BATS - Text2Speech

Also available in:

Deutsch

Text turned into speech in seconds – Kokoro on OpenVINO, straight into an EDIUS tab

Text2Speech is the fast route from idea to spoken track: via a menu entry in EDIUS you open a Text2Speech tab in the player, type or paste text, pick language and voice, audition the output – and the finished WAV file lands directly in the BIN. It is driven by Kokoro on the OpenVINO framework (Intel NPU/GPU or CPU) – with a smart cache (repeated text is loaded from the cache instead of being computed again), a preview mode, simple input dialogs and an activity log.

Adding the menu entry

“Text2Speech” is added through the “Add function...” assistant (right-click in the BIN → Add function... → pick Text2Speech - fast audio speech generation). The assistant’s questions:

  1. Device (“Please select the device to create the subtitle.”): Intel NPU (fastest, on supported hardware) · Intel GPU · CPU (Intel/AMD).
  2. Placement: context menu of the BIN or of the timeline.
  3. Name of the menu entry (suggestion: “Add speech..”), then [Preview '.bat0' file] / [Add menu entry].

Note: BAT Server already ships the system preset “Add Speech....bat0” (BIN, device: AUTO) – run the assistant if you want your own names, a different placement or a fixed device.

How to use it

  1. Right-click an entry in the BIN (or on the timeline) → your Text2Speech entry (e.g. “Add speech..”).
  2. BAT Server prepares the TTS server (toast “Text2Speech server is being prepared for speed…”) and opens the Text2Speech tab in the EDIUS player – analogous to the Cut-by-Text tab.
  3. In the tab: enter text, pick language and voice, preview and save the WAV file – it is read directly into EDIUS (BIN). Repeated texts come from the smart cache and are ready almost instantly.

Good to know

  • 9 languages with their own voices – you see the per-language offerings directly in the tab.
  • EDIUS must be running: the tab opens in the EDIUS player, and the result is taken into the BIN via EDIUS.
  • Smart cache: identically phrased texts are not recomputed but loaded from the cache – for serial speech tracks (e.g. the same on-screen sentences) that saves noticeable time.
  • The device choice is baked in: NPU/GPU/CPU is in the .bat0; the shipped system preset uses AUTO (self-detecting) instead.

Managing the preset

Entries created via the assistant are user presets: the .bat0 file lives under C:\ProgramData\EdiusPowerTools\BAT-server\bat_files\BIN\ or …\bat_files\timeline\ depending on your choice, and can be freely edited, renamed or deleted (deleting removes the context-menu entry). As with all BATS presets, the .bat0 is a template from which BAT Server generates the concrete job file on every run – changes take effect from the next run. The shipped system preset “Add Speech....bat0” is outside this freedom (system presets are meant to be protected from accidental edits, but can be deactivated).