Exposing Speech Transsplicing Forgery with Noise Level Inconsistency

Table 4

SNR estimation for various algorithms.

Algorithm	0 dB	5 dB	10 dB	15 dB
[23]	4.7419	9.8905	14.7146	18.5088
[24]	2.6064	7.7292	12.5088	16.5755
[25]	3.5518	8.2249	12.3533	15.6897
[26]	3.3987	8.7689	13.7923	17.8140
[27]	4.3277	8.5393	12.1390	14.7518
Optimized Sorensen	0.1296	5.0011	10.0396	15.0139

The value in bold type is used for emphasis that the performance of the optimized algorithm is better than other algorithms.