Information Technology Reference
In-Depth Information
Fig. 12.12
The second measure for the example analysis
Fig. 12.13
The third measure for the example analysis
dir ¼ 0
;
1
; þ
1
;
1
f
g
int ¼ 0
;
5
;
2
; f g
dur ¼ 120
;
120
;
120
;
120
f
g
x
¼
1
=
1or1
:
0
Then, the system creates FT2 with musicodes extracted from the second measure
of the training sequence (Fig.
12.12
) as follows:
dir ¼ 0
; þ
1
; þ
1
;
1
;
1
f
g
int
¼
0
;
1
;
1
;
1
; f g
dur ¼ 60
;
60
;
120
;
60
;
60
f
g
x
¼ 1
=
2or0
:
5
Next, MusEng creates FT3, with musicodes from the third measure of the
training sequence (Fig.
12.13
) as follows:
dir ¼ 0
; þ
1
;
0
f
g
int ¼ 0
;
1
; f g
dur ¼ 120
;
120
;
240
f
g
x
¼
1
=
3or0
:
33
fifth measures are processes next, but MusEng does not create
new FTs in these cases because they are repetitions of previous measures; that is,
their respective musicodes have already been seen by the system. In this case, only
the values of
The fourth and
for the respective FTs are adjusted accordingly. Thus, at this point
of the training phase, the
ω
values for each FT are as shown in Table
12.4
.
ω
is memory after the training phase, complete with 3 FTs, is shown in
Table
12.5
. It is important to stress that particular FTs gain or lose perceptual
MusEng
'
Table 12.4
Values of
after three FTs have been created and stored in memory, calculated by
dividing the number of instances of a given FT by the total number of FTs analysed
FT1
ω
FT2
FT3
1/5 = 0.2
2/5 = 0.4
2/5 = 0.4
ω
