9.2 Saving separated sounds to files

Problem

Use this section when listening to separated sounds for checking or when saving a separated sound to a file for experiments.

Solution

Use the SaveRawPCM or SaveWavePCM module. The output format of SaveRawPCM module is the Raw format of Integer with no headers. The output format of SaveWavePCM module is the wave format of Integer with a header. Saving methods differ according to the values to be saved.
1. Saving real number signals (temporal waveforms)
When saving temporal waveforms, connect SaveRawPCM or SaveWavePCM , as shown in Fig. 9.1. If not connecting a module that performs spectral transforms, such as MultiFFT , set the ADVANCE parameter identical to the number of dimensions of the input object (INPUT).
2. Saving complex signals (spectrum)
When saving spectra, connect SaveRawPCM or SaveWavePCM as shown in Fig. 9.2. Both must first be resynthesized to form temporal waveforms. Connect the Synthesize module and convert the spectra into temporal waveforms. Subsequently connect SaveRawPCM or SaveWavePCM , as shown above for real number signals. Here, the ADVANCE parameter must be same as that of the Synthesize module. Determine if the signals have been saved successfully by seeing if a separated sound file was generated at the time of execution.

\includegraphics[width=.8\textwidth ]{fig/recipes/CB_Separation_002_2}
Figure 9.1: Connection example1
\includegraphics[width=.8\textwidth ]{fig/recipes/CB_Separation_002_1}
Figure 9.2: Connection example2

Discussion

The SaveRawPCM and SaveWavePCM modules differ only in their headers. In general, the wave format has a header making is easier to deal with conventional software. Therefore, users should use the SaveWavePCM module.

See Also

Module references for Synthesize , SaveRawPCM and SaveWavePCM .