Transcript Document

Auditory Scene Analysis
Sequential vs. Simultaneous
Organisation
• Sequential grouping involves connecting
components over time to form streams
• Simultaneous grouping involves grouping the
components present at a given moment in time to
create one or more perceived sounds
• This lecture - how these two types of grouping
compete with one another in the formation of
auditory streams
Simultaneous vs. Sequential
Organisation
• Organisation at a point in time may conflict with
organisation over time
• Competition to establish the identity of sounds
• The relationship between two simultaneous tones
can be affected by the relationship between one of
the tones and other previous and/or successive
tones
• Vertical and horizontal relationships seem to
compete with each other
Simultaneous vs. Sequential
Organisation
• The relationships are influenced by:
• The differences (e.g. in pitch) between the
simultaneous components and the
sequential components
• The degree of synchrony of the
simultaneous components
• The number of components involved
Simultaneous vs. Sequential
Organisation
• Asa demo 25 Capturing a tonal component out of
a mixture: Part 1
• 3 pure tones: A, B and C. A is played first
followed by B and C play simultaneously
• A could form one stream (A- A…)
• The simultaneous tones B and C could form
another (BC- BC…)
• OR
• A and B could form a stream of pure tones (A B A
B…)
Simultaneous vs. Sequential
Organisation
• An important determinant of which of these
groupings will emerge in perception is the
frequency separation between the tones A and B.
• Large difference in frequency, A segregates from
B and B fuses with C
• If A and B are close in frequency they group
toegether to form a pure tone stream breaking the
tendency of B to fuse with C
• When A captures B – sequential integration is
stronger than simultaneous integration
Sequential vs. Simultaneous
Organisation
• Asa demo 26 – Capturing a tonal component out
of a mixture: Part 2
• Demonstrate that the fusion of components
depends on their synchronisation
• Pure tones A, BC as previous
• Frequency separation between A and B held
constant
• Vary asychrony of BC
• When B and C are synchronous they fuse into a
single sound – A forms a separate stream
Sequential vs. Simultaneous
Organisation
• B groups with A when B is asynchronous with C
• Asa demo 27 – Competition of sequential and
simultaneous grouping
• Demonstrates that if another component is added
that is close to one of the simultaneous
components – may also work against their fusion
• 3 tone cycle A, BC as before
• Tone D that is close in frequency to C is added to
the cycle – A, BC, D
Sequential vs. Simultaneous
Organisation
• The sequential grouping of C with D competes
and weakens the fusion of C with B
• C groups with D forming a C-D pure tone stream
– difficult to hear C as part of the complex BC –
• weakening of BC fusion has the effect of releasing
B to group more strongly with A in an A-B pure
tone stream
• C-D grouping affects A-B grouping, even though
neither A nor B themselves were changed
Sequential vs. Simultaneous
Organisation
• Asa demo 33 – Creation of a high-pitched residual
by capturing some harmonics from a complex tone
• Tone B, with a rich spectrum, alternates with tone
A (captor) that contains only a few lower
harmonics
• Under some conditions tone B will split - tone A
will capture out of B the lower harmonics it shares
with it - a perceptual stream of low sound
• The harmonics that are unique to B will be left
behind - heard as a separate higher sounding tone
Sequential vs. Simultaneous
Organisation
• The strength of the decomposition of B depends
on the length of the silent gap between A and B
• 1st hear tone B
• 2nd – A B alternate with a 200 msec silent gap
• 3rd – A B alternate with no silent gap
• Capturing of B harmonics by A is best when there
is no silent gap between A and B
Sequential vs. simultaneous
Organisation
• Asa demo 37 – Changing a vowel’s quality by
capturing a harmonic
• hear a synthetic “i” vowel whose third harmonic (a
pure tone) is kept on all the time while the rest of
the vowel goes on and off
• Idea: because the pure tone does not turn on and
off with the vowel, it does not integrate with it
perceptually, and the vowel sounds as if all or
some of the energy of its third harmonic is missing
Sequential vs. Simultaneous
Organisation
• Pure tone – sequentially capturing its counterpart
(the third harmonic) from the vowel.
• The “i” vowel minus its third harmonic sounds
like a different vowel – “e” or “en” (as in bed or
bend)
• 1st - hear 4 cycles of the “i” sound, then 4 of the
“en” sound
• Third harmonic present all the time, “i” vowel
going on and off 8 times, followed by a
comparison vowel.
Sequential vs. Simultaneous
Organisation
• when the “i” vowel comes on, some of the
energy of the third harmonic does not
combine with the rest of the vowel – change
in the vowel’s quality - “en”
• The quality of the comparison vowel
gradually changes from “i” to “en” during
the course of the trials
Memory
• All musical activities involve memory
• E.g. sight reading involves the memory of the
relationship between specific symbols and pitches
and the set of body and finger movements needed
to execute them
• Most knowledge about music has to be acquired
through learning – can be formal or informal
• Memory – a capacity that can be improved – not a
fixed quantity
Memory
• Memory – a practical requirement – such as in the
performance of complex music without a score
• memory often involves the ability to make sense
of incoming material in terms of previously learnt
information - when material cannot be assimilated
into familiar structures, memory performance
declines
• Importance of familiar structures for musical
memory
Memory
• Studies have indicated that listeners can
discriminate better between two short consecutive
pitch sequences when these sequences are drawn
from familiar tonal materials than from unfamiliar
materials
• Lifelong exposure to tonal music – tuning of our
cognitive systems to the structures and regularities
in this music
• Distinguish between two types of memory:
Memory
• Memory for specific pieces of music (episodic,
veridical)
• Memory for norms and prototypes which may be a
shared attribute of many pieces of music
(semantic, schematic)
• If a familiar piece of music is interrupted before
its close, episodic memory – reconstruct the actual
continuation, whereas schematic memory – allows
a guess of a likely continuation
Memory
• A good schematic memory is required for episodic
memory – easier to memorise a piece of music
once the structural interrelationships within the
music are understood
• Memory has two basic forms – recognition and
recall
• Recognition – listening
• Recall - performance
Recognition
• Recognition – a process that operates in
perception to match incoming information to
previously stored information
• One form of recognition is in the perception of
similarity – something heard is perceived as close
to or the same as something heard before
• The ability to recognise similarity between
structures in a piece – crucial in the perception of
form within a piece of music
Recognition
• Recognition of themes under various
transformations – listeners will recognise a
melody, with its interval relationships preserved,
as the same irrespective of its starting pitch –
importance of pitch contour
• evidence that even trained musicians find it hard
to keep the starting-key of a heard composition in
memory if it modulates several times
Recognition
• Some transformations make recogniton more
difficult – e.g. changing the rhythm of a melody
can significantly disrupt recognition even though
the pitch pattern may be unchanged
• Recognition ability – not static - can improve with
experience and training
• Second form of recognition – identification or
naming – retrieval of a verbal label e.g. for a note,
chord, or the musical piece.
Recognition
• Identification – involve recognition of
similarity – although recognition of a
similarity may not lead to identification
• e.g. recognise a familiar chord but unable to
name / identify it
Recall
• Recall – the reproduction of a previously
experienced sequence
• Requires more mental resources than recognition a listener will generally recognise much more than
can be recalled - recall may require some form of
cue to trigger it.
• Cued recall - performing learnt music from a score
- an experienced performer will rarely look at
every note.
Recall
• Can be regarded as either an unintended byproduct of other activities or as the result of
deliberate memorizing efforts.
• Most everyday examples of musical recall are
involuntary and unintended results of other mental
processes. – e.g. due to repeated hearings of radio
/ tv themes and jingles – reproducible without any
particular effort
• However, recall of more complex performance
music – through a process of deliberate and
conscious memorization
Recall
• Characteristics of expert memorizers:
• Long time experience in memorizing
• Possess the ability to represent the material to be
learnt in terms of patterns and structures that have
rich interconnections with each other and with
previously learnt material.
• develop multiple interlocking levels of
representation (e.g. visual, auditory, motor
movements), such that if any one of them is
temporarily lost it can be re-cued from another
level.
Recall
• Representations are flexible – e.g. an
experienced performer could transpose a
piece learnt in one key to another with not
much extra effort