EURASIP Journal on Audio, Speech, and Music Processing

Table 10 Evaluation over all instruments and all Plackett-Burman Designs (factors as in Table 5)

From: A computational study of auditory models in music recognition tasks for normal-hearing and hearing-impaired listeners

	a		b
Fit	\(R^{2} = 0.30, R_{a}^{2} = 0.12\)		\(R^{2} = 0.37, R_{a}^{2} = 0.21\)
Coefficients	Estim.	p value	Estim.	p value
(Intercept)	0.0660	2e ⁻⁰⁸	0.0111	9e ⁻⁰⁴
Interval	−0.0074	0.40	0.0049	0.11
Onsets acc.	0.0056	0.53	0.0037	0.23
Dynamic	−0.0142	0.11	−0.0019	0.54
Acc.	0.0148	0.10	0.0056	0.07
Mean pitch	−0.0068	0.44	0.0062	0.05
Tone dur.	−0.0025	0.78	0.0025	0.42
Mel. − acc.	−0.0185	0.04	−0.0056	0.07

The error rate is the target variable— a: pitch estimation and SVM (auditory model + DFT), b: instrument recognition and SVM (auditory model features)—(italics = significant at 10% level)

Back to article page