|
Experiment Types
ECoS allows you to set up and carry out both deterministic and adaptive listening
tests.
Deterministic Tests
In deterministic tests, the sequence in which stimuli are presented across trials
is fixed at the start of testing. Features of deterministic experiments include :
- Stimuli are presented in a prespecified sequence -- either randomized or fixed
- Stimuli may be presented at a fixed level or at a number of specified levels
- Each test trial may contain from one to as many as six stimuli. Tests may require
identification, discrimination or comparison of signals.
Adaptive Tests
In adaptive tests, the sequence in which stimuli are presented across trials is
adjusted dynamically to reflect the pattern of the listener's responses. As a
result, different listeners hear a different sequence of signals -- these are
adjusted to be most relevant to the given listener.
The adaptive algorithm may be set to converge on 30%, 50% or 70% correct responses
using Levitt's algorithm.
Control parameters are user-specified, so that results can be obtained with specified
speed and/or accuracy. Control parameters include the total number of reversals
in the test, the number of reversals to be used in the estimate, the size of steps
(adjustments) to be used in up to three phases of adaptation (coarse, medium and
fine), and the number of reversals to occur in each phase. Adaptation normally
occurs over file names, using a set of sequentially-named audio files, ordered
with respect to the parameter of interest. Users with TDT System II hardware including
PA4 Series attenuators can also define adaptive intensity experiments.
For Adaptive Experiments, additional forms are provided so you can indicate the
desired details of the adaptive test. These parameters control the speed of the
test, and the accuracy of the final results of the test.
Order ECoS/Win | Download
ECoS/Win Demo |
