Source code for music21.analysis.metrical
# ------------------------------------------------------------------------------
# Name: metrical.py
# Purpose: Tools for metrical analysis
#
# Authors: Christopher Ariza
# Michael Scott Asato Cuthbert
#
# Copyright: Copyright © 2009-2012 Michael Scott Asato Cuthbert
# License: BSD, see license.txt
# ------------------------------------------------------------------------------
'''
Various tools and utilities for doing metrical or rhythmic analysis.
See the chapter :ref:`User's Guide Chapter 14: Time Signatures <usersGuide_14_timeSignatures>`
for more information on defining
metrical structures in music21.
'''
from __future__ import annotations
import copy
import unittest
from music21 import environment
from music21 import stream
environLocal = environment.Environment('analysis.metrical')
[docs]
def labelBeatDepth(streamIn):
# noinspection PyShadowingNames
r'''
Modify a Stream in place by annotating metrical analysis symbols.
This assumes that the Stream is already partitioned into Measures.
>>> s = stream.Stream()
>>> ts = meter.TimeSignature('4/4')
>>> s.insert(0, ts)
>>> n = note.Note(type='eighth')
>>> s.repeatAppend(n, 8)
>>> s.makeMeasures(inPlace=True)
>>> post = analysis.metrical.labelBeatDepth(s)
>>> sOut = []
>>> for n in s.flatten().notes:
... stars = "".join([l.text for l in n.lyrics])
... sOut.append(f'{n.beatStr:8s} {stars}')
>>> print("\n".join(sOut))
1 ****
1 1/2 *
2 **
2 1/2 *
3 ***
3 1/2 *
4 **
4 1/2 *
'''
for m in streamIn.getElementsByClass(stream.Measure):
# this will search contexts
ts = m.getTimeSignatures(sortByCreationTime=False)[0]
# need to make a copy otherwise the .beat/.beatStr values will be messed up (1/4 the normal)
tsTemp = copy.deepcopy(ts)
tsTemp.beatSequence.subdivideNestedHierarchy(depth=3)
for n in m.notesAndRests:
if hasattr(n, 'tie') and n.tie is not None:
environLocal.printDebug(['note, tie', n, n.tie, n.tie.type])
if n.tie.type == 'stop':
continue
for unused_i in range(tsTemp.getBeatDepth(n.offset)):
n.addLyric('*')
return streamIn
[docs]
def thomassenMelodicAccent(streamIn: stream.Stream):
# noinspection PyShadowingNames
'''
Adds an attribute, 'melodicAccent' to each note's `.editorial`
within a :class:`~music21.stream.Stream` object
according to the method postulated in Joseph M. Thomassen, "Melodic accent: Experiments and
a tentative model," ''Journal of the Acoustical Society of America'', Vol. 71, No. 6 (1982) pp.
1598-1605; with, Erratum, ''Journal of the Acoustical Society of America'', Vol. 73,
No. 1 (1983) p.373, and in David Huron and Matthew Royal,
"What is melodic accent? Converging evidence
from musical practice." ''Music Perception'', Vol. 13, No. 4 (1996) pp. 489-516.
Similar to the humdrum melac_ tool.
.. _melac: https://www.humdrum.org/Humdrum/commands/melac.html
Takes in a Stream of :class:`~music21.note.Note` objects (use `.flatten().notes` to get it, or
better `.flatten().getElementsByClass(note.Note)` to filter out chords)
and adds the attribute to
each. Note that Huron and Royal's work suggests that melodic accent has a correlation
with metrical accent only for solo works/passages; even treble passages do not have a
strong correlation. (Gregorian chants were found to have a strong ''negative'' correlation
between melodic accent and syllable onsets)
Following Huron's lead, we assign a `melodicAccent` of 1.0 to the first note in a piece
and take the accent marker of the first interval alone to the second note and
of the last interval alone to be the accent of the last note.
Example from Thomassen, figure 5:
>>> s = converter.parse('tinynotation: 7/4 c4 c c d e d d')
>>> analysis.metrical.thomassenMelodicAccent(s.flatten().notes)
>>> for n in s.flatten().notes:
... (n.pitch.nameWithOctave, n.editorial.melodicAccent)
('C4', 1.0)
('C4', 0.0)
('C4', 0.0)
('D4', 0.33)
('E4', 0.5561)
('D4', 0.17)
('D4', 0.0)
'''
# we use .ps instead of Intervals for speed, since
# we just need perceived contours
maxNotes = len(streamIn) - 1
p2Accent = 1.0
for i, n in enumerate(streamIn):
if i == 0:
n.editorial.melodicAccent = 1.0
continue
elif i == maxNotes:
n.editorial.melodicAccent = p2Accent
continue
lastPs = streamIn[i - 1].pitch.ps
thisPs = n.pitch.ps
nextPs = streamIn[i + 1].pitch.ps
if lastPs == thisPs and thisPs == nextPs:
thisAccent = 0.0
nextAccent = 0.0
elif lastPs != thisPs and thisPs == nextPs:
thisAccent = 1.0
nextAccent = 0.0
elif lastPs == thisPs and thisPs != nextPs:
thisAccent = 0.0
nextAccent = 1.0
elif lastPs < thisPs and thisPs > nextPs:
thisAccent = 0.83
nextAccent = 0.17
elif lastPs > thisPs and thisPs < nextPs:
thisAccent = 0.71
nextAccent = 0.29
elif lastPs < thisPs < nextPs:
thisAccent = 0.33
nextAccent = 0.67
elif lastPs > thisPs > nextPs:
thisAccent = 0.5
nextAccent = 0.5
else: # pragma: no cover # should not happen
thisAccent = 0.0
nextAccent = 0.0
n.editorial.melodicAccent = thisAccent * p2Accent
p2Accent = nextAccent
# ------------------------------------------------------------------------------
class TestExternal(unittest.TestCase):
show = True
def testSingle(self):
'''
Need to test direct meter creation w/o stream
'''
from music21 import note
from music21 import meter
s = stream.Stream()
ts = meter.TimeSignature('4/4')
s.append(ts)
n = note.Note()
n.quarterLength = 1
s.repeatAppend(n, 4)
n = note.Note()
n.quarterLength = 0.5
s.repeatAppend(n, 8)
s = s.makeMeasures()
s = labelBeatDepth(s)
if self.show:
s.show()
class Test(unittest.TestCase):
pass
# ------------------------------------------------------------------------------
# define presented order in documentation
_DOC_ORDER = [labelBeatDepth]
if __name__ == '__main__':
import music21
music21.mainTest(Test) # , TestExternal)