To Ring Or Not To Ring, That Is The Question?

  • 1
  • Question
  • Updated 2 years ago
  • Answered
Attention "Tech Heads"!

(While we all wait for SSDR 2.0 to be "born"...)

I was at a very enjoyable Hamfest this weekend over by Sussex County NJ (and I think that I spied a regular from this Forum there).

I bumped into an old time "moon bounce" EME guy and we were discussing the new tech being developed in the Ham world - he thought that this was primarily from the microwave guys out there (here?).

I mentioned SDR advances and he noted my Flex Radio baseball hat, and we got involved in a chat of the pluses (IMHO) of SDR tech, which he wasn't that sure about, although, he admitted that he hadn't used an SDR ever.

One point he seemed to dispute completely was that digital filters do not experience ringing. He kept saying that they ring as well.

As I have been reading the subject it seems like not all digital filters were created equal. Of course I am aware of FIR  and IIR filters, but there are different techniques that use a Bayesian filter algorithm, or a Fourier filter algorithm. I think that you can build a filter that rings, but my question is whether all digital filters possess some degree of ringing, although in some filters the ringing is so attenuated that it is not even audible.

I get the idea that this "ring or no ring" is going to collapse into the "black hole" of a matter of semantics.

Any thoughts? Ideas?

By the way, I suggested that he look at some Flex Radios on YouTube, but he had no interest in looking at videos of radios, so I offered him a "test drive" with my 6700 "one of these dayz"...

Photo of Roy Laufer

Roy Laufer

  • 451 Posts
  • 45 Reply Likes

Posted 2 years ago

  • 1
Photo of Steve - N5AC

Steve - N5AC, VP Engineering / CTO

  • 1057 Posts
  • 1099 Reply Likes
Official Response
Filter ringing is caused when a stimulus starts a process in the filter that continues once the stimulus is removed.  Filters that ring continue to produce a sound of a particular frequency, typically when that frequency is the stimulus, even after the signal is gone.  Obviously, this can have serious detrimental effects on recovered CW in your brain.

In order for a filter to ring, the output of the filter must be reconnected to the input of the filter (or something "after" in the filter must be connected to something "before").  FIR filters never connect the output back to the input and so they cannot ring.  IIR filters have lower latency typically, but do connect outputs back to inputs and most definitely can ring.

In the SmartSDR, the bandpass filters are all FIR today.  The APF is an IIR filter.

Having said all this, I want to point out something that is often lost on individuals trying to decide if their filter rings.  When you use a small filter (think 0-200Hz wide), you restrict all of the frequencies that are going to come out of your radio to this narrow band.  If you use a 50Hz CW filter and then turn up AGC-T, then the noise floor will be "loud" in your receiver and then it will be filtered to a 50Hz range.  Then people will say "wow that filter rings!"  No, it does not ring!  If you are using a narrow filter and listening to something close to the noise floor, necessarily you will hear noise and the signal at similar levels and frequencies and it is difficult to discern the difference, but this is not ringing.  Turn your AGC-T down until the noise drops off and then the CW will stick out of the noise floor (presuming it's strong enough to do so).  I find that a typical range on AGC-T on the HF bands for me is 35-55.  If you're at 75, you're probably doing something wrong.

Your brain is good at picking signals out of noise and if a 50Hz filter is hard to hear, try widening the filter and see if it is more comfortable to pick CW out when the noise that is present is across a wider range of frequencies.  Also, use the APF, but not with the slider all the way to the right and see if it helps improve SNR of the signal you're listening to.