Datasheet

3 - 7
USER'S GUIDE
The length of a sound depends on many factors: the intrinsic length of the
phoneme - the length of the /oo/ sound in "book" and "groovy" is very different -
, the stress - stressed syllables are higher in pitch and longer - and some other
factors.
Important syntactic borders are marked by a higher intonation, sounds are
longer just before a syntactic border. The type of sentence also plays a role: it
makes a substantial difference whether a sentence ends on a question mark, an
exclamation mark or point as in declarative sentences.
Additionally, the characteristics of spoken words are identified, providing sepa-
rate pronunciations for “rough”, “cough”, “dough”, “plough”, “though”, “thought”,
“through” and “thorough”.
Turning Speech Patterns into Sound Waves
The last phase of acoustic processing takes care of the actual speech pro-
duction: the processed speech patterns from the phonetic phase are converted
into synthetic waves which are reproduced by the PC’s audio card.
EXPERIMENT TO LEARN!
To get a better idea of the highly advanced nature of the speech feature, an
I.R.I.S. test sheet is provided with the IRISPen.
It allows you to experiment with the text-to-speech capability of your IRISPen
Executive and get aware of the linguistic rules you should respect when you want
to correctly read telephone numbers, bank account numbers etc.
LEARNING MORE
These syntax rules are discussed at length in Appendix B. They describe
how the speech synthesis deals with normalized character strings such as bank
account and telephone numbers, date and time indications etc.
4chapter3.p65 2/27/2001, 8:47 AM7