Results Overview
4-class rates
(chance=25%)
Intervals
|
|
S1 |
S2* |
S3 |
S4 |
Mean |
Max |
|
Raw |
59.2 |
57.7 |
56.8 |
66.2 |
59.975 |
66.2 |
|
|
72.6 |
67.9 |
62.7 |
69.8 |
68.25 |
72.6 |
|
|
73.3 |
68.7 |
60.1 |
70.3 |
68.1 |
73.3 |
Melody
|
|
S1 |
S2 |
S3 |
S4* |
Mean |
Max |
|
Raw |
53.7 |
56 |
44.1 |
61.6 |
53.85 |
61.6 |
|
|
56.3 |
47.4 |
49.1 |
74.9 |
56.925 |
74.9 |
|
|
56.1 |
48 |
48.5 |
76.2 |
57.2 |
76.2 |
* Same
subject.
Note that the
subject who participated in both experiments achieved similar rates on both;
thus the overall difference in rates may have more to do with the subject pool
than with the stimuli.
Pair rates –
Intervals
|
|
Varied Dimensions |
S1 |
S2* |
S3 |
S4 |
Mean |
Max |
|
1 vs 2 |
Interval,
timbre, register, loudness |
96.5 |
97.5 |
89.5 |
93 |
94.125 |
97.5 |
|
2 vs 4 |
Register,
loudness |
95.5 |
97.5 |
85.2 |
91 |
92.3 |
97.5 |
|
2 vs 3 |
Interval,
timbre |
94.5 |
94.2 |
83.2 |
94 |
91.475 |
94.5 |
|
4 vs 1 |
Interval,
timbre |
87.5 |
76.7 |
77.3 |
85.2 |
81.675 |
87.5 |
|
1 vs 3 |
Register,
loudness |
77.8 |
72.7 |
71.5 |
84.7 |
76.675 |
84.7 |
|
3 vs 4 |
Interval,
timbre, register, loudness |
68.3 |
69.2 |
63.2 |
70.3 |
67.75 |
70.3 |
Melody
|
|
Varied Dimensions |
S1 |
S2 |
S3 |
S4* |
Mean |
Max |
|
1 vs 2 |
Interval,
timbre, register, loudness |
84.5 |
77.8 |
78.2 |
96 |
84.125 |
96 |
|
2 vs 4 |
Register,
loudness |
71.5 |
72.8 |
75.2 |
91.7 |
77.8 |
91.7 |
|
2 vs 3 |
Interval,
timbre |
71.8 |
70.2 |
74.3 |
86.5 |
75.7 |
86.5 |
|
4 vs 1 |
Interval,
timbre |
78 |
72.2 |
66.7 |
93.2 |
77.525 |
93.2 |
|
1 vs 3 |
Register,
loudness |
74 |
61.2 |
59.5 |
82.3 |
69.25 |
82.3 |
|
3 vs 4 |
Interval,
timbre, register, loudness |
61 |
63.8 |
59.8 |
75 |
64.9 |
75 |
Note that the
bottom-most pair rate (3 vs 4) for both experiments is worse than the 4-class
rate for the intervals experiment.
4-class and Pair Rates
by Subject (raw,
Confusion Matrix for
4-class
Intervals (averaged
|
|
Stim 1 |
Stim 2 |
Stim 3 |
Stim 4 |
|
Stim 1 |
204.75 |
8.5 |
50.5 |
36.25 |
|
Stim 2 |
7 |
269.25 |
10.75 |
13 |
|
Stim 3 |
53.75 |
15.5 |
161.25 |
69.5 |
|
Stim 4 |
29.5 |
15.75 |
72.75 |
182 |
Melody S4 [for better comparison to
Intervals rates]
|
|
Stim 1 |
Stim 2 |
Stim 3 |
Stim 4 |
|
Stim 1 |
248 |
2 |
40 |
10 |
|
Stim 2 |
5 |
263 |
20 |
12 |
|
Stim 3 |
43 |
19 |
182 |
56 |
|
Stim 4 |
4 |
9 |
66 |
221 |
Note the considerable
confusion between 3 and 4 (as in the pair rates). There is also some confusion between stimuli
1 and 3, which share interval (or melody), as well as the plucked-string timbre.
Conditional
Probability Matrix for 4-class
Intervals (averaged
|
|
Stim 1 |
Stim 2 |
Stim 3 |
Stim 4 |
|
Stim 1 |
0.6825 |
0.02833 |
0.16833 |
0.1208 |
|
Stim 2 |
0.02333 |
0.8975 |
0.03583 |
0.0433 |
|
Stim 3 |
0.1792 |
0.05167 |
0.5375 |
0.2317 |
|
Stim 4 |
0.09833 |
0.0525 |
0.2425 |
0.6067 |
Melody S4 [for better comparison to
Intervals rates]
|
|
Stim 1 |
Stim 2 |
Stim 3 |
Stim 4 |
|
Stim 1 |
0.8267 |
0.0067 |
0.1333 |
0.0333 |
|
Stim 2 |
0.0167 |
0.8767 |
0.0667 |
0.04 |
|
Stim 3 |
0.1433 |
0.0633 |
0.6067 |
0.1867 |
|
Stim 4 |
0.0133 |
0.03 |
0.22 |
0.7367 |
Confusion and
Conditional Probability Matrices by Subject
Upon brief
inspection, the
Since
Additionally,
detailed source-by-source