What is the most accurate Morse Code decoder tool for converting Morse Code to English text efficiently?

 

Hi, I'm working on a project using two Arduino microcontroller boards.

They communicate by pulsing a laser and measuring the time each pulse takes. The boards will have a display to show what the other device just transmitted.

You can use light or sound as long as you build a device that is able to decode and demodulate an information carrying signal . Morse code is encoded with "dots and dashes" and the modulation scheme is Pulse Width Modulation.

 

What is the most accurate Morse Code decoder tool for converting Morse Code to English text efficiently?

What is generating the Morse Code signal?
What form is the Morse Code?

 

What is the most accurate Morse Code decoder tool for converting Morse Code to English text efficiently?

A big part depends on how the Morse signal is being produced. If it's being produced by a machine, then you have pretty reliable timings and the decoding process is much easier.

But if it is being produced by a human, then there is quite a bit of variability, both within symbols and between symbols. The variability is sufficient that practiced listeners can distinguish and identify specific senders by their "fist".

You also have drastically different basic rates from one sender to another, ranging from under 5 wpm to over 80 wpm.

This is one area where I think a trained AI could possibly do a very good job.

 

I remember seeing a Morse code decoder back in the mid 1980s published in either Byte or Popular electronics magazines.
The circuit employed a microprocessor and a substantial amount of TTL logic. Quite a complex and challenging circuit.
I would think that, at least from component count, there may be newer, simpler solutions based on very powerful microcontrollers and a few other components.

You may want to search Google for such a kit or fully assembled product.

 

The folks at Elecraft have had a pretty good one in their products (e.g. the K3) for over a decade.

 

One issue to overcome I see is the variable WPM rate,
Unlike TTY that is intended to be machine receivable. Morse was originally intended for Human send/receive method.
Sources such as world news press such as Russian Itar-Tass sent by punched tape originally, are one of a few exceptions.

 

There have been quite a few CW decoder projects published by the ARRL in the QST magazine over the years. Also some code for decoding published, a but farther back. A few packages for sending the signals into a PC have also been published.
The means for handling the variations in speed have most successfully been measuring the relative lengths of the elements and the spaces, and then attaching an identity as "dot", "Dash", "element space" or "word space" before presenting it to either a fast look-up table or an evaluation tree scheme. The big challenge now is that current computers no longer support either serial or parallel I/O, and so require a great deal more logic to input simple ON/OFF digital streams. USB is not our friend at all. (the unknowing will challenge that for sure.)

 

It appears that fast and accurate interpretation of Morse code will have a bit of lag because of needing to see a whole string before deciding what it is and what it means, But with a multi-gigahertz processor that delay will not be obvious to a human reader/user.
So the larger challenge will be the electronic discernment of the received signal against a noisy background/ foreground. Decoding a message with a unity signal to noise ratio seems to be a human ability not easily translated to electronic hardware. I can visualize the logic required but not the best way to implement it. And for faster code speeds the challenge increases.

 

It gets even more difficult when you run into an old timer using a hand key or Vibroplex where they tend to emphasize the dash randomly. Or even worse when it's a newcomer trying to learn how to use a hand key. I've seen a lot of morse "readers" advertised but never tried using one so good luck! Morse is getting to be a "lost" art but there are plenty of Extra Class guys still around that can hit 35wpm or better. My brain just won't work that fast...

 

The guy that I first started learning electronics from could send/receive at 80 WPM (using a speed key). I think the world record is about twice this. He said that he didn't hear dits and dahs, but rather was simply hearing a conversation the same way that we don't hear individual phonemes or syllables, or really even individual words, but rather we hear a conversation. It makes sense, but it still seems foreign to me. I got so that I could send at about 20 WPM, but copying at even 13 WPM was a real struggle. I think the best I ever did over a standard length test message was 8 WPM. At 13 WPM I would get the first fourth or so of it and then get hopelessly lost.

 

Keying back to base, somewhere in the Libyan desert.!

 

The scheme I described should work the best, but even a good scheme will require at least somewhat consistent keying. Unfortunately my hearing response is not adequate for decoding Morse code. Some folks just don't acknowledge anything like that.
And probably Basic would not be fast enough to decode it, anyway.
But if a flow chart can be created then a programmer should be able to write the code and then it is just a matter of getting it compiled to run on a PC using ( Iam not sure which) digital input. The ARRL has a circuit that uses a parallel printer port, but those are less common today. And a USB input adds a whole lot of overhead on both sides. That leaves a sound-card input and a keyboard input, neither a first choice as I see it.

 

Google "Morse Code Decoder" and you will find all kinds of ready built hardware packages. No idea what percentage correct hit rate they have but even by ear you learn to fill in the blank spots fairly easily. Google "Morse Code Decoder Software" and once again you will find plenty of available software packages for the computer. I used to be able to get well over 35wpm sending with a set of Bencher paddles and a home brew iambic memory keyer I made. Problem is that I can't read code that fast so if I was blazing away sending CQ someone would come back at the same speed. So, you learn not to send any faster than you can read so you don't look like an idiot. In the yearly international competition, they use receive "images" (letters, numbers, punctuation) per minute. One word is considered 5 characters for the ARRL. The yearly winners are over 300 "images" per minute which is ~60+ wpm. My extra class ham buddy gets this blank sightless look on his face when he's copying high speed code and yes, he doesn't hear it as code but words. The military guys used to sit at the radio and use a typewriter to copy code so they not only had to learn code but also how to type. All that got replaced by teletype and even the ARRL dropped its code requirement for a Novice license.

 

The military guys used to sit at the radio and use a typewriter to copy code so they not only had to learn code but also how to type.

We had to learn it that way also, I once saw our instructor taking 35wps/m, he got up to close a window, sat down and caught up where he left off.
.The Royal Corp. of Signals.

 

By far the very best decoder for Morse Code is an experienced human operator who has good hearing. The very best computer systems do not come close for several reasons: They require a greater signal to noise ratio, They need time to detect and adjust to speed changes, and they are less able to hear whole words at a time. The only advantage some computer systems have is a greater ability to focus exclusively on decoding, and even that advantage is not universal. Possibly in the future artificial intelligence coupled with machine learning might come close. Maybe.
Machine decoding certainly can be done, and without a lot of extra software, machine decoding requires fairly good sending..