Speech recognition not working even after installing a language pack

[Pfeil]

If your version of Windows did not originally come with a SAPI speech recognition engine installed, but you have installed a language pack that comes with a speech recognition engine (or installed the components for an existing language pack; See below for instructions on doing so), speech recognition will likely not work if VoiceAttack was started for the first time before the speech recognition engine was installed.

To remedy this, make sure that the "Use Built-In SAPI Speech Engines" option is enabled on the "System / Advanced" tab of the VoiceAttack options window, instead of the "Use Installed Speech Platform Speech Engines" option, which should allow speech recognition to operate once you click the "OK" button.


Note that the availability of a speech recognition engine is denoted by a microphone icon*, as seen under the "Preferred languages" heading of the "Language" subsection of the "Time & Language" section of the settings app (or your regionalized equivalent).

If a microphone icon is shown, but a speech recognition engine is not listed in the "Speech Engine" dropdown on the "Recognition" tab of the VoiceAttack options window despite the "Use Built-In SAPI Speech Engines" option being chosen on the "System / Advanced" tab of the VoiceAttack options window (or if that option is unavailable), click the entry for your language pack under the "Preferred languages" heading, and choose "Options", then verify that "Settings" is listed under the "Speech" heading, rather than a download button (otherwise, click said button to install the components).


Note that downloading speech recognition components may not be available, either when installing language packs or afterwards, if the network connection you are using is configured as a metered connection in Windows


*Note that there are some languages that are shown as having speech recognition available (I.E. they have a microphone icon), but may not come with a working SAPI speech recognition engine, E.G.:
Dansk / Danish
English (Australia)
English (Canada)
English (India)
English (Ireland)
English (New Zealand)
Français (Canada) / French (Canada)
Italiano (Italia) / Italian (Italy)
Korean
Português (Brasil) / Portuguese (Brazil)
Português (Portugal) / Portuguese (Portugal)
Español (México) / Spanish (Mexico)

A working speech recognition engine would be around 40 MB; if the download size for a speech recognition engine is shown are being much less than that, it will likely not be functional (E.G. "English (Australia)" is only 4 MB)