35
Operating Manual - nX, nXe, and nXp 150W/75W Power Amplifiers
Background Information
The feedback suppressor uses both notch filters and parametric filters to automatically and quickly suppress multiple feedback
tones and also produce a smooth equalized response which sounds pleasing for music and speech program. Before using the
FBS (Feedback Suppressor), it is helpful to understand the following basic concepts in order to effectively suppress feedback
and achieve good sounding audio program.
Each of the 12 FBS filters available on each FBS input channel can be selected between three different modes:
Restricted
- The filter's attenuation level can only be made deeper, or filter bandwidth made broader by the FBS.
Floating filter
- The filter's parameters are dynamic in that the FBS will slowly remove the filter if no longer needed and
redeploy at a different frequency as needed. Only floating filters are affected by the float time parameter which is the time after
which the FBS will slowly remove the filter.
Manual filter
- The filter is not automatically controlled by the FBS, but can only be controlled by the user.
The FBS will automatically use available restricted filters first. If there are no more restricted filters available, the FBS will
automatically use floating filters. The FBS will not automatically change manual filters.
The sensitivity control simply adjusts how sensitive the FBS algorithm is to detecting feedback and inserting filters. For example,
a setting of <Very Sensitive> means the algorithm will aggressively look for anything that could be feedback and hence it may
often insert filters on musical tones. A setting of <Very Insensitive> means that the FBS would only insert a filter in the presence
of a prolonged feedback tone with very little background noise. The optimal setting of the sensitivity control depends on both
the spectral content of the audio program (how much the musical instruments sound like feedback), and also on the amount
of background music and noise which tends to mask feedback from the detector. It is best to start with an average sensitivity
setting and experiment.
A locked filter means that the filter is prevented from being changed in any way, either automatically by the FBS or manually
by the user. To change or reset a filter, it must first be unlocked.
FIR Filter
(Outputs only)
The FIR (finite impulse response) filter is typically used when a speaker manufacturer develops and provides a proprietary FIR
coefficient file that corresponds directly to their own loudspeaker or speaker cabinet, addressing frequency and phase issues
based on their own measurements and formulations. Plugging in a FIR filter block on a DSP output and then clicking on it will
prompt the user for a *.fir or *.csv text file, which is then loaded and applied to the output DSP. The two file types will have
identical coefficient code, however the *.fir file may add comments (designated by a semi-colon) or a key=value pair definition
such as sample rate. Either file type will work with Protea
ne
software.
Simple
*.fir
file example:
; comments section
sampleRate = 48000
+0.0000042920
+0.0000030236
-0.0000040482
Simple
*.csv
file example:
+0.0000042920
+0.0000030236
-0.0000040482
Protea software does not offer the ability to view or edit the EQ curve produced by a FIR filter. Also, FIR filters are very DSP
intensive, so if a FIR filter is used on an output channel, that same numbered input channel's feedback suppressor will not be
available in order to allocate enough DSP power to the FIR filter. FIR filters can only be used on outputs, only at 48kHz sampling
rate, and can have from 2 up to 384 taps.
