights. To provide different key weights, subclass and override this method. The defaults here are KrumhanslSchmuckler.

>>> a = analysis.discrete.KrumhanslSchmuckler()
>>> len(a.getWeights('major'))
12
>>> len(a.getWeights('minor'))
12

Methods inherited from KeyWeightKeyAnalysis:

Methods inherited from DiscreteAnalysis:

TemperleyKostkaPayne

class music21.analysis.discrete.TemperleyKostkaPayne(referenceStream=None)

Implementation of Temperley-Kostka-Payne weightings for Krumhansl-Schmuckler key determination algorithm.

Values from from http://extras.humdrum.org/man/keycor/, which describes these weightings as “Strong tendency to identify the relative major as the tonic in minor keys. Well-balanced for major keys.”

TemperleyKostkaPayne bases

TemperleyKostkaPayne methods

TemperleyKostkaPayne.getWeights(weightType='major')

Returns the key weights.

>>> a = analysis.discrete.TemperleyKostkaPayne()
>>> len(a.getWeights('major'))
12
>>> len(a.getWeights('minor'))
12

Methods inherited from KeyWeightKeyAnalysis:

Methods inherited from DiscreteAnalysis:

Functions

music21.analysis.discrete.analyzeStream(streamObj: stream.Stream, method: str, **keywords)

Public interface to discrete analysis methods to be applied to a Stream given as an argument. Methods return process-specific data format. See subclasses for details.

Analysis methods can be specified as arguments or by use of a method keyword argument. If method is the class name, that class is returned. Otherwise, the identifiers list of all DiscreteAnalysis subclass objects will be searched for matches. The first match that is found is returned.

Ambitus KrumhanslSchmuckler

>>> s = corpus.parse('bach/bwv66.6')
>>> analysis.discrete.analyzeStream(s, 'Krumhansl')
<music21.key.Key of f# minor>
>>> analysis.discrete.analyzeStream(s, 'ambitus')
<music21.interval.Interval m21>
>>> analysis.discrete.analyzeStream(s, 'key')
<music21.key.Key of f# minor>
>>> analysis.discrete.analyzeStream(s, 'span')
<music21.interval.Interval m21>

Note that the same results can be obtained by calling “analyze” directly on the stream object: >>> s.analyze(‘key’) <music21.key.Key of f# minor> >>> s.analyze(‘span’) <music21.interval.Interval m21>

music21.analysis.discrete.analysisClassFromMethodName(method: str) type[music21.analysis.discrete.DiscreteAnalysis] | None

Returns an analysis class given a method name, or None if none can be found

Searches first the class name, then the .identifiers array for each class, then a subset of any identifier.

>>> acfmn = analysis.discrete.analysisClassFromMethodName
>>> acfmn('aarden')
<class 'music21.analysis.discrete.AardenEssen'>
>>> acfmn('span')
<class 'music21.analysis.discrete.Ambitus'>

This one is fundamentally important…

>>> acfmn('key')
<class 'music21.analysis.discrete.AardenEssen'>
>>> print(repr(acfmn('unknown-format')))
None