2
1. Introduction to the DSP-9+
The DSP-9+ is an audio noise filter for amateur radio voice, data and CW operation. The DSP-
9+ filters and reduces noise and interference to improve radio reception. The DSP-9+ uses
digital signal processing technology to implement algorithms that perform four basic functions:
1) Random noise reduction, 2) Adaptive multi-tone notch filtering (Tone noise reduction),
3) Bandpass filtering, and 4) RTTY remodulation.
Random/Tone Noise Reduction
The noise reduction functions of the DSP-9+ operate by examining a characteristic of signals
and noise called correlation, and dynamically filtering out the undesired signals and noise. The
degree of correlation is relative. Random noise such as white noise or static is uncorrelated.
Speech is moderately correlated. Repetitive noise such as a heterodyne is highly correlated. The
DSP-9+ measures correlation and filters out signals and noise that are outside its correlation
thresholds. There is little degradation of the desired speech signal. The amount of noise
reduction varies according to the correlation characteristics of the noise. Typical noise reduction
ranges from 5 dB to 20 dB for random noise and up to 50 dB for heterodynes.
Bandpass Filters
The DSP-9+ has bandpass filters that are used in voice, data and CW modes. In a typical
example of a voice mode application, a bandpass filter can improve a signal with a poor signal-
to-noise ratio. A bandpass filter removes the high and low audio frequency components that do
not contribute significantly to the speech intelligibility, thus improving signal quality. Another
common voice mode example is the improvement of a SSB signal corrupted by adjacent channel
interference (QRM). The steep skirts of the bandpass filters allow the interference to be
eliminated with minimal effect on the desired signal. In the voice mode, two front panel push
buttons select one of three voice bandpass filter bandwidths from two sets of filters. An internal
jumper behind the back panel selects the filter set, either 1.6, 2.0, and 2.4 kHz., or 1.8, 2.4, and
3.1 kHz.
CW signals require bandpass filters with steep skirts and linear phase response. Linear phase
response maximizes the usable signaling rate for a given bandwidth and minimizes ringing often
heard on extremely sharp filters. The DSP-9+ has 18 different CW filters with skirts so steep
that a signal literally falls off the edge of the pass band as you tune through a CW signal. The
bandwidths of the CW filters are 500, 200 or 100 Hz. A front panel push-button selects either of
two CW bandpass filter center frequencies chosen from a set of 400, 500, 600 and 800 Hz.
Internal jumpers behind the back panel program the two choices. The jumpers also allow the
choice of a special set of filters for the Collins Radio KWM-2. The Collins filters have center
frequencies of 1350 and 1500 Hz. The narrow filters are useful for trying to dig out extremely
weak signals from the noise and QRM. The wider filters allow easy tuning and listening to
multiple CW signals simultaneously.
Data signals also require bandpass filters with steep skirts and linear phase response. There is an
optimum bandwidth for each signaling rate and modulation type. Any wider bandwidth than
necessary will increase the bit error rate of the data communication link by allowing more noise
into the demodulator. The DSP-9+ has four data bandpass filters for five popular data types,
Summary of Contents for DSP-9+
Page 1: ...DSP 9 Audio Noise Reduction Filter Operating Manual Version 3 0A Timewave Technology Inc ...
Page 5: ......
Page 15: ......
Page 19: ......
Page 20: ......