Digital Signal Processing Tools for Loudspeaker Evaluation and Discrete-Time Crossover Design

摘要

Loudspeaker response data can be decomposed into minimum- and excess-phase components. Subjectively less significant elements can then be filtered and a new family of response curves computed. Applications to both cumulative-decay spectra (CDS) and energy-time curves (ETC) are discussed, where noocausal attributes are corrected. A versatile technique for producing digital crossover alignments is presented together with the implications on CDS and ETC. A novel method of masking polar response errors within the crossover transition band is also proposed.