Start with the result you want, not with the name of an editor. A voice edit changes how a shared voice sounds. A rule changes which items receive that voice or other feedback. Output settings control delivery. The distinctions below help avoid making a wider change than intended.
Edit the active palette’s default named voice to change normal text and
labels that do not request an explicit voice. Other entries, such as the
voice used for headings, keep their own settings. Choose this route for a
more comfortable everyday reading voice, not for one exceptional heading.
See Edit a named voice.
Use a presentation rule when the change belongs to matching headings, prompts, message fields or events. Inspect a representative item, choose the criteria, and decide how long the rule should last. A rule for the current buffer can match several items there; it is not automatically limited to the one under point. The heading exercise is a reversible starting point.
Use voice mapping to choose a different named voice for those items. Use a guided feedback change to add, remove or alter a voice, sound, tone or spoken label. Neither operation requires retuning the shared named voice.
Edit a named voice when its other uses in the palette should change too.
For example, adjusting bolden affects items that resolve to that entry,
not only the item from which the editor was opened. Inspect its known uses
before saving. For a change limited to the original context, use a rule instead.
See Edit a named voice and voice adjustments in context.
Use the palette-wide review when you want a physical voice to be preferred
across all its effective named entries while retaining their different
adjustments. Review the proposed result for several entries before saving.
This is a one-time edit, not a policy automatically inherited by future names.
It is broader than editing default.
Use engine preference when you want to change the shared engine order. It does not override explicit physical choices already stored in named voices. If one item still uses another engine, inspect that voice’s physical choices before changing the engine order again. Voice availability and activation are separate from preference: an installed or allowed voice is not necessarily ready to play.
For reading speed or punctuation detail, use Speech settings. For the place where notifications play, use notification routing; for their loudness, use output volumes. These controls do not decide which text an application speaks automatically. Its focused speech settings control that behaviour.
For one unwanted cue, follow Silence one cue instead of disabling all sound. To change the recordings used by a family of cues, choose a sound pack.
After choosing the operation, decide whether to keep a draft, use the result now or retain it for startup. Choose what persists explains those choices; the scope shown in a menu does not make every Save action equivalent.