Timbre

Question 1

Define Timbre

Answer 1

“that attribute of auditory sensation which enables a listener to judge that two nonidentical sounds, similarly presented and having the same loudness and pitch, are dissimilar” - Acoustical society of America

Question 2

What is timbre independent of?

Answer 2

Loudness

Question 3

What does timbre depend on?

Answer 3

The object’s physical properties

Question 4

What does timbre play a crucial role in?

Answer 4

Speech-based communication -

Question 5

What is timbre useful for?

Answer 5

• Timbre ends up being a very useful cue for making sense of complex auditory environments, when there are many different sound sources at the same time
• The brain uses timbre as a cue to organise sound into discrete auditory streams, where each stream might correspond to a particular speaker in a conversation, or to a different musical part in a piece of polyphonic music.

Question 6

What are free-text methods?

Answer 6

Ask them to write free text about what they heard - this can get very rich data

Question 7

What are the downsides to free-etxt?

Answer 7

However, this kind of free-form data is complex to analyse quantitatively because of its unstructured format. Two participants can easily write descriptions that on the surface look rather different, even though their subjective experience was relatively similar.

Question 8

What are semantic rating scales?

Answer 8

Get the participant to evaluate a single perceptual dimension of the sound at a time, for example its ‘brightness’. The participant rates this dimension on a numeric scale, for example from ‘very dark’ to ‘very bright’. This means that we can then average results over multiple participants to get a more reliable outcome.

Question 9

What are the downsides to semantic rating scales?

Answer 9

This comes at the expense of having to restrict the available response options a priori to one or more terms preselected by the researcher. This is feasible for domains that have already been studied in some detail, but it is not very practical for domains about which we currently know little.

Question 10

What are similarity judgements?

Answer 10

In the classic similarity judgement paradigm, we play participants pairs of sounds, and we ask them to rate the similarity of each pair.

Question 11

What are the benefits to similarity judgements?

Answer 11

The appealing feature of this approach is that similarity judgements do not need to be mediated by any verbal vocabulary; participants are free to compare the sounds based on whatever perceptual features they deem relevant, even if they don’t have good words for them.

Question 12

What does a similarity matrix show?

Answer 12

Each number in a similarity matrix tells us the average perceptual similarity of a pair of stimuli.

Question 13

What is an issue with a similarity matrix?

Answer 13

They can get big very quickly

Question 14

What is dimensionality reduction?

Answer 14

Some kind of statistical technique to distil this data down and create some meaningful quantitative insights

Question 15

What is multidimensional scaling (MDS)?

Answer 15

MDS takes the similarity matrices as its input, and uses them to compute a low-dimensional space, constructed so as to preserve the similarities between stimuli as well as possible.
Each stimulus is automatically given a location in the MDS space: the idea is that stimuli with high similarity are located close to each other, and stimuli with low similarity are located far apart.

Question 16

What is low-level temporal structure?

Answer 16

For a pitched musical sound, these fluctuations happen very fast: for example, a sine wave corresponding to middle C will repeat itself approximately 262 times every second.

Question 17

What do we mean when we talk about temporal aspects of timbre perception?

Answer 17

Interested at temporal dynamics at longer time scales shown in the ‘envelope’ within which the waveform oscillates.

Question 18

What is the ADSR model?

Answer 18

An attempt to capture the key components in which these different envelopes vary.

Question 19

What are the four portions that the ADSR model splits the envelope into?

Answer 19

1. The attack portion - where the envelope rises from zero amplitude to the maximum amplitude
2. The decay portion - where the envelope decays from its maximum amplitude to a lower amplitude, over a time period corresponding to the decay parameter.
3. The sustain portion - the amplitude remain constant, at a level determined by the sustain parameter.
4. The release portion - the amplitude decay to zero over a time period specified by the release parameter.

Question 20

What is a negative to ADSR model?

Answer 20

Is a big simplification compares to the real instatement envelopes

Question 21

What do the temporal aspects timbre concern?

Answer 21

the amplitudes of these different frequency components

Question 22

What does a spectral centroid describe?

Answer 22

How much of the spectral energy is concentrated in higher rather than lower harmonics. A high centroid tends to make the sound appear bright and piercing

Question 23

What is spectral irregularity?

Answer 23

Discrepancies in amplitudes between adjacent harmonics

Question 24

When is sound perceived as nasal?

Answer 24

When it has a lot of energy in the region spanning from 2000 to 5000 Hz.

Question 25

What are formants?

Answer 25

An aspect of timbre that determine vowel sounds in speech. A formant corresponds to a peak in the acoustic spectrum located somewhere above the fundamental frequency. The first formants are particularly important for determining vowel identity.

Question 26

What are formant frequencies determine as in human vocalisations?

Answer 26

The resonances of the vocal tract, which the speaker manipulates in large part through changing the shape and position of the tongue

Question 27

What is a vocoder?

Answer 27

A technology that enables musicians to incorporate formants (and other speech sounds) into their musical performances in real time. As the performer vocalises into the microphone, a computer extracts the spectral characteristics of their vocalisation and uses it as the timbre for a more traditional musical instrument, for example a keyboard or a guitar.