When you pass a string of characters to the Speech Manager via Speech Pack the string is converted into phoneme code. This code contains special characters to indicate how a word must be pronounced. The second step is to convert this phoneme code into allophone code and then into actual sound-data and create the audio output.
The first step of converting a string into phoneme control code takes a lot of memory and time and is also the most important step since the quality of the voice depends largely on the pronounciation. If you provide phoneme code directly this pronounciation is used and you can actually use phoneme code created by a better speech synthesizer with possibly a larger dictionary.
You can use the SP Voice code function to convert a string of characters into phoneme code and store this either in the structure file (as a ”STR#“ resource for use with the SP Voice String function), a field inside the data-file or your own documents.
