|
The Speech Performance Analyser provides almost all the measurements that DSLA makes in a convenient package
for codec developers and software engineers. The SPA dynamic linked library can be called from a user application program with reference
and degraded filenames. The SPA returns key numerical data to the application program along with a wealth of detailed graphical information
to a results file if desired.
The Speech Performance Analyser delivers:
- Predictions of Speech Quality using one or more objective evaluation metrics including PESQ ITU-T Recommendation P.862/P.862.1
- Mean Active Speech Level and Activity Factor for reference and degraded speech files to ITU-T Recommendation P.56, Method B
- Noise in the presence of speech
- Peak levels
- Average Time Offset between reference and degraded speech files
- Range and standard deviation of time offsets
- Lost speech or temporal clipping data
- Graphical result data for reference, degraded and error surfaces, transfer function estimate, speech quality and time offset variation by utterance or frame
Contact sales@malden.co.uk for more information about SPA or for a 30 day Evaluation Pack that
includes a simple user interface to Speech Performance Analyser with PESQ and the Speech Performance Viewer to look at the results.
SPA is supplied under a software usage agreement for applications running under Windows or VxWorks.
|
