SAPI Notes

I use Microsoft's Speech Application Programming Interface (SAPI) for both text-to-speech (TTS) as well as automatic speech recognition (ASR). There are various versions of SAPI along with various engines and voices. This page is for me to store notes about SAPI in general. I use SAPI v5.1 for XP which is becoming old and information is getting harder to find as URL's and sites change and with info getting confused with newer versions of SAPI.

Everything on this page is in refrence to version 5.1 unless noted.

SAPI Versions

  • v1 - v4: These old versions started in 1995 and was used till about 2000.
  • v5.0: Released in 2000, this was a complete redesign from previous versions and is the foundation for current verisons.
  • v5.1: Released in 2001, this is the version this page is about. It was shipped with Windows XP.
  • v5.2: Released in 2004, this is a special version for Microsoft Speech Server.
  • v5.3: This is the version shipped in Windows Vista. It can not be downloaded and installed in another OS.
  • v5.4: This is the version shipped in Windows 7. It can not be downloaded and installed in another OS.

Detailed information on SAPI and versions can be on on Wikipedia.

SAPI vs Engines

One of the important things to keep in mind about SAPI is that it is an API layer between an application and ASR & TTS engines. This allows an single app to work with multiple TTS & ASR with a single API. When you install SAPI v5.1 SDK, it comes with the Microsoft English Recognizer v5.1 ASR engine.

However, Microsoft released a much better ASR engine called Microsoft English (U.S.) v6.1 Recognizer. The only place I know how to get it though is from the Office 2003 disc (I've only verified with the Pro version). To install it without installing the rest of Office 2003, follow the below…

  • Run the CD through Autorun or by clicking Setup.exe within the disk
  • Choose Custom Install when prompted and click next
  • Uncheck all applications including Word, Excel, Outlook, PowerPoint, Publisher, Access & InfoPath (where applicable) and check the box to Choose advanced customization of applications and click next
  • Click the + next to Office Shared Features and the + next to Alternative User Input
  • Verify Speech is set to "Run from My Computer"
  • Click Next and proceed with your installation

TTS Voices

TODO

Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-ShareAlike 3.0 License