v3.0

The first stable release of the add-on.
Full Changelog: v3.0-beta.3...v3.0

v3.0-beta.3

Bug fixes

• Small bug fixes #46 #42

v3.0-beta.2

What's new

• Fixed some bugs related to voice variants
• Fixed voice manager shows invalid options for some voices

If you face any issues updating to this version, follow these steps:

1. Open the task manager and kill any tasks with the name sonata-grpc
2. Go to NVDA's addons directory, open the sonata_neural_voices directory, and remove the file named installTasks.py
3. Try updating using the usual way by clicking the add-on file or from within the add-on store

v3.0-beta.1

What's new

• NVDA 2024.1 compatibility
• Changed the name to Sonata since we plan to support additional TTS models besides Piper in the future.
• Support for fast variants for Piper voices. These fast variants improves responsiveness significantly
• Speaker and variant lists are now updated when changing voices from within NVDA's Speech Settings GUI
• Improvements to responsiveness and speed across the board
• The TTS server is now built as a single, statically linked executable

Important notice!!!

• After installing this version, you will lose all of your installed voices. Please use the voice manager to re-install the voices again.

v2.0-beta2

More responsive than ever

v2.0 of Piper-for-NVDA introduces a separate process for the TTS. This significantly increases responsiveness.

Noticeable improvements and fixes

• Installing voices is now easier using the integrated voice manager, which allows you to preview and download available voices.
• The TTS no longer introduces unnatural pauses at the end of the line during say all
• You can now configure synthesis parameters, such as length scale and noise scale
• Fixed some bugs related to voice switching
• Support for diacritizing Arabic text

What's new in v2.0 beta2

• Fixed the issue where sentences are broken by index command
• Upgraded onnxruntime to v1.16.1
• Fixed some bugs with voice preview
• Installing voices from local files no longer requires a specific file name format
• Use any free port for the grpc server rather than a fixed port.
• Some house keeping work

Full Changelog: v2.0-beta...v2.0-beta2

v2.0-beta

What's Changed

• Separated TTS process from NVDA. GRPC is used to communicate between the TTS process and the add-on. This provides significant performance improvement.
• Restore local installation by @rmcpantoja in #17

v1.0-beta2

New features

• Moved the core logic from Python to Rust. This provides significant speedup in synthesis, and hence responsiveness
• Implemented rate boost functionality up to 5x
• Added pitch adjustment
• Implemented speaker selection for multi-speaker voices
• Added the ability to remove installed voices

Known issues

• During say all, the synthesizer speaks each new line as if it is a separate sentence
• Some voices could not be installed. For workarounds please see this comment
• Some lower quality voices may skip syllables, especially at sentence boundaries

Full Changelog: v1.0-alpha2...v1.0-beta2

v1.0-alpha2

Updated the readme.

v1.0-alpha

Fixed minor bugs related to cancel and pause.