Morse Code Signal Shaping Filter

[DF7TV]

Hello,

just to share:

Recently I changed my setup for iCW using a software Shaping Filter (combination of three filters) in the RX-path and the TX-path which seems to work nicely in both paths for CW speeds from 20 to 60 wpm.

The filters (VST-plugins) are provided by Robin Schmidt, www.rs-met.com

For a side tone frequency of f_s = 666 Hz, Filter 1 is a 2-nd order bandreject (notch) filter (Bessel) centered at f = 2 * f_s with a bandwidth of 0.5 oct., Filter 2 is a 4-th order lowpass (Papoulis) with cutoff frequency at f = 2 * f_s and Filter 3 is a 3-rd order bandpass filter (Bessel) centered at f = f_s with a bandwidth of 0.5 oct.

Some documentation available at

https://www.qsl.net/df7tv/iCW-at-DF7TV.html

73

Tom DF7TV

[Andrew Moore]

Tom, kudos on profile pic :)

--

---
You received this message because you are subscribed to the Google Groups "iCW - internet CW" group.
To unsubscribe from this group and stop receiving emails from it, send an email to i_cw+uns...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/i_cw/083ffb18-62a8-41bd-990a-53691f499f91n%40googlegroups.com.

[Chuck Vaughn]

excellent !

that Engineer's Filter VST PLUGin is amazing !

one of my most favorite AUDIO DSP VSTs ... EVER...

sure wish it was ported to LINUX as a LV2 ...but it has to use W.I.N.E., and it always causes audio glitches from HIGH CPU use...

interesting combo of filter types you have put together ... was that done by ear or by math/theory calculations/calculus  etc.. ?

--

[DF7TV]

Hi Chuck,

Well there was a bit all ingredients you mentioned...

1. Ear:

I always like to listen to the nice sound of W5UXH's Ebook Code Practice Room at the iCW Chicago Server and the so far here used Bessel bandpass filter only did not come near to that sound.

2. Reading:

I read the original article on the HI-PER Audio Filter by D. Cripe KC3ZQ (73 Amateur Radio, May, 1994) and tried to let me inspire by his ideas.

3. Analyzing:

I looked into the spectrum of the W5UXH Ebook signal and found that the slope towards lower frequencies is not as steep as to the higher ones. I played around with setups to get this part (lower frequencies) of the spectrum more present than with a single bandpass filter. I assumed that the "light"/"smooth" character of the sound of the Ebook signal is at least in part due to this enhanced low frequency part of the spectrum.

So an idea comming was to use a less sharp bandpass filter and to add a lowpass filter to get rid of the harmonics above the fundamental frequency (side tone f_s). The addition of the band-reject filter helped to suppress the second harmonic (2 * f_s) which is especially present for a switched square-wave signal.

4. Testing

Using my setup with Voicemeeter Potato, I may easily switch between CW-regeneration and the original sound of the source (Ebook). This showed that my combination of the three filters leads to a

lovely sound (in my ears) and that the spectrum of the Ebook signal and of the regenerated signal is comparable. In the time domain there are differences (at the falling edge of the DITs and DAHs), but the sound is comparable.

I have tried to keep the orders of the filters relatively low and the Papoulis lowpass helped to achieve a good attenuation at high frequencies (even with a cutoff frequency at 2 * f_s) while keeping the attenuation in the passband low.

...well just experimenting

73

Tom DF7TV