TI macro itemList

Tie the items in itemList and assign them to the specified macro name. This command applies to any item type but all of the items in itemList must be of the same type. The detailed method of tying depends on the item type as follows:

state(s)

the state with the largest total value of gConst in stream 1 (indicating broad variances) and the minimum number of defunct mixture weights (see MU command) is selected from the item list and all states are tied to this typical state.

transitions(t)

all transition matrices in the item list are tied to the first in the list.

mixture(m)

all mixture components in the item list are tied to the first in the list.

mean(u)

the average vector of all the mean vectors in the item list is calculated and all the means are tied to this average vector.

variance(v)

a vector is constructed for which each element is the maximum of the corresponding elements from the set of variance vectors to be tied. All of the variances are then tied to this maximum vector.

covariance(i)

all covariance matrices in the item list are tied to the first in the list.

xform(x)

all transform matrices in the item list are tied to the first in the list.

duration(d)

all duration vectors in the item list are tied to the first in the list.

stream weights(w)

all stream weight vectors in the item list are tied to the first in the list.

pdf(p)

as noted earlier, pdf's are tied to create tied mixture sets rather than to create a shared pdf. The procedure for tying pdf's is as follows

1. All mixtures from all pdf's in the item list are collected together in order of mixture weight.
2. If the number of mixtures exceeds the join size J [see the Join (JO) command above], then all but the first J mixtures are discarded.
3. If the number of mixtures is less than J, then the mixture with the largest weight is repeatedly split until there are exactly J mixture components. The split procedure used is the same as for the MixUp (MU) command described above.
4. All pdf's in the item list are made to share all J mixture components. The weight for each mixture is set proportional to the log likelihood of the mean vector of that mixture with respect to the original pdf.
5. Finally, all mixture weights below the floor set by the Join command are raised to the floor value and all of the mixture weights are renormalised.