TTS-driven Synthetic Behavior Generation Model for Embodied Conversational Agents - Coverbal Synchrony in Human-Machine Interaction - page 345

Figure 6. The functioning of the inner fl uidity deque, using HRG structure.
phrases. In the case of CU-1, the NA syllable is ' ve ', and the starting and
ending points of the movement phase are determined by the syllables
' ve ' and ' dno '. In the case of CU-2, there is no word preceding the
prosodic gesture trigger that would also contain an NA syllable. As by
the aforementioned rule, this indicates that no preparation movement
phase is required. However, since both CU-1 and CU-2 represent
similar meaning (in terms of semiotics and gesture affi liation), the
preparation phase over the word ' in ' is artifi cially inserted into the
movement propagation scheme. Finally, the sixth process ' Align hold/
retraction ' aligns the hold (both pre- and post-stroke), and retraction
movement phases. The retraction movement phase is identifi ed by the
last meaningful word phrase by the word that contains the NA syllable
and precedes the B3 phrase break or precedes a longer pause. In Figure
6, the starting and ending points are NA syllable ' ra ' and syllable ' ze '
of the word 'obraze'. The hold movement phase is then determined by
the pauses ( sil units predicted and stored within the Syllable relation
layer), and by the residual of the meaningful words not yet aligned with
the stroke movement-phase. In addition to the sil units, the residual
of word phrase WP-1 is the unused syllable ' te ', and the residuals of
the word phrase WP-2 are the syllables ' is ', ' te ', ' ob '. As shown in the
Movement Phases relation layer within the HRG structure, the residual
content and the sil units are assigned as hold movement-phases. The
detailed movement structure, as generated by the inner fl uidity deque, is
presented in Figure 7. The movement phases, preparation (P), stroke (S),
and retraction (R), indicate a physical manifestation of a (new) shape,
whereas the hold (H) movement phase indicates the maintaining of
the closest previous physical manifestation.
As can be seen in Figure 7, the conversational expression CE-1
(gesture 1) is aligned with the prosodic boundaries (phrase breaks)
of the word phrase WP1 ('vedno iste' ( always the same )). During