6.7.29 SegmentAudioStreamByID Outline of the node

This node extracts acoustic streams that use ID information and output them while adding ID information to them. Necessary files

No files are required. Usage

When to use

This node is useful in extracting and processing specific parts of acoustic streams, such as audio parts, not in processing all entire acoustic signals as one stream. Since extractions are performed with IDs as keys, ID information is essential for inputs. This node extracts sections in which the same ID is continuously seen as a signal and outputs these sections as Map data in one channel, adding ID information to it.

Typical connection

We presume that the input stream is a mixture of sounds from two audio signals. When comparing the signals of the original sound mixture separated by GHDSS from a temporally similar signal, the acoustic stream of 1 channel is input, as is the sound source detected by sound source localization to this node. In this case, the outputs are in a format (Map ) identical to those of sounds separated by GHDSS and PostFilter .

\includegraphics[width=.8\textwidth ]{fig/modules/SegmentAudioStreamByID}
Figure 6.140: Example of a connection of SegmentAudioStreamByID Input-output and properties of the node



: any , Matrix<complex<float> > or Matrix<float> type.


: Vector<ObjectRef> type. Sound source directions with IDs. The contents of each Vector are Source type indicating the sound source information with IDs. Feature vectors are stored for each sound source. This parameter must always be designated.



: Map<int, ObjectRef> and ObjectRef are slender pointers to Vector<float> and Vector<complex<float> > . Details of the node

This node extracts acoustic streams using ID information and outputs them while adding ID information. Note that this node converts Matrix<complex<float> > and Matrix<float> into Map<int, ObjectRef> while adding IDs, although at present supports data from only 1 channel as input.