8.7 Localization results are fragmented / Isolated sounds are connected

Problem

Read this section if

Despite sounds being continuous, localization results are discontinuous.
All acoustic localization results are output continuously.
The user wishes to set an appropriate value for the PAUSE_LENGTH property of the SourceTracker module.

Solution

Connect the SourceTracker module to the DisplayLocalization module and display the localization results.
Read an appropriate sentence aloud, and see the localization results.
Localization results break off: Increase the value of PAUSE_LENGTH.
Localization results are too close:

Discussion

The purpose of the PAUSE_LENGTH property is to recognize speech appropriately, even if the power of a MUSIC spectrum in the LocalizeMUSIC module localizes it as continuous speech. Since this is applicable only to human speech, such a sound can be used. If your purpose is to localize human speech, use the default value.

8.7.1 PAUSE_LENGTH units

PAUSE_LENGTH is measured in milliseconds. Therefore, the maximum PAUSE_LENGTH depends on the sampling frequencies of the AudioStreamFromMic and AudioStreamFromWave modules (SAMPLING_RATE) and the step size (ADVANCE) of FFT. If all parameters are set at their default settings (sampling frequency, 16000Hz; step size, 160 pt), changing PAUSE_LENGTH by 1 corresponds to changing it 1 msec.