Free complete embedded speech recognition software for windows

The Windows Speech Recognition Macros tool ? or WSR Macros for short ? extends the usefulness of the speech recognition capabilities in Windows Vista.
Users can create powerful macros that are triggered by spoken commands which can perform a series of tasks from as simple as inserting your mailing address to as complex as providing a completely different speech interaction with applications.
While we have tried to make it easy to use, this release of WSR Macros is a technical preview of technology we are planning to release in the future. Not all the features we have planned are included, and some are incomplete. Users are cautioned to treat this release as "pre-beta." While creating simple macros is very straightforward, the creation and editing of more complex macro files is best suited for advanced users. We welcome your feedback on how we can improve WSR Macros to best suit your needs.
System requirements:
- A computer running a genuine copy of Windows Vista (x86 or x64)
- Currently, only English-language editions of Windows are supported.
- A working microphone
- Optionally, speakers or headset for sound and voice output
We recommend the following in addition:
- A minimum of 1GB of memory is recommended when using speech alongside of other applications.
- Speech recognition works best when the computer can hear you clearly. We recommend high-quality microphones for best results. Microphones that are placed near the speakers mouth are optimal, such as a headset microphone.
- We recommend that you complete the Microphone Wizard and additionally the Windows Speech Recognition tutorial before using WSR Macros. This will ensure that your microphone is setup properly and help the speech engine become adapted to your voice.
- Please note that most Bluetooth microphones do not function well with Windows Speech Recognition due to limited audio bandwidth.

Text-Independent Speaker Recognition Based on Neural Networks

The Text to Speech application was designed to a generation of synthesized speech from text. Our goal is to make synthesized speech as intelligible, natural and pleasant to listen to as human speech and have it communicate just as meaningfully.
We have developed a novel TTS system, Naxpres, built on IBMs successful work in data-driven methodologies for speech recognition. Our system obtains its parameters through completely automated training on a few hours of speech data, which is acquired by recording a specially prepared script. During synthesis very small segments of recorded human speech are concatenated together to produce the synthesized speech.