Analysis of Two-Piano Teaching Assistant Training Based on Neural Network Model Sound Sequence Recognition

Table 4

Low signal-to-noise ratio model identification results.

Input noise type	Whitt (%)	F16 (%)	HF channel (%)
Noisy	64.2	36.7	52.4
AET	53.7	31.2	44.5
R-CED	48.5	25.4	38.7
UNetSMB	43.6	20.8	32.6