EURASIP Journal on Audio, Speech, and Music Processing

Table 6 Comparison of median and mean absolute error for volume as well as RT\(_{60}\) parameters among the best-performing models

From: Exploring the power of pure attention mechanisms in blind room parameter estimation

Method	Median		Mean absolute error
Method	vol (\(\varvec{m}^{\varvec{3}}\))	RT\(_{\varvec{60}}\) (seconds)	vol (\(\varvec{m}^{\varvec{3}}\))	RT\(_{\varvec{60}}\) (seconds)
CNN [21]	728	0.64	2481	1.32
CRNN	329	0.39	2265	0.71
Proposed method w/ pretrain	294	0.31	2208	0.61

Back to article page