EURASIP Journal on Audio, Speech, and Music Processing

Table 2 Statistics for Experiments 4 and 5

From: ViSQOL: an objective speech quality model

		E4			NOIZEUS			P.Sup23
	Pearson	Spearman	Std. err.	Pearson	Spearman	Std. err.	Pearson	Spearman	Std. err.
Without fit
ViSQOL	0.80	0.74	0.20	0.87	0.74	0.23	0.77	0.64	0.47
PESQ	0.78	0.57	0.18	0.90	0.57	0.20	0.92	0.90	0.29
POLQA	0.81	0.65	0.26	0.77	0.65	0.29	0.96	0.96	0.20
LLR	0.25	0.27	0.20	0.88	0.27	0.22	0.44	0.18	0.65
fwSNRSeg	0.21	0.14	0.21	0.86	0.14	0.24	0.48	0.38	0.64
With linear fit
ViSQOL	0.85	0.77	0.22	0.84	0.86	0.25	0.81	0.70	0.43
PESQ	0.78	0.57	0.26	0.90	0.88	0.20	0.92	0.90	0.29
POLQA	0.81	0.65	0.24	0.77	0.79	0.29	0.96	0.96	0.20
LLR	0.25	0.27	0.40	0.88	0.88	0.22	0.44	0.18	0.65
fwSNRSeg	0.21	0.14	0.40	0.86	0.85	0.24	0.48	0.38	0.64

Back to article page