User Control for Content Reconstruction in Speech Compression

This Army SBIR Phase I effort is intended to examine and document the feasibility of expanding a novel proprietary technique of lossy speech compression that allows reconstruction of a perceptually equivalent signal through the extraction of salient bits of data. This expansion is meant to provide user-input at the stage of extraction from an uncompressed signal. Essentially, this would permit a user to adjust the means for defining the template for reconstruction. We envision four related adjustments, which could be viewed like dials on an old-fashioned ham radio, implemented in this case as software-based controls. Combined, these adjustments would allow an end-user sufficient control over the variables involved to permit reconstructions based on various parts of the signal in flexible relation to each other.