This node is used to output only sound source separation results with IDs over a designated value among multiple sound source separation results. In particular, when setting the FIXED_NOISE property of the GHDSS node to true, negative IDs are given to stationary noise separation results. This node is used to filter the sounds other than these.
No files are required.
When to use
The sound source separation node GHDSS switches on the robot although it does not move it. When performing sound source separation under conditions in which the noise (e.g. the sound of the fan) is both stationary and known, an ID of the separation result of the noise is output as -1. Thus, when connecting SourceSelectorByID after the GHDSS node and setting the threshold value to 0, the separated sound of the stationary noise can be ignored in subsequent processing.
Typical connection
Figure 6.94 shows a example of a connection, in which the node is connected to the posterior half of GHDSS .
Input
Map<int, ObjectRef> type. Since this node is usually connected after the sound source separation node is connected, sound source IDs correspond to int , the key of Map . ObjectRef is of Vector<float> type (power spectra) or Vector<complex<float> > (complex spectra), indicating a separation .
Output
Map<int, ObjectRef> type. Only data with sound source IDs greater than MIN_ID are extracted as output. The content of Map is same as that of INPUT.
Parameter
Parameter name |
Type |
Default value |
Unit |
Description |
MIN_ID |
0 |
Sound sources with IDs greater than this value are passed. |
int type. Separated sounds with sound source IDs greater than this parameter value are passed. The default value is 0. When this node is connected after GHDSS , no changes to the default value are required.