8

Table Of Contents
Chapter 6 Filter 83
U/V Detection Section
The U/V Detection section detects the unvoiced portions of the sound in the analysis
signal, improving speech intelligibility. Please refer to “Unvoiced/Voiced (U/V)
Detection on page 165, for a full explanation of the U/V Detection principle. More
information about improving speech intelligibility can be found in “Tips for Better
Speech Intelligibility on page 169.
 Sensitivity knob: Sets the degree of responsiveness of U/V detection. By turning this
knob to the right, more of the individual unvoiced portions of the input signal are
recognized.
When high settings are used, the increased sensitivity to unvoiced signals can lead to
the U/V source—determined by the Mode parameter—being used on the majority
of the input signal, including voiced signals. Sonically, this results in a sound that
resembles a radio signal which is breaking up and contains a lot of static or noise.
 Mode pop-up menu: Choose the sound source(s) which can be used to replace the
unvoiced content of the input signal. The choices are: Off, Noise, Noise + Synth, or
Blend.
 Noise: Uses noise alone for the unvoiced portions of the sound.
 Noise + Synth: Uses noise and the synthesizer for the unvoiced portions of the sound.
 Blend: Uses the analysis signal after it has passed through a highpass filter, for the
unvoiced portions of the sound. This filtered analysis signal is then mixed with the
EVOC 20 TrackOscillator output signal. The Sensitivity parameter has no effect on
this setting.