Paradigms

Todo

Give better description of the available paradigms.

Available Paradigms

Transformed Up-Down

This paradigm implements the transformed up-down adaptive procedures described by [Levitt1971]. It can be used with \(n\)-intervals, \(n\)-alternatives forced choice tasks, in which \(n-1\) “standard” stimuli and a single “comparison” stimulus are presented, each in a different temporal interval. The order of the intervals is randomized from trial to trial. The “comparison” stimulus usually differs from the “standard” stimuli for a single characteristic (e.g. pitch or loudness), and the listener has to tell in which temporal interval it was presented. A classical example is the 2-intervals 2-alternatives forced-choice task. Tasks that present a reference stimulus in the first interval, and therefore have \(n\) intervals and \(n-1\) alternatives are also supported (see [GrimaultEtAl2002] for an example of such tasks)

Transformed Up-Down Interleaved

This paradigm implements the interleaved transformed up-down procedure described by [Jesteadt1980] .

Weighted Up-Down

This paradigm implements the weighted up-down adaptive procedure described by [Kaernbach1991].

Weighted Up-Down Interleaved

This paradigm combines the interleaved procedure described by [Jesteadt1980] with the weighted up-down method described by [Kaernbach1991].

Constant m-Intervals n-Alternatives

This paradigm implements a constant difference method for forced choice tasks with \(m\)-intervals and \(n\)-alternatives. For example, it can be used for running a 2-intervals, 2-alternatives forced-choice frequency-discrimination task with a constant difference between the stimuli in the standard and comparison intervals.

Constant 1-Interval 2-Alternatives

This paradigm implements a constant difference method for tasks with a single observation interval and two response alternatives, such as the “Yes/No” signal detection task.

Constant 1-Pair Same/Different

This paradigm implements a constant difference method for “same/different” tasks with a single pair of stimuli to compare.

Multiple Constants 1-Pair Same/Different

This paradigm implements a constant difference method for “same/different” tasks with multiple pairs of stimuli to compare.

Multiple Constants ABX

This paradigm implements a constant difference method for “ABX” tasks with multiple pairs of stimuli to compare.

Odd One Out

This paradigm implements a three-alternatives oddity procedure (see [VersfeldEtAl1996]).

PEST

This paradigm implements the PEST adaptive procedure described by [TaylorAndCreelman1967]. However, beware that support for this procedure in pychoacoustics is very experimental. Its implementation has received very little testing.

PSI

This paradigm implements the PSI+ and PSI-marginal adaptive procedures described by [Prins2013].

UML

This paradigm implements the updated maximum likelihood (UML) adaptive procedure described by [ShenAndRichards2012].