As per Mickaël's suggestion, I'll try to list the requirements for a read aloud navigator that I'm aware of, feel free to add something or correct me if I'm wrong.
Terminology
- I call read aloud navigator a component reading the content aloud, without any visual rendition. Requirements for interactions with visual navigators are listed too though.
- I call meaningful fragment of content a guided navigation node, a sentence or a chunk of sentence. Supporting words is not mandatory. In the case of a rendition using some prerecorded content (e.g media overlays), only guided navigation nodes are supported.
Prerecorded content
- Support discontinuities in playback: some chunks of an audio file can be skipped.
- Mix prerecorded content and text to speech.
Voices
- Use different voices for different content parts, up to a different voice for a foreign word.
- Support online voices.
- Change preferred voices at runtime.
Playback modes
- Continuously read the publication content.
- Automatically pause the playback between meaningful fragments of content.
- Set a silence between the playback of meaningful fragments of content. These silences are not taken into account in the content duration.
Playback enhancements
- Add silences in the range of 200ms-2s between nodes depending of their roles. Those silences are taken into account in the content duration.
- Change speed and pitch.
Navigation
- Inspect the content tree and the node roles.
- Jump to a specific node.
- Start from a node matching a given ref, from the beginning of the node or after skipping some text.
- Jump to the previous or next meaningful fragment of content.
Skippability and escapability
- Escape the current node if it's escapable.
- Skip the current node if it's skippable.
- Change the set of escapable and skippable roles at runtime.
- Automatically skip skippable nodes.
Interaction with visual rendering
- Highlight the meaningful fragment currently being played.
- Go to the current playback location in the visual rendition.
- Start playback from some location from the visual rendition.
As per Mickaël's suggestion, I'll try to list the requirements for a read aloud navigator that I'm aware of, feel free to add something or correct me if I'm wrong.
Terminology
Prerecorded content
Voices
Playback modes
Playback enhancements
Navigation
Skippability and escapability
Interaction with visual rendering