USC dyslexia project speech perception experiments (cont.) STILL IN PROGRESS!!
Here are the scripts, sound
files, and picture files for the experiments run in Matlab. Sound files
are in .wav format.
These materials are posted here for the convenience of project members; others please do not use without first asking either keating@humnet.ucla.edu or manis@usc.edu.
Zip
of some functions used by multiple scripts below.
Experiment 1 (spy-sky continuum identification and discrimination;
this is the same as the Psyscope version of spy-sky but with the addition of
discrimination of the long transition stimuli, and between-block reinforcements
“advance to next level”). Stimuli differ as short (45 ms) vs.
long (95 ms) transition durations and by F2 onset frequency (1100 to 1800 Hz).
zip of
whole thing including Matlab scripts, sound files, picture files
zip
of just picture files (.jpg): spysky, skyspy, congrats, level1, level2,
level3, level4, level5, level6
Experiment 2 (identification
in noise of endpoint synthetic short-transition spy-sky stimuli,
naturally-spoken spy and sky, and two environmental sounds, clapping and
stream; many SNRs have been made for piloting but any one script uses only a
few of them)
zip
of scripts for Nov. 28 version of experiment
zip
of all picture files (only clap and stream are new, others are same as
Exp1)
zip
of all sound files (table shows all combinations made and some sample
filenames, but files are available only all together; stimuli called
“clear” are originals normalized-equalized to nominal 80 dB)
|
S/N |
Natural spy or sky |
Synthetic spy or sky |
Environmental clap or stream |
|
CLEAR |
natsky-clear.WAV |
|
|
|
+20 |
|
|
envCLAP20.WAV |
|
+15 |
|
|
|
|
+10 |
|
|
|
|
+5 |
|
|
|
|
+4 |
|
|
|
|
+3 |
|
|
|
|
+2 |
|
|
|
|
+1 |
|
|
|
|
0 |
|
|
|
|
-1 |
|
synsky-1.WAV |
|
|
-2 |
|
|
|
|
-3 |
|
|
|
|
-4 |
|
|
|
|
-5 |
|
|
|
|
-6 |
|
|
|
|
-7 |
|
|
|
|
-8 |
|
|
|
Experiment 3 (draft) (tone detection in noise by
adaptive staircase procedure; there are no sound files because tone, noise, and
their combination are all generated by the script during the experiment; and no
picture files)
zip of script and functions (draft)
Experiment 4 (gating).
There are no picture files for this experiment.
Matlab
script
(includes a parameter for omitting short gates in practice and/or test items)
zip
of sound files (all gates)
Experiment 5 (draft) (target word recognition under
several variability conditions, final items not yet determined)
Matlab
script
zip
of all sound files (for all conditions, possibly some extras as well)
zip
of all picture files (screen displays used in running experiment)
Experiment 5a (draft) (subset of above with only
conditions 0 and 1, testing one vs. multiple speakers) Still to come:
audio hints about the identity of target words, to eliminate mis-hearings by
subjects.
This experiment's scripts (both versions) include 2 features intended to speed up RTs, which were far too long in initial pilot: variable ISI so that subjects cannot set up a rhythm of responses, and an on-screen display of a time bar which shows time remaining to an arbitrary deadline for response (which in fact is not enforced). Together these features do encourage subjects to respond quickly. There are no picture files for stimuli in this experiment.
Experiment 5sh
(extension of 5a that adds items comparing same
vs. different allophones by the same vs. different speakers). This
version includes an audio hint for each target word, and the on-screen time
bar. There are 2 versions of the experiment, 1 and 2, which
counterbalance the one-speaker vs. multi-speaker conditions. Each version
has its own script and file list, but uses the same sound and picture files.
Matlab script
1 and script
2
list1
and list
2
zip
of all sound files (including prompt files)
(zip of pictures files
is same as above though not all are used in this version)
zip
of the whole thing
Experiment 6 (segments vs features) (with the original script this
experiment failed in that the expected basic RT effect that it builds on was
not obtained - all RTs were far too long; see above for speeding up RT - this
is why there are 2 versions of the script; the second version has not yet been
tested)
original
Matlab script
Matlab
script
incorporating countdown bar
list.txt
file
created by P. Keating Dec. 2004, last updated Mar. 2005
Return
to the page of Psyscope experiments
Return to Pat Keating’s homepage