A voice reported by a speech engine, such as an installed system voice or a downloaded model’s voice. Installation alone does not prove that it is enabled, active, or audible. See Installed voices and engines.
A presentation name such as default or bolden, resolved
through a voice palette to physical choices and adjustments. Editing a
shared named voice can affect several kinds of text that use it. See
Edit a named voice.
A collection defining named voices, including their physical choices, fallbacks, and adjustments. Palettes can inherit values; editing one is different from changing an engine preference. See Voices and routing.
A match between application meaning or context and the feedback to produce, such as a voice, cue, or silence. A rule chooses presentation; it does not install a physical voice. See Manage feedback rules.
A short sound conveying an event or state. A sound pack supplies cues, while presentation rules determine when to use them. Tones can convey information without a recorded sound file. See Sounds and space.
A saved selection of presentation resources and options. It refers to resources rather than copying everything into a full configuration backup. Applying a profile also retains its selection for startup; individual palette operations have different rules. See Presentation profiles.