S E T U P F O R M C H A P T E R 8
The Automatic Notch Filter (ANF) attempts automatic computation of a filter to remove one or more
carrier tones that are interfering with the signal of interest.
#Taps: This determines the length of the computed notch filter. The longer the filter, the larger the
number of tones that can be cancelled and the more effective the cancellation will be. It also
introduces latency (signal delay) equal to the number of filter taps divided by the sample rate in
samples per second and is in addition to Delay (see below). The larger the number of taps, the
longer it takes for the filter to converge but upon achieving convergence, the better the filter will be.
Delay: Determines how far back to look in the signal before beginning to compute a cancellation
filter. The larger the delay, the less the impact on normal speech, and the more likely the filter will
be able to concentrate only on longer term coherent signals such as carrier tones. Latency is
introduced that is equal to the Delay.
Gain: Determines the adaptation rate of the filter. The larger the number, the faster the filter will
converge and the less stable it will be.
Block LMS: We have implemented a fast block LMS routine. This routine is faster than the old
routine, gives more taps in the filter, and will work on I/Q signals which enables it to work in BIN
mode.
Buffer Size
This controls the size of the DSP buffers, which determines the size of the FFT filter and therefore the
group delay (latency) through the digital filter. Higher values will result in more latency and sharper (“brick
wall”) filters. Lower values will allow nearly real time monitoring with filters that “roll off” (as opposed to
the typical “brick wall” filters).
Note 1: Using lower values will negate the effects of some of the narrowest filters
(25 or 50Hz).
Note 2: The DSP Buffer is completely independent of the Audio buffer size.
Noise Blanker
This controls the detection threshold for impulse noise. If a signal sample exceeds this detection threshold,
the sample will be set to zero and the filtering in the radio serves to interpolate through this zero sample.
This noise blanker is identical in theory to those in traditional radios. The detection threshold in our noise
blanker has the unique feature that they are signal strength dependent. This enables them to function
properly at all signal levels.
This control is preferable when the spikes are very large in comparison to the average signal. However,
when the spike is smaller, Noise Blanker 2 provides a much cleaner reconstruction of the signal since the
signal is more likely to look like the mean. For this reason, the Noise Blanker 2 threshold should always be
about four or five less than the Noise Blanker threshold.
86 FlexRadio Systems
To Next Page
Loading...
Need help?
General
