SPro is a toolkit for processing speech signals that offers real-time commands for executing standard algorithms for feature extraction.
Written in ANSI C, the SPro library provides a range of functions, including waveform signal input, low-level signal processing (FFT, LPC analysis, etc.), low-level feature processing (lifter, CMS, variance normalization, deltas, etc.), and feature I/O. However, the library doesn't provide high-level feature extraction functions, which directly convert a waveform into features, mainly because such functions would entail a tremendous number of arguments.
The latest version of SPro (Version 4.0) comes with some significant changes. Most notably, the toolkit now offers feature streams, much like the signal streams used for input. This new feature I/O mechanism enables the processing of continuous (or very large) audio streams. Introducing feature streams required a major redesign of the source code, and all the library files have been greatly modified. To prevent confusion, old spro_*.c names have been changed. Additionally, the SPRO FEATURE FILE FORMAT HAS CHANGED to accommodate streams (e.g. number of frames unknown).
Some of the other features of the latest version include feature I/O via feature streams, a redefined SPro feature file format, which now includes an optional variable length ASCII header. Besides, it comes equipped with added LSF computation in slpc, normalization qualifier (R), and direct computation of delta, delta-delta, and CMS in the main tools.
Overall, SPro is an excellent toolkit for anyone seeking to implement speech-related applications. With its improved functionality and user-friendly interface, it is a must-have for all software developers. So, if you are in the market for an efficient speech signal processing toolkit, look no further than SPro!
Version 4.0: N/A