Quality Enhancement of Compressed Audio Based on Statistical Conversion

EURASIP Journal on Audio, Speech, and Music Processing

Table 8 Subjective listening test scores.

Preference ratio of enhanced file over 64-kbps file
	95% confidence interval
Rock	0.40	± 0.27
Electronic	0.50	± 0.25
Symphony	0.47	± 0.26
Jazz	0.46	± 0.27
Flute	0.53	± 0.25
Violin	0.44	± 0.23
Bass	0.46	± 0.27
Piano	0.47	± 0.24
Harp	0.54	± 0.27
Drums	0.56	± 0.23

Preference ratios (normalized to a maximum value of 1) of the enhanced file over the 64-kbps MP3 file along with the 95% confidence intervals, for the case of 32-kbps MP3 enhancement. The enhanced file is always preferred over the 32 kbps MP3 file for all signals.