[TenTec] Noise Reduction Setting

[Merle Bone]

Kim said:
In playing with it on CW, this doesn't appear to be the case: 
different settings definitely result in different amounts of noise 
reduction, and the lower settings seem to do their job immediately, 
but not as aggressive as the higher settings. Intuitively, this is 
what I'd expect.

Has the NR changed in later firmware versions and the manual failed 
to keep? Or, has the manual always been in error?
-------------------------------------------------------------------------------------------------------
Kim, your observations are correct. The V1 and early V2 noise reduction DSP code implemented different filters from the current V2 DSP firmware in the bandpass. What happens with NR depends on the setting of the multi-knob for NR. In V1 and early V2 releases, setting 1 was either a lowpass or highpass filter (depending on the frequency of the signal in the bandpass) and settings 2 thru 9 were bandpass filters (centered around the signal) with more poles (steeper skirts) as you increased the setting from 2 thru 9. The idea was that setting 1 would very quickly form a low or high pass filter for use in SSB so there was minimal voice distortion as the filter was built and removed ( The LMS (Least Mean Square) filter system has some forward leakage to cause the filter to collapse between audible sounds). Settings 2 through 9 were increasingly slower in being built but sharper filters and "more effective"(Always debateable when compared to just narrowing down the DSP bandwidth with the bandwidth control) for CW. During the V2 Beta process Ten-Tec decided to change the design of NR and just use the Lowpass filter. Settings 1 through 9 then set the "roll-off rate" of the lowpass filter and some level of fixed attenuation - slow roll-off at setting 1 with little attenuation and steeper roll-off at 9 with much more attenuation (Creating and removing this filter is still faster then the bandpass filters that were used in V1 and they use fewer DSP cycles) (and presumably less memory since you  don't have to store multiple filters). The Orion manual has not really been updated in this area. You can observe the effect of NR by using a spectrum analyzer to look at the output of the Orion. This was all part of the process of rewriting the V2 DSP code to make it more like the code in the Jupiter and earlier TenTec DSP efforts.

The assessment of "what's helpful" in NR is probably mostly subjective and also depends on what modes you predominately operate and how you set up the Orion when you use different modes. In the end, the idea is to remove "some" of the noise and still have the signal be easy to correctly copy.  Since this is mostly done with signals we copy with our ears, and people "hear differently," what is best in NR is often seen differently by people. 
73, Merle - W0EWM