Note
Please see Azure Cognitive Services for Speech documentation for the latest supported speech solutions.
Microsoft Speech Platform
Registry Settings
The Speech Platform defines a standard set of attributes for tokens in the registry, and applications can query for these attributes. This allows applications to find the most suitable token within any category on the computer.
Voices category
Applications can query the attributes of voice tokens to find and select the most suitable voice token on the computer. Voice tokens are located under the following key:
- HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Speech Server\v11.0\Voices\Tokens
See Enumerate and Inspect Tokens for more information about attributes and querying for them.
Table 1 lists and describes the registry entries that constitute a sample voice token called VoiceToken1, located under HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Speech Server\v11.0 Server\v11.0\Voices\Tokens.
RegKey | Entry | Comments |
---|---|---|
VoiceToken1 | The registry key for the Token. | |
(Default) | The language-independent name. | |
/Attributes | Attributes for the Token are under this key. | |
Age | The value may be Child, Teen, Adult, or Senior, depending on the age of the TTS Voice. Senior indicates an elderly voice. Vendors may choose to classify some voices as both Senior and Adult. | |
Gender | The gender of the voice. Value can be Male or Female. | |
Language | The identifier (in hex) of the language this token speaks. | |
VendorPreferred | The default installed voice for the vendor. |
Table 1. Registry Settings and Attributes for Voice Tokens: Local Machine Subtree
The Voices category in the HKEY_CURRENT_USER\Software\Microsoft\Speech Server\v11.0 subtree stores the default voice to use for speech synthesis operations. Table 2 provides a listing of the registry entries that constitute a voice token in HKEY_CURRENT_USER\Software\Microsoft\Speech Server\v11.0\Voices\.
Entry | Value | Comments | |
---|---|---|---|
VoiceToken1 | The registry key for the token. | ||
DefaultTokenID | HKEY_CURRENT_USER\Software\Microsoft\Speech Server\v11.0\Voices | This points to the default voice for speech synthesis operations. |
Table 2. Registry Settings for Voices: User Subtree
Note:
- The TTS engine does not need to store any of these values, the Speech Platform takes care of that.
- The DefaultTokenID for the Voices category (or other categories) can be set programmatically using ISpObjectTokenCategory::SetDefaultTokenId.
Recognizers category
Applications can query the attributes of recognizer tokens to find and select the most suitable recognizer token on the computer. The Speech Platform enumerates all the tokens for speech recognition installed on the computer under the following registry key:
- HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Speech Server\v11.0\Recognizers
Table 3 lists and describes the registry entries that constitute a sample recognizer token called RecognizerToken1 under HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Speech Server\v11.0\Recognizers\Tokens.
RegKey | Entry | Comments |
---|---|---|
RecognizerToken1 | This is the registry key for the token. | |
(Default) | The language-independent name | |
/Attributes | Attributes for the Token are under this key. | |
Language | The identifiers (in hex) of the language(s) this engine recognizes. Typically this is only one per recognizer token. A value of 409;9 indicates that the Recognizer recognizes generic English (American, British, Australian, and so on). | |
VendorPreferred | The default installed recognizer for the vendor. |
Table 3. Registry Settings and Attributes for a Recognizer Token: Local Machine Subtree
There is also a Recognizers category in the HKCUS subtree that stores the selected default Recognizer. This is done exactly as for Voices, as shown in Table 2. The CategoryID is HKEY_CURRENT_USER\Software\Microsoft\Speech Server\v11.0\Recognizers\.
Entry | Value | Comments | |
---|---|---|---|
RecognizerToken1 | The registry key for the token. | ||
DefaultTokenID | HKEY_CURRENT_USER\Software\Microsoft\Speech Server\v11.0\Voices | This points to the default voice for speech recognition operations. |
Table 4. Registry Settings for a Recognizer Token: User Subtree
Note: The DefaultTokenID for the Recognizers category (or other categories) can be set programmatically using ISpObjectTokenCategory::SetDefaultTokenId.