Inside Macintosh: Sound /

Chapter 4 - Speech Manager / About the Speech Manager

Speech Channels

To indicate to the Speech Manager which voice or attributes you would like it to use in generating speech, your application must use a speech channel. A speech channel is a data structure that the Speech Manager uses when processing text; it can be associated with a particular voice and particular speech attributes. Because multiple speech channels can coexist, your application can create several different vocal environments (to simulate a conversation, for example). Because a synthesizer can be associated with only one language and region, your application would need to create a separate speech channel to process each language in bilingual or multilingual text. (Currently, however, only English-producing synthesizers are available.)

Different speech channels can even generate speech simultaneously, subject to processor capabilities and Sound Manager limitations. This capability should be used with restraint, however, because it can be hard for the user to understand any speech when more than one channel is generating speech at a time. In general, your application should generate speech only at the specific request of the user and should allow the user to turn off speech output. At the very least, your application should include an option that allows the user to view text instead of hearing it. Some users might have trouble understanding speech generated by the Speech Manager, and others might have a hearing deficit. Even users who are able to clearly understand computer-synthesized speech might prefer to read rather than hear.

Using the Speech Manager, you can identify how many voices are available and sort through an index of the voices to get information about a specified voice such as its gender, age, or the synthesizer with which it is associated. In general, your application does not need to know which speech synthesizer it is using, and in most cases, you do not need to be concerned with which speech synthesizer a voice is associated. Sometimes, however, a speech synthesizer may provide special capabilities beyond that provided by the Speech Manager. For example, a speech synthesizer might allow you to select an option to speak numbers in a nonstandard way. The Speech Manager allows you to determine which synthesizer is associated with a voice for these circumstances and provides hooks that allow your application to take advantage of synthesizer-specific capabilities.

In general, your application can achieve the best results by not making assumptions about which synthesizers might be available. The user of a 2 MB Macintosh Classic might use a synthesizer with low RAM requirements, while the user of a 20 MB Macintosh Quadra 950 might take advantage of a synthesizer that provides better audio quality at the expense of memory usage. The Speech Manager makes it easy to accommodate both kinds of users. Currently there are three synthesizers available with the Speech Manager. Each synthesizer has its own RAM requirements. To be compatible with all three synthesizers, you must reserve enough space in your application's heap to accommodate their requirements. In general, reserving around 250 KB per channel that you anticipate using provides enough space for the MacinTalk Pro synthesizer.