Abstract
This article presents the results of an experimental study evaluating the effectiveness of two approaches to digital audio file compression: fractal and fractal-spectral codecs. The primary objective is to compare compression ratio, reconstructed signal quality, and computational efficiency for each method on a standardized audio dataset covering diverse genres and durations. Objective and subjective quality metrics—including signal-to-noise ratio (SNR), compression ratio (CR), and perceptual audio quality assessment—are analyzed and compared. The results demonstrate that the fractal-spectral approach, which integrates frequency-domain analysis via wavelet decomposition with classical fractal modeling, achieves a higher compression ratio with comparable perceptual quality relative to the classical fractal codec. This renders the fractal-spectral method a promising technique for modern multimedia systems where maximum compression efficiency and acceptable audio restoration quality are simultaneously required.