Configurable registry values

Following are all configurable registry values used by this program, documented for reference. Some of them can be changed via the installer UI, others cannot (yet) but can be changed with the registry editor.

These registry values might be changed or removed in future versions. Also, some of the features below are "experimental", so you should only change a value when you are sure what can happen.

Values marked as "Boolean" types can be either true (any non-zero value), or false (zero). Boolean values default to false (zero), unless otherwise stated.

Values in HKEY_CURRENT_USER\Software\NaturalVoiceSAPIAdapter

Value Name	Type	Description
LogLevel	REG_DWORD	Sets the log level. Logs below this level will be ignored. 0: Trace; 1: Debug; 2: Info (default); 3: Warning; 4: Error; 5: Critical; 6: Off. Note that logs in Trace and Debug levels may contain personal information.
LogFlushLevel	REG_DWORD	Flushes the log file when a log at or above this level appears. Default is 3 (Warning).
LogFlushInterval	REG_DWORD	Flushes the log file at specific intervals (milliseconds). Default is 0 (disabled).
LogFile	REG_SZ	Overwrites the default path for the log file (%LOCALAPPDATA%\NaturalVoiceSAPIAdapter\log.#.txt).
EnableWebsocketppLogs	REG_DWORD (boolean)	Enables Websocketpp logs. Disabled by default.
LogToConsole	REG_DWORD (boolean)	Opens a console window to display the logs. Note that this might break some applications.
EnableCrashDump	REG_DWORD (boolean)	Catches unhandled exception and generate a mini dump file when an unhandled exception occurs. The dump file will be in %LOCALAPPDATA%\NaturalVoiceSAPIAdapter\minidump.dmp.
CrashDumpType	REG_DWORD	What kinds of information will be included in the dump file (the DumpType parameter of MiniDumpWriteDump). Default is MiniDumpWithIndirectlyReferencedMemory MiniDumpWithDataSegs MiniDumpFilterModulePaths.
ForceEnableAzureSpeechSDK	REG_DWORD (boolean)	Forces the Azure Speech SDK (and related features such as local Narrator voices) to be enabled on unsupported systems (earlier than Windows 7).
UseAzureSpeechSDKForAzureVoices	REG_DWORD (boolean)	Uses the Azure Speech SDK, instead of the implementation in this engine, when using an Azure voice.

Values in HKEY_CURRENT_USER\Software\NaturalVoiceSAPIAdapter\Enumerator

Value Name	Type	Description
Disable	REG_DWORD (boolean)	Disables the enumerator. Nothing will be enumerated.
NoNarratorVoices	REG_DWORD (boolean)	Disables local Narrator voices, including installed voices and voices in NarratorVoicePath.
NarratorVoicePath	REG_SZ	The path to a folder whose subfolders contain unzipped Narrator voice models. Defaults to <this program's folder>\NarratorVoices.
NoEdgeVoices	REG_DWORD (boolean)	Disables Microsoft Edge online voices.
NoAzureVoices	REG_DWORD (boolean)	Disables Azure voices. You should also have AzureVoiceKey and AzureVoiceRegion to enable Azure voices.
EdgeVoiceAllLanguages	REG_DWORD (boolean)	Whether to include all supported languages for online (Edge & Azure) voices.
EdgeVoiceLanguages	REG_MULTI_SZ	The list of the included languages for online (Edge & Azure) voices, separated into substrings. Uses IETF BCP 47 language tags, e.g. en-US for American English, en for all English accents/dialects. Defaults to the user's preferred/display languages when EdgeVoiceAllLanguages is false.
AzureVoiceKey	REG_SZ	An API key for accessing the Azure Speech service, in clear text.
AzureVoiceRegion	REG_SZ	The region for AzureVoiceKey.
EdgeVoicesOverrideAzureVoices	REG_DWORD (boolean)	When there are an Edge voice and an Azure voice with the same name (e.g. Jenny), the Azure voice will be hidden from the list.
DefaultErrorMode	REG_DWORD	Defaulted to 0. When set to 2, a message box will be shown when an error occurs. By default, errors will be written to the log, and reported to the client application if possible.
LanguageForUnknownLocales	REG_SZ	The language ID(s) provided to SAPI for voices in language without a language ID. For example, jv-ID (Javanese, Indonesia) and su-ID (Sundanese, Indonesia) do not have language IDs even on Windows 11, but SAPI requires each voice to have a language ID. By default, voices in those languages will be disabled. If you want to use them, you can set LanguageForUnknownLocales to a hexadecimal string (without 0x prefix) of a language ID, such as 409 to pretend that the voices are American English voices. This may work or not, depending on the client application.