Little games to play with Morse Code in the browser | Hacker News Reader

Little games to play with Morse Code in the browser

alexanderell.is

51 pointsotras21 comments

The timing sounds off (incorrect dot-dash-pause ratio) and the characters are not well shaped keyed, not very pleasant. There are much better options, among others:

https://morsecode.world/international/trainer/trainer.html

https://lcwo.net/

http://www.elkins.org/

Here's a comprehensive list of Morse resources:

https://morsecode.ninja/resources/index.html

Thanks for the feedback about the timing and the other resources! I'm very much not an expert in Morse code (as I'm sure you can tell).

I tried to follow the ratios here (https://morsecode.world/international/timing.html) as follows:

    var DOT_TIME = 300;               // Dit: 1 unit
    var DASH_TIME = DOT_TIME * 3;     // Dah: 3 units
    var SYMBOL_BREAK = DOT_TIME;      // Intra-character space: 1 unit
    var LETTER_BREAK = DOT_TIME * 3;  // Inter-character space: 3 units
    var WORD_BREAK = DOT_TIME * 7;    // Word space: 7 units

But it's interesting to hear that the timing still sounds off and the characters are not well shaped. I wonder if there are some artifacts in the web audio API related to how it's playing that would cause the difference.

Either way, thanks again!

Mh, timing sounds correct now - I listened to it on my mobile earlier and it appeared to have excessively long dots (Firefox Mobile). Cannot reproduce now. Sorry.

You may want to run the output through a low-pass filter of the Web Audio API, or manually "ramp up" and "ramp down" the edges, so they don't sound so rough/hard.

Glad to hear it - I may have snuck in a source code change since you last listened after one of the other commenters suggested the Farnsworth method! Thanks again for the suggestions.

I recently learned Morse code. Long story short, for a while my life took a strange turn where I had full reign over someone else's Airbnb, and was free to modify the place in my image. We found this big light fixture meant for the ceiling, but in a moment of divine inspiration put it out on the fire escape and ran the wire to the house server as our "semaphore". Originally I was going to add server code to translate messages from some API endpoint into Morse. But just to get things off the ground, I set it up as a normal light switch and started keying in messages by hand.

I spent a few afternoons keying out random silly / vulgar messages. Stuff like "- .... . / ..-. -.-. -.-. / .-- - -. - / .-.. . - / -- . / -...", "- .... . / .-.. .. --. .... - ... / -. --- - / - .- .-.. -.- .. -. --. / -.-- --- ..- / .... .- ...- . / --. --- -. . / -- .- -.." and ".. -- / -- . ... ... .- -.. .. -. -.. / .- -... --- ..- - / -.-- --- ..- .-. / -.-. .- .-. ... / . -..- - . -. -.. . -.. / .-- .- .-. .-. .- -. - -.--". At first I had to look at a Morse reference sheet while keying these in, but after a few hours I was no longer looking at the sheet.

Moral of the story, it only takes a few bored afternoons to learn Morse. You should do it.

Your second message would have gotten a chuckle out of me if I'd seen it out in the wild.

I think the second word in the third message is supposed to have a couple of `--.` in place of the `-..`s. And the first one's third word I think is meant to be `.-- --- -. -` (`---` instead of `-` as second letter) instead?

I've been learning the letters of Morse code via Anki for months now (casually, on and off), but I feel like reading them as words here instead of as individual letters cemented them in a different way in my brain. Maybe it's time I graduated on from the letters deck to a words-based deck, it's a pretty fun task too, decoding these.

For me, I'm far better at patterns than memorization. Words are superior to letters because I can often think back to an earlier word/phrase to find a letter. Like, suppose you've done nothing but memorize how to key in the curse words. Starting from "... .... .. -" you're just 4 more letters from learning ".- ... ... .... --- .-.. ." And three away from "-.-. ..- -. -". From there it's just 2 letters to "..-. ..- -.-. -.-". 1 more gets you to "-... .. - -.-. ....". 2 more gets both ".--. . -. .. ..." and ".--. ..- ... ... -.--" (obviously can't have one without the other). At that point you've picked up enough of the alphabet that it's just two more to "-- --- - .... . .-. ..-. ..- -.-. -.- . .-." If you ever forget a letter, just think back to the curse words.

I also like to organize the letters by complexity/similarity. The least complex are the letters that are the same symbol n times. E I S H for (1-4) dots and T M O for (1-3) dashes. Just ask yourself "IS HE TOM?" The next layer up is letters that are n dots followed by m dashes or vice versa. These are subgrouped into length 4 down to length 2 and ordered from left heavy to right heavy. So the taxonomy is [(1,1) A/N], [(2,1) U/G,(1,2) W/D], [(1,3)B, (2,2) Z, (3,1)V/J]. The third group is actually smaller and thus easier. [(1,1,1) R/K], [(2,1,1)F/Q, (1,2,1)P/X, (1,1,2) L/Y]. The final group is just (1,1,1,1) C.

Writing out the table in terms of these similarities helps. Eg I'd often remember letters in terms of other letters that I use more frequently like "k is the opposite form of r (-.- vs .-.)". Or I would know where the letter sat in terms of layer or skew and I knew what letters it couldn't be. Eg for a while I knew N was length 2, I knew it wasnt .. or .- since I use I/A a lot, so it was just a matter of learning the difference between m (--) and n (-.), Which happens to be the same as the difference between w (.--) and u (..-) after you remove the first dot and turn the page upside down. Obviously u is upside down n and w is upside down m. For the occasional clusters of easily confused groups, learning one forces the others to snap into place. .-.. is off balance to the Left, so it must be L (..-.). -.. begins with a Dash so that one is D. .-- has twice the dashes as ..- so that makes it "double the U" aka W. P (.--.) is similar in shape and origin to rho so it makes sense for it to be similar to R (.-.).

That's about as dense as I can compress it. The rest is rote learning.

I had a similar experience learning to send, but learning to copy is significantly more difficult. I can probably send at 20 WPM or more, but only copy (listen) around 10 WPM.

The trick to getting good at Morse code is to learn how the letters (and eventually words) sound, rather than parsing individual dits and dahs. I wouldn't recommend using this on anything but the "fast" speed, and even then, it's almost too slow.

I'll second the recommendation of https://lcwo.net . That's how I learned it, and I was able to get up to about 10 wpm after a couple months of practice.

I believe the technique you're referring to is called the Farnsworth method where the letters are keyed at a faster speed than the overall transmission speed (eg: at 5 words per minute the letters are sent at 13; when learning at 13 the letters are sent at 20).

Another thing about learning Morse code is to NEVER try to learn it visually if plan to send/receive aurally. Before I joined the Navy I had been proficient with Morse code at around 15WPM; I was able to read the flashing light codes being sent between ships but I had to "hear the light." The Signalmen had the opposite phenomenon: since they had learned Morse code visually in order to copy Morse code by ear they had to "see the sound." I'm sure between the fields of neuroscience and psychology there could be an interesting discussion as to how that cross-sensory translation takes place but it certainly does.

I'll second this as someone who NEVER learned Morse code. We did lessons on it time and again during scouts, but, not only was there a near fetishistic emphasis on the dots and dashes, there was also an angry rejection of the idea that rhythm and pauses were a part of the process. Yes, multiple combinations of dots and dashes lead to the same pattern, but you were just supposed to determine the difference from context. The adults would give us a message like the one below and then despair that our generation was full of idiots.

..... .-... -..-- -.--- --.-. .-..- ..

Sometimes they took it further and had us try to decode call signs and license plates, which gleefully removed that extra layer of context.

-.-.- .-.-. .---- ----- --... .---- -..

> Yes, multiple combinations of dots and dashes lead to the same pattern,

Sorry, but this is nonsense.

There should always be character spaces and word spaces to break up the pattern.

Without character spaces, it's not morse!

Some people don’t know what they’re talking about!

If you do want to learn code, try the Ham Morse iOS app.

That's a great perspective. I'm so unaccustomed to parsing it that even the "Fast" setting felt lightning fast to me! I went ahead and added a new "Faster" speed. It may be running into the limits of how precise the Web Audio API can be, but hopefully it should be a little closer to a more proper speed. Thanks!

Edit: after a little more research and experimentation, the Farnsworth method you mentioned is actually an even better experience for both listening and inputting - went ahead and added an option for it, defaulting to on. Great call!

Thanks! And it's great to have more and different games for practicing, so thanks for making this.

This is a fun game - more Morse code games are a good idea!

The slow setting is painfully slow though - that isn't teaching the right sound of the morse code.

So I'd suggest the minimum letter speed should be 20wpm but with spacing between the letters to make it easier to understand when you get up to words.

Note dot time in ms is 1200/wpm if that helps!

PS speaking as someone who can do morse at about 15 wpm now after having studied on LCWO for quite a few months!

Great call on varying the spacing between letters. I added that as an additional option and it's a much better experience for both listening and inputting. Thanks!

I'm seeing a bug where if you interrupt a listening exercise, i.e. by pressing "try again" because you missed something, it starts playing it at max speed. Any ideas what could be causing this?

Ah if this is happening when you try to replay it before it finished, it may be due to timing issues with how the “try again” is implemented. My bet is that it sometimes doesn’t correctly stop the first playthrough before starting the replay, so you may be hearing the two dueling playthroughs on top of each other. If you wait for them to stop, you should be able to try again and hear just one copy of the correct message.

I think it could be fixed with more intelligent stopping (e.g. a strictly increasing play counter that it could check before playing, or something like that). Thanks for pointing that out!

Edit: yup the increasing play counter (and a check everywhere) was enough to avoid the duplicate replays, and it should be better now. Thanks again!

A question: Is there a shortcut for the "Try again" button?

And a suggestion: Pressing that button clears the input, which is annoying when you have partial input correctly entered there.

Control on desktop should trigger the “Try again” for both games, although there’s no similar shortcut for mobile.

Great suggestion about the clearing! I think it would be helpful to listen while you look at your input to verify, too.

Edit: fixed the Try Again button (newly renamed "Replay") so it doesn't clear your input. Good call!

Reply on news.ycombinator.com