Someone brought this up on Discord asking about it. I figured it prudent to repeat what I had to say here as well.
I only got as far as modifying the NATO alphabet to have different ending vowel sounds. I was planning for that to be a diagetic part of a game, where military voice controlled hardware was controlled via menu only and only by this modified NATO alphabet to reduce voice recognition errors, because, in universe, those errors could be deadly.
I had a rough idea of a crashed pilot pinned to their seat via shrapnel and their controls are smashed. A utilitarian drone is strapped into the back, able to be voice controlled. As the pilot, you have to navigate the drone to facilitate your rescue. There would be three voice actors. One acting as both the narrator and the internal thoughts of the protagonist, one of the audio feedback from the drone, and one of the pilot themselves, uses sparingly, say perhaps in the intro where you’d hear their mayday.
If the voice recognition of the game failed to recognize which letter you said, it would diagetically have the drone respond in a very dispassionate, “I’m sorry, Dave. I’m afraid I can’t do that” sort of way, and perhaps the narrator voicing the irritated thoughts of the protagonist, or even the pilot’s voice swearing under their breath, which cues the drone to innocently say they can’t hear them, which invites the pilot to react, you get the idea.
Any voice recognition errors on the player’s part gets folded diagetically into the story. That’s as far as I got before life rather rudely intruded and made me drop it like it was hot.
The idea was to borrow how *Lost Pig* used its protagonist to diagetically disguise parser failures as part of the story, as Grunk no understand. I figured if it worked for Grunk, I could make it work for janky voice recognition as well.
Here’s the NATO Alphabet:
I no longer have my altered version, it was lost on my smashed laptop in the crash. I don’t fully recall it either. But it basically altered the words to avoid reusing the same end sound.
So instead of Bravo and Echo and Kilo and Tango… …you had Bravo, Echeye, Kilos, and Tangle.
You had to be careful. You couldn’t just carelessly go with, say Tangor, because you already had Victor, and now you accidentally created a new redundant ending sound. I had it worked out so each had a unique ending syllable, in addition to the already fairly unique beginning syllables. In theory, that should make the spoken words even more distinct and useful for a voice recognition system in war conditions, with static and interference and gunfire and munition explosions going on in the background.
It was this thread that originally got me thinking in this direction. You can probably see that now if you read my posts with this information in mind.