Neural Voice Manager

The Neural Voice Manager add-on provides a convenient interface for managing Microsoft neural voices available for the NVDA screen reader. With it, you can browse, install, remove, and download voice packages grouped by language and country.

Note: To use the voices, you must install the Neural Voices driver.

Purpose of the Add-on

Microsoft neural voices provide the most natural and high-quality synthesized speech. This add-on allows you to centrally manage these voices without the need for manual searching and installation of files. All voices are structured by language groups and regional variants, making it easier to choose the right voice for your needs.

Key Features

• Browse available Microsoft neural voices with filtering by language and country.
• Install selected voices with automatic downloading and extraction of necessary files.
• Remove installed voices.
• Download voice files to your computer for later use or backup.
• Display the status of each voice: installed, will be installed after restart, will be removed after restart, or not installed.
• Work with local voice data cache when offline.
• Flexible table column display settings with the ability to show or hide desired columns.

Add-on Interface

The add-on is integrated into the "Tools" menu of the NVDA system tray. To open the management dialog, follow these steps:

1. Press NVDA+N to open the NVDA menu.
2. Select "Tools".
3. Select "Neural Voice Manager" (To quickly navigate to this item, press "M" in English or "Н" in Russian interface).

Once the dialog opens, you will have access to the following controls.

Language and Country Selection

At the top of the window, there are two drop-down lists: "Language" and "Country". They allow you to filter voices by language and regional affiliation. When you select a language, the list of available countries updates automatically, and when you select a country, the voice list updates.

Voice List

The central part of the window displays a table with the following columns:

• Name: The display name of the voice.
• Gender: Male or female voice.
• Age: Age category of the voice.
• Type: The type of voice package (Neural, RUS, HD, and others). Detailed descriptions of types are provided in the "Voice Type Explanations" section.
• Status: Current status of the voice (installed, not installed, will be installed after restart, or will be removed after restart).
• Version: The version of the voice package.
• Size: The voice file size in a human-readable format (KB, MB, GB).

To select a voice, click on it with the left mouse button or use the list navigation keys.

Column Display Management (View)

For convenience and to reduce information overload, you can customize the voice table display. This feature is available through the context menu:

1. Right-click on any voice in the list.
2. In the context menu that appears, select the "View" submenu.
3. The submenu displays items for each column (except "Name" and "Status", which are always visible):
   • Gender
   • Age
   • Type
   • Version
   • Size
4. Each item has a check mark indicating whether the column is visible.
5. To hide or show a column, simply select the corresponding menu item — the change takes effect immediately.

All column display settings are automatically saved and applied when you reopen the dialog, even after restarting NVDA.

Control Buttons

Below the voice list, there are three main buttons:

• Install: Downloads and installs the selected voice. Available only for voices that are not yet installed.
• Remove: Removes the installed voice. Available only for installed voices.
• Download to PC: Saves the voice file to your computer's "Downloads" folder without installing it. Available for all voices.

Context Menu

When you right-click on a selected voice, a context menu opens with the same commands: "Install", "Remove", or "Download to PC", as well as the "View" submenu for adjusting column display.

Voice Installation Process

When you click the "Install" button, the voice file download process starts from the server. Progress is shown in a progress window. The installation time depends on the file size and your internet connection speed. After the download completes, the voice is automatically extracted to a special folder. To complete the installation, you need to restart NVDA. After the restart, the voice becomes available in the speech synthesizer settings.

It is important to note that installed voices are preserved even when updating NVDA, unless you remove them manually.

Voice Removal Process

When you click the "Remove" button, a confirmation dialog appears. After confirmation, the voice is marked for removal. To complete the process, you also need to restart NVDA. After the restart, the voice will be completely removed from the system.

Downloading a Voice to Your Computer

The "Download to PC" function allows you to save the voice file to your system's "Downloads" folder. This is useful in the following cases:

• You want to save the voice for later installation without an internet connection.
• You plan to transfer the voice to another computer.
• You want to create a backup of the voice file.

A progress window is displayed during the download. After the download completes, you will receive a notification with the path to the saved file.

Voice Type Explanations

The "Type" column displays various designations that characterize the features of voice packages. Below are descriptions of the most common types:

• Neural — neural voices. They use deep learning technologies for the most natural and realistic speech quality.
• HD (High Definition) — high-quality voices. They feature enhanced clarity and sound detail compared to standard voices.
• RUS (Restricted Use Subscription) — a Microsoft software licensing format that provides access to a predefined set of features or language models (voices) for a fixed fee, but without the right to further distribute such a product.

Please note that the list of types may expand as new voice packages become available.

Working with Data Cache

The add-on saves information about available voices in a local cache. This allows you to work with the add-on even when offline, provided the data was previously loaded. The cache updates automatically each time you open the dialog if the data is outdated or missing.

Behavior When Offline

If there is no internet connection, the add-on attempts to load voice data from the local cache. If the cache is missing or corrupted, you will see a message about the need to connect to the internet.

Usage Notes

• The add-on requires a stable internet connection when first launched or when updating voice data.
• Voice installation may take significant time for large files. Do not interrupt the installation process.
• After installing or removing a voice, you must restart NVDA for the changes to take effect.
• The add-on automatically removes temporary files after installation completes.
• All voices installed through the add-on appear in your speech synthesizer's voice list.
• Column display settings are saved in the settings.json file in the add-on configuration folder and are applied each time the dialog is opened.

Contacts

If any voice fails to download while others work, please report it through the feedback form on the main community website:

Go to Feedback Form You can also report it to the community leader in the community Telegram group: https://t.me/nvda_group

Note: If you discover new Microsoft Neural voices, please also report them and send the voice files.