Research Article
Multiengine Speech Processing Using SNR Estimator in Variable Noisy Environments
Table 2
Speech/nonspeech classification results (%).
| Audio class | (dB) | | −10 | 0 | 10 | 20 | 30 | Clean |
| White noise | | | | | | | Nonspeech | 84.0 | 95.2 | 98.1 | 98.3 | 98.6 | 98.6 | Speech | 32.0 | 93.2 | 96.8 | 97.3 | 97.3 | 97.3 | KING | 33.0 | 93.7 | 98.0 | 98.6 | 98.6 | 98.6 | Average | 58.2 | 94.3 | 97.7 | 98.1 | 98.3 | 98.3 | Pink noise | | | | | | | Nonspeech | 87.0 | 94.0 | 97.5 | 97.8 | 98.4 | 98.6 | Speech | 27.1 | 90.0 | 96.0 | 95.9 | 95.9 | 96.7 | KING | 29.2 | 89.1 | 97.5 | 98.2 | 97.4 | 97.5 | Average | 57.5 | 91.7 | 97.1 | 97.8 | 97.9 | 97.9 |
| Speech/nonspeech average | 57.8 | 93.0 | 97.4 | 98 | 98.1 | 98.1 |
|
|