Filter contour - roll your own?

  • 1
  • Idea
  • Updated 5 years ago
I'd like to have a slider to change the receiver filter contour. Something that starts off fairly soft, then sharpens up to the current shape.

My reason for asking is because at least to my ears the current filter shape is too sharp for casual use and causes fatigue over time.

Thanks,
--jim (N7CXI)
Photo of Jim - N7CXI

Jim - N7CXI

  • 124 Posts
  • 8 Reply Likes

Posted 5 years ago

  • 1
Photo of Jim Gilliam

Jim Gilliam

  • 462 Posts
  • 71 Reply Likes
I am not an expert on filters, but it seems if there is fatigue, you could use the equalizer to gradually roll off the lower and higher frequency content.
(Edited)
Photo of Jim - N7CXI

Jim - N7CXI

  • 124 Posts
  • 8 Reply Likes
Yes, you could approximate it that way. It would be a little clumsy, since you'd need to manipulate a number of sliders to change it, and the results would not jibe with what you see in the panadapter.

I don't know what method Flex uses to generate their filter coefficients, but most common methods allow for manipulating the stopband to passband transition area. Some even allow for arbitrary response "sketching", which would be nice for custom digital-mode filters.

Since $1500 ICOM radios offer soft/sharp contours and since it *probably* wouldn't be a huge technical challenge, I'm thinking it would be nice to have a similar feature in a considerably more expensive radio.  I don't know much about the Flex team's design compromises though - implementing and displaying it in the GUI might raise an issue or two. Fortunately for me, I can make suggestions without much consideration for the work required. ;-)

--jim (N7CXI)
Photo of Ken - NM9P

Ken - NM9P, Elmer

  • 4166 Posts
  • 1329 Reply Likes
Please describe more the effect that a gentler slope has, and whether you are referring to CW or SSB.  I am intrigued.  Having never experienced variable slope skirts on my filters, I don't have a frame of reference.  I always felt that a sharper filter kept more interference out of the passband, which means less fatigue.

I get most of my fatigue when I haven't adjusted the AGC-T to the correct spot.  Adjusting it to the "sweet spot" just below the "knee" reduces the noise but still allows the signals to get through.  On CW, PSK and RTTY, I use the narrowest filter that will suit my purpose, which also reduces the noise fatigue.  The RX EQ is also very effective, as has been mentioned.
Photo of Jim Gilliam

Jim Gilliam

  • 462 Posts
  • 71 Reply Likes

Also, although the filters are sharp, you hear only what is transmitted and many of the transceivers do not have a "cliff" roll off and it seems it would not matter how sharp the filter is if it's made slightly wider than the filter of the transmitter. Of course, the background noise would stop abruptly.  

(Edited)
Photo of Jim - N7CXI

Jim - N7CXI

  • 124 Posts
  • 8 Reply Likes
The effect is a little hard to describe. With very sharp filter transitions there is an abrupt discontinuity in the response that bothers people that are sensitive to it. Softening the corners makes those transitions sound much more natural - especially the upper one.

I'm thinking more in the context of SSB operation, because softer filters for CW would be a problem in many more cases. On the other hand, **casual** CW operation would be less fatiguing for me at least because the band noise wouldn't have the sharp transition bump.

I can't tell you whether this person or that one would even notice the difference. For me, it's a fairly "big deal" - but then I've spent a good part of my life fiddling with audio and DSP subtleties like this. My tolerance slider is no doubt set lower than most. ;-)

--jim (N7CXI)
Photo of Jim - N7CXI

Jim - N7CXI

  • 124 Posts
  • 8 Reply Likes
Jim - I just saw your last post.

I'm talking more about the band/white noise than the received signal, although the latter is affected to a lesser degree.

I understand it might not be a problem for everyone, but the percentage must be fairly high if virtually all recent Yaesu rigs have a contour control, and most ICOM rigs above entry level do as well. (even those with receiver EQ's)

--jim (N7CXI)
Photo of Ken - NM9P

Ken - NM9P, Elmer

  • 4166 Posts
  • 1329 Reply Likes
I think i understand.  It is kind of like the difference between listening to a sound system that has a smooth response vs. one that has a "peaky" response.  A person can listen without offending the ears at a much higher volume with a smooth system than with one with spikes and poor EQ, regardless of the high end or low end roll off.  As long as it is smooth.  (This has been my experience with church sound systems anyway.)  I can see that for some ears the "brick wall" might be like a rock upside the head.
Photo of Jim - N7CXI

Jim - N7CXI

  • 124 Posts
  • 8 Reply Likes
Yes, that's pretty much it. We may need an acronym - "Trained Ear Syndrome (TES) " ;-)
I like your "rock upside the head" analogy.
Photo of DrTeeth

DrTeeth

  • 1687 Posts
  • 389 Reply Likes
From what I have read in this thread and others about filters, we each seem to have our own preferences so I would second the need for as much customisation as possible. For example, Fldigi will only put a notch filter in the middle of the two RTTY signals if the rig supports it (or words to that effect). I'd LOVE a RTTY notch.
Photo of Jim - N7CXI

Jim - N7CXI

  • 124 Posts
  • 8 Reply Likes
A case could certainly be made for an arbitrary response filter designer, but that goes quite a ways beyond my original request. A front-end RTTY filter would be a neat thing though, as you point out.

FWIW, I have DSP EQ software that I can use to impose whatever shape I want on the audio, but it adds yet another complication to the station. I'd much rather have all the filter fidgets in SmartSDR and not have to juggle as many balls.

--jim (N7CXI)