Discrimination of MP3-compressed musical instrument tones.
Chung Lee, Andrew Horner · The Journal of the Acoustical Society of America · 2010
Eight musical instrument tones were compressed using an MP3 codec to determine how the detection of compressed sounds varies with bit-rate and instrument. Sounds with harmonically flattened frequencies were compressed with bit-rates of 32, 40, 48, 56, 64, 80, 96, 112, 128, and 160 kbps. Listeners were asked to discriminate the compressed sounds from the original sounds. Averaged over the eight instruments, discrimination was very good (above 80%) for bit-rates of 32 and 40 kbps, moderate (above 70%) for 48 and 56 kbps, and poor (around 50–60%) for bit-rates above 64 kbps. Statistical analysis showed that discrimination was significantly affected by the instrument for intermediate bit-rates of 40, 48, 56, and 80 kbps. Discrimination scores were strongly correlated with the signal-to-mask ratio and spectral irregularity of the original tone. Relative spectral error as a predictor for discrimination accounted for more than 80% of the variance in the discrimination scores.