From: Music detection from broadcast contents using convolutional neural networks with a Mel-scale kernel
Label | Korean drama (dev) | Korean reality | British broadcast | Spanish broadcast | MIREX 2015 | Dafx 07 |
---|---|---|---|---|---|---|
Pure music | 35.2 | 19.8 | 26.9 | 29.5 | 74.3 | 43.0 |
Music + noise | 15.4 | |||||
Music + speech | 15.7 | 56.5 | 20.0 | 20.5 | 2.1 | |
Music + speech + noise | 1.0 | |||||
Total music | 67.5 | 76.3 | 53.0 | 50.0 | 76.5 | 43.0 |
Total non-music | 32.4 | 23.6 | 46.9 | 50.0 | 23.5 | 57.0 |
Total duration (h) | 2.9 | 1.6 | 7.9 | 12 | 5.24 | 9 |