Tuning Data Architecture

This content is no longer actively maintained. It is provided as is, for anyone who may still be using these technologies, with no warranties or claims of accuracy with regard to the most recent product version or service release.

Tuning data originates in a Speech Server speech application. It is first stored in event trace log (ETL) files and then imported to a Microsoft SQL??Server database, where it can be used for performance tuning and reporting.

Data Flow

Data Entities and Tools

Use ETL data to analyze and tune speech applications. This is accomplished through instrumentation and ETL conversion tools.

Speech Application Instrumentation

Instrument the speech application by adding events and structuring turns and tasks so that data is modeled in a useful way. For more information, see Design Speech Applications for Easy Reporting and Tuning.

ETL Files

Tuning and reporting data is initially stored in ETL files on the computer where Speech Server is installed. For more information, see Logging Administration. To use data in ETL files for tuning and application analysis, first export the data to SQL??Server.

ETL Conversion Tools

The following command-line tools are provided to convert data in ETL files.

Tool	For More Information
MssLogToDatabase.exe	Import Log Files into the Tuning Database
MssContentExtract.exe	Extract Audio and Grammar Streams to File
MssLogToText.exe	Log File Conversion to Text

Analytics and Tuning Studio

Use Analytics and Tuning Studio to view reports and perform rapid, intuitive analysis of speech application log data. For more information, see Speech Application Data Analysis and Speech Recognition Tuning.

Reports

Speech application reports provide basic reports that analyze the logging data imported from ETL files to the tuning database. For more information, see How to: View Analytics and Tuning Studio Reports.

See Also

Other Resources

Management of Tuning Data
Speech Application Data Analysis
Speech Recognition Tuning