Skip to main content

Table 2 RIR and noise for training data augmentation

From: Progressive loss functions for speech enhancement with deep neural networks

  Room impulse responses
  Small Medium Large
Probability 0.5 0.3 0.2
Size (x,y,z)[m] xU(1,6),yU(1,6),zU(2,3.5) xU(6,10),yU(6,10),zU(3,5) xU(10,20),yU(10,20),zU(4,6)
RT60[s] RT60U(0.1,0.25)
Distance[m] 0.5, 1.0, 1.5, 2.0, 2.5
Microphone type Bidirectional, hypercardiodid, cardioid, subcardoid, omnidirectional
  Noise
Music 659 files
Noise 929 files
SNR [dB] SNRU(5,25)