A§E
Intonation Tests
A§E Prosody
The A§E prosodical tests are based on pseudo-words and pseudo-sentences.
They aim to assess the coding of the temporal finestructure of
complex sounds. They are constructed in such a way that the fundamental frequency
F0 is 200 Hz.
They are used in a discrimination (pseudo-sentences) and identification (pseudo-words)
task.
Intonation is imposed as a spectral glide of the fundamental frequency from
F0=200 Hz to F0 = (200+∆)Hz.
An adaptive algorithm is used to seek the just noticeable difference, i.e. the smallest
∆ that is perceived by the listener.
Low-pass filtering is possible to rule out all spectral information above 300 Hz.
Sentence Intonation:
The sentence intonation test is a discrimination task to assess the temporal coding
of sound.
A pseudo-sentence is presented twice, once with a flat ending (∆ F0 is 0 Hz) and
once with a ∆ F0 between 0 and 208 hertz. The listener is asked to respond whether
both are equal or different. An adaptive algorithm changes ∆ F0 in function of the
listener’s response. It converges to the smallest ∆ F0 that is still perceived by
the listener. This is the just noticeable difference or JND.
The graph shows the responses and how the algorithm converges to threshold. The
cumulative responses for each ∆ can be followed on the histogram.
When viewing the report, double click the graph to show the histogram.
Word stress Pattern:
The word stress pattern test is an identification task to assess the temporal coding
of sound.
The listener has been familiarized with the three different stress positions during
the training. It is a closed set. After each presentation the listener should identify
the syllable containing the accent. He or she will typically respond one, two, or
three, corresponding to the first, second or third syllable containing the accent.
An adaptive algorithm changes ∆ F0, in function of the listener’s response. It converges
to the smallest ∆ F0 that is still perceived by the listener. This is the just noticeable
difference or JND. After a correct answer ∆ becomes smaller and vice versa. Internal
controls and stochastic processes improve the test efficacy and accuracy.
The graph shows the responses and how the algorithm converges to threshold. The
cumulative responses for each ∆ can be followed on the histogram.
When viewing the report, double click the graph to show the histogram.
Both tests have been developed by a European consortium funded by a 7th Framework
SME-grant (7th FP-SME 222291 Dual-Pro) from the European Commission.
Disclaimer: The present reflects only Otoconsult's view and the European Community
is not liable for any use that may be made of the information contained herein.
A§E Synthetics
The A§E synthetic tests are based on synthesized complex sounds
(harmonic complexes, harmonic intonating sounds and disharmonic intonating sounds).
They aim to assess the coding of the temporal finestructure of
complex sounds. They consist of a fundamental frequency, three harmonics and a low
level background noise. The reference fundamental frequency F0 is 200 Hz.
The sounds are used in a discrimination task.
Spectral differences are imposed at the level of the fundamental frequency. This
is either stationary in the sense that a second sound is presented with a fundamental
frequency F0 = (200+∆)Hz , or intonational in the sense that the second sound shows
a spectral glide of the fundamental frequency from F0=200 Hz to F0 = (200+∆)Hz.
An adaptive algorithm is used to seek the just noticeable difference, i.e. the smallest
∆ that is perceived by the listener.
Normative data intonation tests
(based on 40 dutch speaking hearing subjects validated in 30 romanian and 30 italian
speaking hearing subjects)
