How the Player Works
This page explains what visitors see and how to use the TTS player.
What the visitor sees
When a page has TTS enabled, a player bar appears with:
- A play/pause button
- A progress scrubber showing elapsed and total time
- A speed selector (0.5x to 2.0x)
- An optional download button (if enabled in settings)
The player also announces state changes to screen readers via ARIA live regions.
First play
Clicking play for the first time on a page without cached audio triggers background generation. The player shows "Generating speech..." and polls the server every 3 seconds. Once the audio is ready, playback starts automatically.
Pages with cached audio begin playing immediately.
Resuming playback
The player saves your position to the browser's localStorage every 5 seconds. When you return to the same page, the player shows a "Resume from X:XX?" link. Clicking it jumps to where you left off. Progress is cleared when playback reaches the end.
Keyboard shortcuts
The player supports standard keyboard interaction:
| Key | Action |
|---|---|
| Space / Enter | Play or pause |
| Arrow left / right | Seek backward / forward |
| Tab | Move between player controls |
How text is extracted
The module never receives text from the browser. Instead, the server loads the entity and renders it, extracting plain text from the configured fields. This means computed fields, Views output, and block field content are all included in the audio. It also prevents content injection: the client sends only an entity reference, not the content itself.
Long content
Content over 2,000 characters is split at sentence boundaries. Each chunk is generated separately, then all chunks are concatenated into a single OGG file in one ffmpeg pass. The result is seamless playback regardless of content length.
Caching
Each audio file is cached with a key derived from the text hash, voice, speed, and language. When content is edited, the hash changes and the old audio is invalidated automatically. See Keeping audio current for details.