6.5.2 DeltaPowerMask

6.5.2.1 Outline of the node

This node generates mask values for dynamic logarithmic power, which is a kind of acoustic feature. The generated mask value is added to the mask vector element of an input.

6.5.2.2 Necessary file

No files are required.

6.5.2.3 Usage

When to use

This node is used to perform speech recognition by masking features depending on reliability based on missing feature theory. It is usually used for the posterior half of DeltaMask .

Typical connection

\includegraphics[width=100mm]{fig/modules/DeltaPowerMask}
Figure 6.71: Typical connection example of DeltaPowerMask 

6.5.2.4 Input-output and property of the node

INPUT

: Map<int, ObjectRef>  type. A pair of the sound source ID and mask vector of the feature as Vector<float>  type data. The mask value is a real numbers from 0.0 to 1.0. 0.0 indicates the feature is not reliable and 1.0 indicates it is reliable.

Output

OUTPUT

: Map<int, ObjectRef>  type. A pair of the sound source ID and mask vector of the feature as Vector<float>  type data. The mask value is a real numbers from 0.0 to 1.0. 0.0 indicates the feature is not reliable and 1.0 indicates it is reliable. The dimension size is one more than the input dimension.

Parameter

6.5.2.5 Details of the node

This node generates a mask value of the dynamic logarithmic power, which is one of the acoustic features. The mask value generated is 1.0 consistently. The dimension of the output mask is the mask’s dimension+1.