User Manual

18 User Manual v 1.5
UNVEIL
t[ADAPTATION]
PRESENCE
This determines the adaptation time of the analysis network
("forgetting factor"). The setting of this parameter should be
approximately equal to the reverb tail length of the signal.
You can get an idea of how to set this parameter by visually
comparing the related decay curve in the display to the
amplitude decay of the input signal. If the steepness looks
similar, you’re good. “Release Time” is a suitable analogy to
get the idea of this control (though technically, it is very
different.). If the adaptation time is lower than the reverb tail
time of the reverberated signal you will hear some reverb
come back on a few msec after the incident sound. In this
case you should increase the adaptation time until the reverb
is gone.
This parameter adds a random signal to the network's input
on each pass of the analysis recursions. This changes the
starting conditions of the evaluation slightly, and causes the
output to be more generalized. The effect on the audio signal
is mainly a gain in high frequency content and a more natural
sound, as the noise and short time scale fluctuations in the
signal are averaged away and remain largely untouched.
With some "well behaved" signals this setting might not have
any effect, some signals (especially voice) can benefit greatly
from increasing this value. Start low & raise as needed.