Overview Of Noise Reduction; Nr1 (Spectral Subtraction Method) (Af Processing) - Kenwood TS-590SG Manual

04 DSP
Thus, NB2 is also effective for pulse noise that NB1 has no effect on, and causes very little unnaturalness in the auditory sense
as the target signals are not truncated. Also, unlike the NB2 of the analog era, it is not intended for a specific pulse cycle.
Meanwhile, however, when there are strong target signals, pulse sound may be heard, as differences in the amplitude of the pulse
noise is detected and attenuated to the level of the target signal.
In other words, the NB2 on the TS-590S/SG operates differently and possesses different characteristics compared to the NB2
of the analog era, and is able to effectively attenuate pulse noise that the conventional NB is unable to handle. You are thus
recommended to make use of them according to the different conditions. TS-990S is also equipped with the same NB2 as the
TS-590S/SG.

4.6.3 Overview of Noise Reduction

There are two methods available for noise reduction on TS-590S: NR1 and NR2. You can select the noise reduction that
is more effective depending on the operation mode and reception conditions.
NR1 has different algorithms that operate according to the operation mode: in the voice modes (SSB, FM and AM), a
newly-developed noise reduction method featuring audio signals based on spectral subtraction is used. In the non-voice
modes (CW and FSK), noise reduction is based on a line enhancer using an adaptive filter that emphasizes the periodic
signal. The noise reduction is automatically switched over when an operation mode is selected.
On the other hand, NR2 employs what is known as SPAC (speech processing by auto correlation) to piece together only
the periodic components detected from the RX signal and to produce the result as audio output. Table 4-1 provides the
relationship between the RX modes and NR algorithms used.
Noise Reduction
NR1
NR2

4.6.4 NR1 (Spectral Subtraction Method) (AF Processing)

Unlike the conventional noise reduction methods, the spectral subtraction method of NR1 is a brand new approach of noise
reduction developed for the TS-590S. NR1 estimates the noise component and takes away (subtracts) this estimated noise
component from the RX signal to make the target signal emerge (Figure 4-14).
This method was developed with a focus on improving the intelligibility of a weak SSB signal. Compared to the conventional
NR1 (based on a line enhancer), influences on high-frequency sound are reduced, and the audio output obtained has minimal
deterioration in the sound quality while noise is attenuated. In the spectral subtraction method of NR1, the musical noise
(tonal blip blip sound) that is inherently generated is significantly reduced through the development of a new technology.
Figure 4-14 Conceptual Scheme of NR1 Based on Spectral Subtraction
34
Table 4-1 Reception Modes and NR Algorithms Used
SSB/ SSB DATA
Spectral subtraction
SPAC
Receive Mode
FM/ AM
Spectral subtraction
SPAC
CW/ FSK
Line enhancer
SPAC