you joke, but looking at old cyrillic handwriting, I've literally seen both "т" and "ѕ" written "ꙅ", "k" written as "ιι", as well as about three different versions of every letter (ж being the second worst offender behind к) (и goes from being "n"-shaped to being "u" shaped so н can be "n"-shaped instead of "N"-shaped)
2^30 is actually not all that many, relatively speaking, only about 1.07 billion. You could write a computer program that looped over every possible configuration of pixels in a 6x5 grid, and still have it complete in a fairly reasonable amount of time, as long as you weren't doing anything too expensive with each pixel configuration.
This reminds me of recently when I was attempting to fit a lot of text in a tiny amount of space. While trying to take it to the absolute extreme, I made each letter 2x2 pixels. This means there were only 16 possible combinations, so some letters had to look the same. Knowing what the text was, I could have sworn I could still read it, but no one else had any idea what it said.
I've done something similar, but not quite: I took the few characters I could find of a 3x3-per-character font in a rhythm game and expanded upon it. Writing with it leaves behind sometthing near illegible for even me until my eyes finally adjusted to recognition, and even more so if you don't leave a space between each row of text, that became half the joy of the project by the time I considered myself finished with it. The other half being the joy of finally getting a larger project of my own finished despite a crippling lack of motivation. There's no capital-lowercase distinction, since I had to take advantage of either form of each character to fit the 3x3 requirements, but I did manage to make an & symbol. :) And the number 5, despite it's distinctly hindering configuration.
Where the fuck do you speak that? The name looks very Bavarian, so I'd assume somewhere there?? I only recognized it as some dialect of German because Fichtenholz-wasauchimmer, Karotte and Weizen-kerne(?)
Honestly, I was surprised at how much of the minecraft text I could decipher!! It really shows us how our brains don’t go through each letter when reading a word… and when the shape of the entire word is somehow recognisable, we can still read it!!
I used to tutor a boy with dysgraphia who took this to extreme levels. Almost half of the letters he wrote looked _exactly_ the same. Literally, his n, m, u, h, t, o, c, e, v, b, l, s, and w all looked like something halfway between u and n. It was honestly kind of impressive.
I had a very similar problem in my handwriting (though with Cyrillic script that I use more often), and I had to systematically change the way I write every single letter until they became readable. I specifically tried to see how much I can stray from the grapheme forms so that I had more distinctive lines for each of my letters, while still having it possible for others to understand It was one hell of an exercise
Latin letters: er Ɛɛ ir Řř or Ɔɔ ur Ůů Ergative(Ɛgative) Invader Zim(Invadɛ Zim) Irken(Řken) Girls(Gřls) Orphan(Ɔphan) Translator(Translatɔ) Urinate(Ůinate) Yogurt(Yogůt)
I absolutely love how when you showed a cyrillic alphabet, I, as a native ukrainian speaker, didn't catch like half of those vile symbols and some that I know were missing. Absolutely cursed.
@@piotrbojkoff Most of them are completely legible and familiar to me, even some of the older ones like ѣ, but the last row is absolutely cursed. I've read some old texts in my time, and even a little of (newer) church slavic. But that last row is really alien to me, as well as that O and V mix and some in the third row.
@@piotrbojkoff Good to know, but beware saying that Ukrainian or Belorussian are dialects of Russian. Old Russian? Sure. Russian? No way -- by the time of more modern Russian variations we had our own languages. And as for Greek characters, I should've mentioned that they're still common and recognisable on their own to avoid wasting your time, sorry.
Can confirm. Each pixel can be black or white, independent of the other pixels, so there are 2 options for the first pixel, times 2 options for the second pixel, and so on until you end up multiplying 2 by itself 30 times.
I'd say, that there are no distinct boundaries of acceptability, it's more of a spectrum, affected by one's personality, mood, health conditions and weather in Slough. upd: it also might be cool to make a map of all symbols based on the Levenshtein distance between their pixel representations.
I have an idea on how to do this - Agma pointed out that every symbol in Minecraft is bound to a 6×5 pixel grid. We can create a 30-digit binary string - 0 denoting no fill-in, 1 denoting a filled pixel. We can then simply attach a 30-digit string to every Minecraft glyph, and find the Levenshtein distances between said strings :)
@@TokuMGTT Yeah, that's almost exactly my idea! It's just that in this case it would make a bit more sense to get Hamming distance and it wouldn't allow for visual translation movements. What I'm interested in is the smooth movement from one letter to another (with a bit of addition and subtraction, mb), akin to 3b1b. Actually, come to think of it, if you are only interested in the difference, you could take XOR of symbols. Thank you for showing interest in my idea!
@@the_multus idea: make each letter with a certain amount of Bézier curves, then calculate how much you have to massage one letter to get to another somehow (??)
The minecraft font is, for most characters, 5x7 -- I'm pretty sure that some glyphs (at least for those outside of the ASCII range) have ascenders, descenders, and diacritics that extend outside of that zone, though.
Even some ASCII letters, such as the lowercase j, are 5x8 due to having a descender, and @ and ~, which are also part of ASCII, are both 6 wide. This video uses the old font from Java Edition 1.12 and prior, where characters can be 8x8 at most. In the new font for 1.13 and later, characters can even be 9x12. For characters unsupported by the standard font, the unicode font has a character size of 16x16, but is also displayed smaller.
yeah I make pixel fonts on the side and everybody gangster until they have to design capital letters with diacritics. I normally just bake in like 3 pixels of dead space above letters reserved for diacritics and some punctuation. It also makes the default line spacing a little nicer
A couple of things: 1. You picked a fairly legible text for your thumbnail. I could read it without too much thought, so good job on that one. This is likely due to the high amounts of 's's and relatively unchanged 'i', 't', and 'l' and the newly opened 'o' (IPA helping legibility, baby!) which allows for pattern brain to pick up on it. 2. I don't remember if I mentioned this or not in my submission for the CCC2, but I intended to use a logography that has multiple words mapped to the same logograph, like a reverse Chinese. For example, but certainly not practice, there could be a word which has the same logograph as a different, inflected word. In hindsight, I think I would have called the writing system a "featural logography", but that's really just a flavor thing.
small correction at 11:11 binary situations such as this have 2^number of spaces combinations, so it would actually be 2^30 (a small amount over a billion if I did my math right) combinations
In my previous comment, I calculated the permutations based on a 5x7 pixel grid as observed in the video, which resulted in 34,359,738,368 permutations. However, as another user (@YatYusYae) mentioned, the font sizes in this version of Minecraft can actually be up to 8x8. This adjustment means the true number of permutations for the font being used would be 18,446,744,073,709,551,616. This figure includes numerous permutations with empty spaces in the boundary layers.
This might just be me, but the changing light levels throughout the video end up being very distracting. There should be a setting on your camera to lock the exposure level, you might want to set that to a specific amount instead of auto? Just might help with shot consistency. Super interesting topic!
I found the background fakescreen meme a distraction in itself, and I somewhat made a correlation between the meme changes and the light level. That, with a blue dot-like effect on the screen. (I ended up hiding that part of my screen with a paper.)
the german FE-Schrift font used for number plates all over the world is specifically designed so that you cant modify it to easily turn one letter into another (like a B to a P etc using black and white tape), its pretty interesting
Speaking of heterography, compact fonts, and Minecraft, EthosLab came up with an elegant way to invent your own alphabet in Minecraft using the 4 stair rotations and 2 slab positions, which when arranged in a 1*2 block configuration, gives you just enough variations to fit all 36 figures and letters of the alphabet in a compact manner. He demonstrates it in his HermitCraft S9#10 video at 7:19. Also, a 30 binary pixels font wouldn't amount to 30! possible combinations, but rather 2^30, or about 1 billion. 30! would describe the different possible ways to sort 30 different items in a sequence, whereas the Minecraft font corresponds to the different possible ways to keep 30 light bulbs powered or unpowered.
I tried to read the test in the thumbnail, and my brain _obviously_ processed it as it would normal words, and my inner voice read it like some wizard chanting a spell.
Great video! Just wanted to point out a couple of details for accuracy. At 10:50, you mention the pixel density being 5x6, suggesting an area of 30. But at 12:15, it's evident that the minimum space is actually 5x7, which brings the area to 35. Also, at 11:12, while discussing the permutations of images, you correctly identify it as a binary calculation. However, instead of using factorial, for binary strings, we use 2 raised to the power of N (where N is the string length, here 35). This means the number of permutations for the pixels in the image is actually 34,359,738,368 - a significant difference from 30!. Keep up the good work, just thought these clarifications might be helpful.
In my previous comment, I calculated the permutations based on a 5x7 pixel grid as observed in the video, which resulted in 34,359,738,368 permutations. However, as another user (@YatYusYae) mentioned, the font sizes in this version of Minecraft can actually be up to 8x8. This adjustment means the true number of permutations for the font being used would be 2^64 = 18,446,744,073,709,551,616. This figure includes numerous permutations with empty spaces in the boundary layers.
I’ve never been more surprised and slightly confused for how I was able to read the thumbnail sentence so easily, having never seen this channel before. Excited to fall down a binge session 🙂.
I think I found your heterography, for numerals at least, by accident. When I worked booking-in samples from a clinical trial with handwritten dates on their labels, which I had to compare to a typewritten manifest, I swear I saw every single numeral 0 to 9 rendered in bad handwriting in such a way as it could be mistaken for any other digit. I honestly wish I'd made copies of all the worst ones.
If "hetero" is a good antonym of "ortho", then the opposite of an "orthonormal basis" (a concept seen in vector spaces with an inner product, a notable example of which is the usual Cartesian basis for Euclidean vector spaces) is a "heteronormal basis". We can analyse this further by noting that a _non-orthonormal basis_ isn't necessarily _the opposite_ of an orthonormal basis. Instead, we must directly counter each part of the definition. - _Ortho-_ refers to the basis vectors being _orthogonal_ - meaning their inner product is 0 (for Euclidean vector spaces, they're at 90° to each other). The opposite of _orthogonal_ is _parallel_ as parallel vectors have no components that are orthogonal with each other. It is worth noting that a basis of vectors which are all parallel can only ever describe a 1D line, since vector-vector addition doesn't affect the angle of the vectors, so two parallel vectors will always sum to another parallel vector. - _Normal_ refers to the basis vectors being of magnitude (for Euclidean vectors, length) 1. The opposite of this could either be magnitude 0 or perhaps magnitude ∞. The latter is not considered a number in the usual Real numbers, but can exist in _projective geometry_ where infinite points act as the 'points of perspective' one finds in perspective art. Hence: _Heteronormality_ implies an inability to work with more than a single dimension at a time, and a magnitude of 0. The only way to achieve a non-zero magnitude is through projection. In both cases, even the single dimension that did exist is reduced to only one trivial point. QED: Heteronormality is lame, LGBTQIA+ rights are cool. The mathematics says it all.
To get every English letter in distinct form including upper/lower distinction, you only need a 4 or 5 wide and a 4 high pixel grid - 5x5 if you want things to be very clear for letters like M, W and E.
This is pretty sick, i like it. Also, I'm one of the reasons you jumped to 10k suddenly. I don't follow a lot of linguistics channel but your original CCC got algo'd a little bit so i saw it. I mean, i saw like a third of it, I'm still working my way through.
I created a language/orthography called ■● [/sə/] which consisted solely of the two phonemes. The consonant was written as ■ and the vowel as ●, and anything else (i.e. punctuation) was▲
I think a more general approach is understanding that your goal is the reduction of redundant information. Consider the 7-segment display. Did you know you can make segmented displays for numbers use more than 7 segments? They can be stylized to be more interesting and legible than 7-segment. And did you know you can use fewer than 7 segments? There are 10 digits, which is greater than 2^3, but not greater than 2^4, so you need at least 4 segments to technically distinguish the 10 digits. There's a video by 'Posy' called "Segmented Displays" which runs through various segmented displays. Apparently you can achieve reasonable legibility on numbers with as few as 6 segments, while legibility suffers greatly with fewer still. 92 characters is more than 2^6, but fewer than 2^7, thus you should only need 7 distinct graphemal features to compose all characters. Segmented displays require gaps between distinct segments, whereas in writing, the overlap may grant more legibility. You might note further that our goal is not preventing characters from being read, but from them being distinguished. Furthermore, if we consider the 13375p34k dialect, and only concern ourselves with one case (upper or lower) and the digits, we only need 26+10-10=26 characters, which is not greater than 32. (In 13375p34k, both I and L are replaced by 1, O is replaced by 0, B may be 8, G may be 6, and R may be 2, along with other rules noted in the proper name of the graphical dialect. Indeed, only 9 has no letter, while 1 has two, effectively removing the contribution of the 10 digits from graphemal information complexity.) 5 segments is very difficult to make numbers for. Let's try making the numbers *and* alphabet in only 5 distinct figures!
Thank you for putting the definitions for the jargon, ive been super interested in conlangs and linguistics for a while but sometimes the stuff just goes over my head
I thought I was a language nerd, but this is proof that I have a long way to go before I achieve master form. Also, Hyperlegible made me think of a game I just played recently: Syrup and the Ultimate Sweet. It's a short, cute game, obviously made on a low budget, and so I was absolutely charmed when they included a font for dyslexics (like me). I definitely didn't expect it, but it was very welcome!
I'm surprised that there was no mention of cursive Cyrillic. Its basically a naturally occurring heterography. Seriously, how does anybody read that squiggly mess?
My name is Yoshikage Kira. I'm 33 years old. My house is in the northeast section of Morioh, where all the villas are, and I am not married. I work as an employee for the Kame Yu department stores, and I get home every day by 8 PM at the latest. I don't smoke, but I occasionally drink. I'm in bed by 11 PM, and make sure I get eight hours of sleep, no matter what. After having a glass of warm milk and doing about twenty minutes of stretches before going to bed, I usually have no problems sleeping until morning. Just like a baby, I wake up without any fatigue or stress in the morning. I was told there were no issues at my last check-up. I'm trying to explain that I'm a person who wishes to live a very quiet life. I take care not to trouble myself with any enemies, like winning and losing, that would cause me to lose sleep at night. That is how I deal with society, and I know that is what brings me happiness. Although, if I were to fight I wouldn't lose to anyone.
Faffing about with orthography I devised a sci-fi font made only of straight borders, corner-to-corner diagonals, and two dots, one per center of the top and bottom halves. The whole uppercase alphabet represented by 8 morphemes; the idea being they can be scanned instantly by space ships at a great distance like a barcode, while still being readable enough for a human viewing it directly. There would be some wierd compromises though, while A would look like a tilted Atlantis A without the crossbar, other letters like P would be an upside-down L, and R would be that but with a semi-centered dot hovering by its base as a reference to its missing half-diagonal that the font doesn't support.
@@AgmaSchwa I'm not going to lie, the lighting change happens throughout the entire video, and I'm watching in a relatively low-lit room... I thought you were intentionally directing me to look at the computer :)
Reminds me of what would happen during certain (now fortunately resolved) severe anxiety spikes several years ago during which letters and would start to warp/contort in demented ways due to my pattern recognition going haywire from subconsciously trying to visually locate non-existant threats. Reading any text I wasn't already familiar with was next to impossible and even sentences and phrases I encountered regularly remained only barely decipherable. All that to say that from experience actively having to read this kind of thing in real life sucks horribly (especially when your job requires reading important disclosures verbatim on a recorded line) and is monumentally cursed. 0/10 then but 10/10 here, can't wait to see what other unholy abominations people are inspired to make from this haha.
Oof anxiety is no joke, it can really bring the brain to some extremes. I know under enough anxiety I can't physically make myself talk, and even lost motor controls and had to be practically picked up off the ground, I wish more people would get how much severe anxiety can impair a person. I hope you never have to be in that state again
The minecraft font's dimensions are 7x5, not 6x5. If you look at the "af" on the screen, you can count the 5 rows that make the a which reach up to the dash on the f, after which another 2 rows are needed to make the upper arch on the f.
The first message was pretty clear up to the last word. "Hey everybody, welcome to n???" the last two letters both LOOK like N but can't be, because N was already used. It took a bit of puzzling to figure out that your username was InsaneInbox, but once I had that I could figure out the rest more easily. The second message starts, obviously, with "My name is Yos" but is more or less illegible from there. The third message, after staring at it for a bit, I realized, was "The square root of 49 is 7" which gave me a lot more letters to work with. Now that I have those letters, I can go back to message two and see if I can puzzle that out. "My name is Yoshikage Kira" I think? The first message, now that the second one gave me a G and the third one gave me the U and H, says "Hey everybody, welcome to nguh" That was fun to figure out! I'm liking this whole alternate alphabets thing. 20:09 Now that I've been staring at this alphabet for a while, it's immediately obvious that the carrot's name is "Karotte". You switched to German.
Looking at the alphabet at 20:42, the last word of the first line appears to be “nguh”, yes, and the second line ends in “Yoshikage Kira”. Also, you didn’t do the 4th or 5th messages, so I’ll decode those too. The 4th is “do you have any idea what i am capable of?” And the 5th is “thanks for watching!” The only difference between “!” and “?” is that “?” is shifted one pixel to the right compared to “!”. Someone else can decode the text on the screen at 21:53, I won’t do it yet, I’ll do it tomorrow if it isn’t done by then.
... a bang, but with a whimper light fades on this feable attempt at... (spelling not mine) ...s the next one will be kinder to ... way, we lived ...at is beautiful ... be not, and it will all play out the game ... x dee return to monke ...ush me on the cursed conlang circus ...ing on it, I swear. Just give me and ... I think this is all right
the objectively best -eme is the one that best performs your purpose in communicating to your audience. If you want minimum ambiguity that means one thing (sans-serif on a screen and maybe serif fonts in print but who knows), and if you want to stack up entendres maybe you want to commit horrible crimes like this video.
Ooohhh ideasss! Idea a. Make the pixel font as small as possible, so that each letter is naturally only one pixel off from a bunch of others. You would need at least 7 pixels (2^6 < 90 < 2^7), but since 7 is a prime number and wouldn't form a rectangular grid, either a 2x4 grid or a 3x3 grid should work. Idea α. Reshape the 7-segment display so all characters can be shown. (Relatedly, I remember watching a video trying to reshape the segments so all 10 digits could be displayed with only 4 bits of information.) Idea א. Break up each letter of the alphabet into "strokes", like how Chinese characters can be broken into radicals. I tried this for the lowercase letters in my cursive style, and got it down to a list of 13 strokes. These have been swimming around in my brain since I first watched the video a week ago, and by now I've filled one page of physical paper, one page on the Notes app, and two pages of spreadsheets exploring ideas ay and aleph. Anyways, yeah. Cool video.
Oh I love making alphabets like this I’ve made one where I take away one line from each letter I’ve made one where they are variations on a weird spiral shape (a couple letters didn’t work like that)
3:53 I don't think anyone argued about the number in the comments yet, but there's actually 94 under this definition - there are 32 punctuation symbols on the US keyboard. 10:45 the Minecraft font is 7 pixels tall, not counting descenders, and others have pointed out this makes the number of combinations 2^35.
what's funny to me is ive never actually been able to register that uppercase cursive G as actually being a g, i always write it differently because it looks unrecognizable as g to me
It is more like a heavily stylized lowercase g, which is what makes it confusing, because that's the last thing you expect to see when you start off with basic script, where uppercase G looks like this
what that strange article puts forward is actually pretty intuitive and freeing. essentially, there are no rules to break; everything, like all things it seems, more and more, is an infinite matter of unknowable prerequisites only seeming to enact, or perhaps truly enacting themselves in the present, in the becoming, through life and conscious memory
insaneinbox: Hey everybody, welcome to nguh insaneinbox: My name is Yoshikage Kira insaneinbox: The square root of 49 is 7 insaneinbox: do you have any idea wgat i am capable of? Seems like you picked, a very weird language to Try and figure out what it is.. It's some form of Swabian.
I really liked this, subjects like this are fascinating to me! I'm subbing. One piece of constructive criticism I have is that it was pretty difficult to focus on you speaking when the video brightness kept shifting every time you moved. You seem to have auto-exposure locked On on your camera, and it might be better to set the exposure once each time you set it up, then keep it on that setting while you film.
Am I the only one that could pick up many words? But yeah, I loved the idea. I'd do as such: I would challenge myself to only change x pixels per character, increasing x in every iteration. What is the minimal x that screws everything real bad?
If only I knew every channel that gets 10K subscribers because it provides massive content (squeegee cleaning every corner of a topic so much tht if it were a black hole, every fold in the blanket of space time would be spotless), but the World and Himanity is not ready for - which, consequently, causes the YT algorithm to disfavor such releases… yeah there needs to be a specific home to YT for social intellectuals
Things that I didn't expect to not be great to watch before bed: It's like the uncanny valley for text, its like I lost the ability to read my native language, I can imagine having a weirdly specific nightmare of this.
17:56 “Hey everybody, welcome to ___” “My name is Y____ ___” “The square root of 44 is 7” (I know it’s ‘49,’ but I originally read it as 44) “Do you have any load what I am capable of?” (again, ‘idea,’ but my brain originally read it as load. in hindsight this sentence was pretty easy once I glanced at ‘capable’) kinda crazy how some of the common words the brain instantly fills, then immediately fizzles to a stop on some of the others because it just couldn’t find one that made sense with the ‘letters’ being presented. (‘course, I sat through the sped up footage of the editing process, so it’s possible some of that my brain held on to enough to remember for later.)
My favorite allographic variation has to be lower case Greek sigma, which is commonly written as , but turns into when word-final. Example: γνῶσις, gnosis; 'knowledge'. What I find so interesting about it is that unlike the orthographic allographs you mentioned, which display allography based on the lexeme itself, lower case sigma comes in a complementary distribution based upon its position in a lexeme. You could write down a rule for sigma (using || for graphemes and for graphs, something like |σ| -> / _#), which is something I'd find quite impossible to do for the orthographic allographs. Capitalization at the beginning of a word seems to also go into this direction, but is also different, as it is both positional and based on the lexical category of the word (proper nouns in English or all nouns in German) in which it occurs. Not to mention sentence initial capitalization. Furthermore, in contrast to the allographs in free variation (like the different fonts), in this case it seems to me that there is an underlying grapheme |σ|, which is realized as the graph when at the end of the word, as the graph at the beginning of proper nouns and sentence initial words, and as the graph otherwise.
this randomly popped in my reccomended then i noticed you're the dude that made that hunger games sim that everyone puts characters or their friends in so thank you for this information and that wonderful website
I just want to show my appreciation towards the different images you have showing on the desktop behind you, it really adds to the perceived quality of the video, which feels like that of a million subscriber channel. Especially the one at 7:45, I found that particularly funny, LMAO.
I used to use the Minecraft font so much that I would memorize it so I could draw it in MS Paint without having a font installed also aren’t the letters actually 7×5? the ‹o› is 5px tall, and the ‹b› extends 2px higher?
In a game Axiom Verge there are notes written in an "alien" language that is actually just english with very deformed letters. When I played I spent some time learning to read this deformed alphabet by recognizing familiar words. Later in the game you find an item that translate those notes to the standard english...
Do you know that -7 is also a solution to sqrt(49)? Yeah, now you have an idea what I am capable of! When a entire country learn cursive, but few write it well, you end up with most population to "decode" what was the meaning of the sentence. Plus, including grammatical, conjugation and other errors, you really have to be deductive. A lot of teacher in my country would not struggle for too long to play with your font. Heck, give me 20 minutes to be used to it, and I pretty sure I could play as if I was just a slow-reading child.
(im late but) -7 is not equal to sqrt(49) as the "sqrt()" function/operation only takes the principal root (the positive one) however, x=-7 is a solution to x^2=49 along with x=7 but not "-7 is a solution to sqrt(49)" i feel like i couldve explained this better but im supposed to be asleep rn lmao edit: if i missed that this was supposed to be a joke i apologize in advanced as i am very sleepy zzz
i've had to change the way i write two letters over the years, and done a third unintentionally. the first one i did around 5th grade, and it was lowercase y. i wrote it like a sans serif font with two strokes and it was hard to tell apart from lowercase x, so i decided to adopt the hook of a lowercase g like i saw some other people doing. then in 10th grade i didn't like that my lowercase t was hard to distinguish from a plus sign and that the first stroke frequently trailed into the second, so i started writing it like a sans serif font with a hook at the bottom. when i was quite young, i wrote lowercase a like a sans serif font, but i started writing it without the top hook when i got into school. and recently i've also started occasionally dropping the cross bars from capital i for writing speed and when i don't have much room
This one it would be great for someone to make a program that changes each letters, minimizing both the count of bit flips and the confidence level when shoved into OCR. That way the surprise is on you. I would like to do that, but I don't think I got any free OCRs that can reliably detect pixel glyphs...
As someone who loves seeing pixel fonts that attempt to infuse a recognizable style while keeping legibility all at really low resolutions this was painful, but I like it, it's definitely interesting to see where is the breaking point. It reminded me of when I struggled really badly at creating a functional font while only giving 4 by 4 pixels to every character, specially cos I wanted to cover some diacritics.
i think i caught something wrong with the math you did. In the minecraft font every pixel can either be on or off, since there are 30 total pixels that means there is 2³⁰ different combinations of pixels, a number way smaller than 30!
Your comments on the connection between the shape of letters being directly associated with the sounds they make (around 9 mins) cause me to think about how current AI works to understand images as an input, and make me wonder if this sort of thinking could how AI training models might be able to interact with each other to make it easier to get what you want out of the system.
Agma, all you did was create the most legible Cyrillic cursive ever known.
real
Truly
I did think it looked kind of what it would be like to look at Cyrillic while on acid...
@@tankermottind this sounds like you have experiences of looking at the Cyrillic alphabet whilst on acid.
you joke, but looking at old cyrillic handwriting, I've literally seen both "т" and "ѕ" written "ꙅ", "k" written as "ιι", as well as about three different versions of every letter (ж being the second worst offender behind к) (и goes from being "n"-shaped to being "u" shaped so н can be "n"-shaped instead of "N"-shaped)
11:11 If it got 30 spaces it would be 2^30 combos not 30!
since every pixel can be in 2 states, you multiply 30 choices between 2 states
Yeah that was a certified math blunder on my part, lmao
Also, it looks like the graphemes are contained in 7x5 rectangles not 6x5 lol
was just about to comment this thanks for my sanity
Bro you were like “i’ll overdub myself if i make a mistake” but you didnt plan for if the overdub had mistakes lol
2^30 is actually not all that many, relatively speaking, only about 1.07 billion. You could write a computer program that looped over every possible configuration of pixels in a 6x5 grid, and still have it complete in a fairly reasonable amount of time, as long as you weren't doing anything too expensive with each pixel configuration.
So you basically just showed everyone what dyslexia looks like. Which is a quite valuable input into general accessibility education. Wow, bravo
dyslexia isn't seeing letters as different letters, its not being able to make sense of or understand letters, numbers or understand what each word is
@Poland4life
Yeah, I think he did that. Exactly what your saying, I think it did.
@@lavabiscuits but dyslexia doesn't make letters just change shape. you won't look at an o and it'll suddenly become horribly deformed
@@Poland4life I know, but it gives a good concept for non-dyslexic people. To see "text" but not know what it means. Or am I wrong in this thought?
@@lavabiscuits yes, you are wrong
It's like the enchantment font escaped from the enchanting table and cursed the entire game.
Was thinking the exact same thing.
It's the Standard Galactic Alphabet, seriously, look it up
@@webpombo7765 I didn’t know people didn’t know this XD
I'm just happy to see nguh finally embrace his destiny as a minecraft youtuber.
Thought I was going crazy trying to read that
hey, be careful with that
Vötgil
The fact that the monitor in the background has a different meme every single damn cut is what firmly sets this video into being a masterpiece
This reminds me of recently when I was attempting to fit a lot of text in a tiny amount of space. While trying to take it to the absolute extreme, I made each letter 2x2 pixels. This means there were only 16 possible combinations, so some letters had to look the same. Knowing what the text was, I could have sworn I could still read it, but no one else had any idea what it said.
I've done something similar, but not quite: I took the few characters I could find of a 3x3-per-character font in a rhythm game and expanded upon it. Writing with it leaves behind sometthing near illegible for even me until my eyes finally adjusted to recognition, and even more so if you don't leave a space between each row of text, that became half the joy of the project by the time I considered myself finished with it. The other half being the joy of finally getting a larger project of my own finished despite a crippling lack of motivation. There's no capital-lowercase distinction, since I had to take advantage of either form of each character to fit the 3x3 requirements, but I did manage to make an & symbol. :) And the number 5, despite it's distinctly hindering configuration.
the language you switched to was Allgovian German apparently, known as Oschdallgaierisch in minecraft
Where the fuck do you speak that? The name looks very Bavarian, so I'd assume somewhere there?? I only recognized it as some dialect of German because Fichtenholz-wasauchimmer, Karotte and Weizen-kerne(?)
@@kleingrrmpf it's the dialect spoken in Allgäu in Bavarian Swabia
I did know what it was but I knew it was Germanic bc it looked like it said "Karett" with the k. That's sick though!
@@FabsterCola "Bavarian Swabia" okay, haha, very funny, now seriously what's the name of the place it's from?
@@the_real_Kurt_Yarish that is, unfortunately, the true name of the region.. Bayrisch Schwaben. greetings from Munich.
Honestly, I was surprised at how much of the minecraft text I could decipher!! It really shows us how our brains don’t go through each letter when reading a word… and when the shape of the entire word is somehow recognisable, we can still read it!!
I used to tutor a boy with dysgraphia who took this to extreme levels. Almost half of the letters he wrote looked _exactly_ the same. Literally, his n, m, u, h, t, o, c, e, v, b, l, s, and w all looked like something halfway between u and n. It was honestly kind of impressive.
I knew several girls who would always add an extra arch to both n and m.
@@bacicinvatteneaca That's how cursive works, so they probably took it from there
@@-starrysunrise-2908 cursive adds a leading arch to connect to the previous letter, not a full arch that touches the ground.
I had a very similar problem in my handwriting (though with Cyrillic script that I use more often), and I had to systematically change the way I write every single letter until they became readable. I specifically tried to see how much I can stray from the grapheme forms so that I had more distinctive lines for each of my letters, while still having it possible for others to understand
It was one hell of an exercise
@@alsy6813 Cyrillic is too weird. Лишили Лилии. Like, bruh…
Latin letter Infinity §¤
Capital §
Lowercase ¤
Pronounced:Inf
Examples:
Information(§ormation)
Steinfeld(Ste¤eld)
s§ul co§lation
@@_profile sINFul coINFlation
Lowercase not capital!
Latin letters:
er Ɛɛ
ir Řř
or Ɔɔ
ur Ůů
Ergative(Ɛgative)
Invader Zim(Invadɛ Zim)
Irken(Řken)
Girls(Gřls)
Orphan(Ɔphan)
Translator(Translatɔ)
Urinate(Ůinate)
Yogurt(Yogůt)
Hello I'm typing my new alphabet with the 27th letter Inf!
The quick brown fox who jumped over the lazy dog ¤inite times!
@@bonzibuddyauttpi hate it
I absolutely love how when you showed a cyrillic alphabet, I, as a native ukrainian speaker, didn't catch like half of those vile symbols and some that I know were missing. Absolutely cursed.
As a non-native Russian speaker, it absolutely is cursed.
As a native russian i immediately thought of those 3 minutes of native language lesson in like 5th grade with mentions of cyrill and methodius
@@piotrbojkoff Most of them are completely legible and familiar to me, even some of the older ones like ѣ, but the last row is absolutely cursed. I've read some old texts in my time, and even a little of (newer) church slavic. But that last row is really alien to me, as well as that O and V mix and some in the third row.
@@piotrbojkoff Good to know, but beware saying that Ukrainian or Belorussian are dialects of Russian. Old Russian? Sure. Russian? No way -- by the time of more modern Russian variations we had our own languages.
And as for Greek characters, I should've mentioned that they're still common and recognisable on their own to avoid wasting your time, sorry.
Idk where you pulled the factorial from but the actual number of combinations for 30 pixels, black or white, is 2^30, 23 orders of magnitude smaller
Can confirm. Each pixel can be black or white, independent of the other pixels, so there are 2 options for the first pixel, times 2 options for the second pixel, and so on until you end up multiplying 2 by itself 30 times.
Was about to comment this
I've been watching too much VSauce lately
I wonder what is something where the possibility space would actually be 30! 🧐
@@emilyrln
If you had to uniquely order all 30 pixels there'd be 30! ways of doing that.
I'd say, that there are no distinct boundaries of acceptability, it's more of a spectrum, affected by one's personality, mood, health conditions and weather in Slough.
upd: it also might be cool to make a map of all symbols based on the Levenshtein distance between their pixel representations.
I have an idea on how to do this - Agma pointed out that every symbol in Minecraft is bound to a 6×5 pixel grid.
We can create a 30-digit binary string - 0 denoting no fill-in, 1 denoting a filled pixel. We can then simply attach a 30-digit string to every Minecraft glyph, and find the Levenshtein distances between said strings :)
@@TokuMGTT Yeah, that's almost exactly my idea! It's just that in this case it would make a bit more sense to get Hamming distance and it wouldn't allow for visual translation movements. What I'm interested in is the smooth movement from one letter to another (with a bit of addition and subtraction, mb), akin to 3b1b. Actually, come to think of it, if you are only interested in the difference, you could take XOR of symbols.
Thank you for showing interest in my idea!
oh no not slough
@@the_multus idea: make each letter with a certain amount of Bézier curves, then calculate how much you have to massage one letter to get to another somehow (??)
Also maybe looking into signed distance fields might help?? Maybe not but it came to mind
The minecraft font is, for most characters, 5x7 -- I'm pretty sure that some glyphs (at least for those outside of the ASCII range) have ascenders, descenders, and diacritics that extend outside of that zone, though.
Even some ASCII letters, such as the lowercase j, are 5x8 due to having a descender, and @ and ~, which are also part of ASCII, are both 6 wide. This video uses the old font from Java Edition 1.12 and prior, where characters can be 8x8 at most. In the new font for 1.13 and later, characters can even be 9x12. For characters unsupported by the standard font, the unicode font has a character size of 16x16, but is also displayed smaller.
yeah I make pixel fonts on the side and everybody gangster until they have to design capital letters with diacritics. I normally just bake in like 3 pixels of dead space above letters reserved for diacritics and some punctuation. It also makes the default line spacing a little nicer
A couple of things: 1. You picked a fairly legible text for your thumbnail. I could read it without too much thought, so good job on that one. This is likely due to the high amounts of 's's and relatively unchanged 'i', 't', and 'l' and the newly opened 'o' (IPA helping legibility, baby!) which allows for pattern brain to pick up on it.
2. I don't remember if I mentioned this or not in my submission for the CCC2, but I intended to use a logography that has multiple words mapped to the same logograph, like a reverse Chinese. For example, but certainly not practice, there could be a word which has the same logograph as a different, inflected word. In hindsight, I think I would have called the writing system a "featural logography", but that's really just a flavor thing.
> Different words mapped to the same logograph.
So, essentially, Japanese
2^30, not 30!
small correction at 11:11
binary situations such as this have 2^number of spaces combinations, so it would actually be 2^30 (a small amount over a billion if I did my math right) combinations
I was gonna comment this as well
In my previous comment, I calculated the permutations based on a 5x7 pixel grid as observed in the video, which resulted in 34,359,738,368 permutations. However, as another user (@YatYusYae) mentioned, the font sizes in this version of Minecraft can actually be up to 8x8. This adjustment means the true number of permutations for the font being used would be 18,446,744,073,709,551,616. This figure includes numerous permutations with empty spaces in the boundary layers.
This might just be me, but the changing light levels throughout the video end up being very distracting. There should be a setting on your camera to lock the exposure level, you might want to set that to a specific amount instead of auto? Just might help with shot consistency. Super interesting topic!
I found the background fakescreen meme a distraction in itself, and I somewhat made a correlation between the meme changes and the light level.
That, with a blue dot-like effect on the screen. (I ended up hiding that part of my screen with a paper.)
the german FE-Schrift font used for number plates all over the world is specifically designed so that you cant modify it to easily turn one letter into another (like a B to a P etc using black and white tape), its pretty interesting
I looked it up, & that’s an incredibly nice font. Aesthetically pleasing, easy to read, my complements to Karlgeorg Hoefer
I'm far more fascinated by your Letter-H esque lightstring placement and your decision to use the Lapis pillow as a guitar stand.
When I first saw the title, I didn't notice it was weird and read it as "so this is how it all goes"
Speaking of heterography, compact fonts, and Minecraft, EthosLab came up with an elegant way to invent your own alphabet in Minecraft using the 4 stair rotations and 2 slab positions, which when arranged in a 1*2 block configuration, gives you just enough variations to fit all 36 figures and letters of the alphabet in a compact manner. He demonstrates it in his HermitCraft S9#10 video at 7:19.
Also, a 30 binary pixels font wouldn't amount to 30! possible combinations, but rather 2^30, or about 1 billion. 30! would describe the different possible ways to sort 30 different items in a sequence, whereas the Minecraft font corresponds to the different possible ways to keep 30 light bulbs powered or unpowered.
Reminds me a bit of what happens when those neural network image generators try to make small, non-specific text. just vague graphemes
I tried to read the test in the thumbnail, and my brain _obviously_ processed it as it would normal words, and my inner voice read it like some wizard chanting a spell.
Perhaps there is a way to calculate the "robustness" of a font by measuring how much mutations in the font is needed to confuse most people
Great video! Just wanted to point out a couple of details for accuracy. At 10:50, you mention the pixel density being 5x6, suggesting an area of 30. But at 12:15, it's evident that the minimum space is actually 5x7, which brings the area to 35. Also, at 11:12, while discussing the permutations of images, you correctly identify it as a binary calculation. However, instead of using factorial, for binary strings, we use 2 raised to the power of N (where N is the string length, here 35). This means the number of permutations for the pixels in the image is actually 34,359,738,368 - a significant difference from 30!. Keep up the good work, just thought these clarifications might be helpful.
In my previous comment, I calculated the permutations based on a 5x7 pixel grid as observed in the video, which resulted in 34,359,738,368 permutations. However, as another user (@YatYusYae) mentioned, the font sizes in this version of Minecraft can actually be up to 8x8. This adjustment means the true number of permutations for the font being used would be 2^64 = 18,446,744,073,709,551,616. This figure includes numerous permutations with empty spaces in the boundary layers.
I’ve never been more surprised and slightly confused for how I was able to read the thumbnail sentence so easily, having never seen this channel before. Excited to fall down a binge session 🙂.
HOW?!
I actually use the hyperlegible font because of this video, it's one of my favorite fonts! Thanks ŋə!
I think I found your heterography, for numerals at least, by accident. When I worked booking-in samples from a clinical trial with handwritten dates on their labels, which I had to compare to a typewritten manifest, I swear I saw every single numeral 0 to 9 rendered in bad handwriting in such a way as it could be mistaken for any other digit. I honestly wish I'd made copies of all the worst ones.
If "hetero" is a good antonym of "ortho", then the opposite of an "orthonormal basis" (a concept seen in vector spaces with an inner product, a notable example of which is the usual Cartesian basis for Euclidean vector spaces) is a "heteronormal basis".
We can analyse this further by noting that a _non-orthonormal basis_ isn't necessarily _the opposite_ of an orthonormal basis. Instead, we must directly counter each part of the definition.
- _Ortho-_ refers to the basis vectors being _orthogonal_ - meaning their inner product is 0 (for Euclidean vector spaces, they're at 90° to each other). The opposite of _orthogonal_ is _parallel_ as parallel vectors have no components that are orthogonal with each other. It is worth noting that a basis of vectors which are all parallel can only ever describe a 1D line, since vector-vector addition doesn't affect the angle of the vectors, so two parallel vectors will always sum to another parallel vector.
- _Normal_ refers to the basis vectors being of magnitude (for Euclidean vectors, length) 1. The opposite of this could either be magnitude 0 or perhaps magnitude ∞. The latter is not considered a number in the usual Real numbers, but can exist in _projective geometry_ where infinite points act as the 'points of perspective' one finds in perspective art.
Hence: _Heteronormality_ implies an inability to work with more than a single dimension at a time, and a magnitude of 0. The only way to achieve a non-zero magnitude is through projection. In both cases, even the single dimension that did exist is reduced to only one trivial point.
QED: Heteronormality is lame, LGBTQIA+ rights are cool. The mathematics says it all.
It's actually kinda crazy that "hetero", depending on the context, can be an antonym of "ortho" and a synonym of "straight".
To get every English letter in distinct form including upper/lower distinction, you only need a 4 or 5 wide and a 4 high pixel grid - 5x5 if you want things to be very clear for letters like M, W and E.
Atkinson Hyperlegible is my favorite font! It changed the way I draw lowercase js and is.
This is pretty sick, i like it. Also, I'm one of the reasons you jumped to 10k suddenly. I don't follow a lot of linguistics channel but your original CCC got algo'd a little bit so i saw it. I mean, i saw like a third of it, I'm still working my way through.
I created a language/orthography called ■● [/sə/] which consisted solely of the two phonemes.
The consonant was written as ■ and the vowel as ●, and anything else (i.e. punctuation) was▲
"Arial is the ideal font"
Every English Teacher: TIMES NEW ROMAAAAAAAN! TWELVE POINT!
Calibri be like: ☕
I think a more general approach is understanding that your goal is the reduction of redundant information.
Consider the 7-segment display. Did you know you can make segmented displays for numbers use more than 7 segments? They can be stylized to be more interesting and legible than 7-segment.
And did you know you can use fewer than 7 segments? There are 10 digits, which is greater than 2^3, but not greater than 2^4, so you need at least 4 segments to technically distinguish the 10 digits.
There's a video by 'Posy' called "Segmented Displays" which runs through various segmented displays. Apparently you can achieve reasonable legibility on numbers with as few as 6 segments, while legibility suffers greatly with fewer still.
92 characters is more than 2^6, but fewer than 2^7, thus you should only need 7 distinct graphemal features to compose all characters. Segmented displays require gaps between distinct segments, whereas in writing, the overlap may grant more legibility. You might note further that our goal is not preventing characters from being read, but from them being distinguished.
Furthermore, if we consider the 13375p34k dialect, and only concern ourselves with one case (upper or lower) and the digits, we only need 26+10-10=26 characters, which is not greater than 32. (In 13375p34k, both I and L are replaced by 1, O is replaced by 0, B may be 8, G may be 6, and R may be 2, along with other rules noted in the proper name of the graphical dialect. Indeed, only 9 has no letter, while 1 has two, effectively removing the contribution of the 10 digits from graphemal information complexity.)
5 segments is very difficult to make numbers for. Let's try making the numbers *and* alphabet in only 5 distinct figures!
Thank you for putting the definitions for the jargon, ive been super interested in conlangs and linguistics for a while but sometimes the stuff just goes over my head
I thought I was a language nerd, but this is proof that I have a long way to go before I achieve master form. Also, Hyperlegible made me think of a game I just played recently: Syrup and the Ultimate Sweet. It's a short, cute game, obviously made on a low budget, and so I was absolutely charmed when they included a font for dyslexics (like me). I definitely didn't expect it, but it was very welcome!
I love how you made the Hypolegible to the Hyperlegible
so this is how it all ends 😉
This comment with the wink was the most ominous thing I’ve ever read in my life my heart literally seized up
tysm!
I'm surprised that there was no mention of cursive Cyrillic. Its basically a naturally occurring heterography. Seriously, how does anybody read that squiggly mess?
I saw some russian doctors writing in English and Latin using cursive, and it is not really easier to read...
American doctors writing in English also produce variable legibility 😂
As someone who reads Cyrillic, I can’t make heads or tails of it. It’s just cubluuuuub for слышишь, for example.
Power of a habit 😅
My name is Yoshikage Kira. I'm 33 years old. My house is in the northeast section of Morioh, where all the villas are, and I am not married. I work as an employee for the Kame Yu department stores, and I get home every day by 8 PM at the latest. I don't smoke, but I occasionally drink. I'm in bed by 11 PM, and make sure I get eight hours of sleep, no matter what. After having a glass of warm milk and doing about twenty minutes of stretches before going to bed, I usually have no problems sleeping until morning. Just like a baby, I wake up without any fatigue or stress in the morning. I was told there were no issues at my last check-up. I'm trying to explain that I'm a person who wishes to live a very quiet life. I take care not to trouble myself with any enemies, like winning and losing, that would cause me to lose sleep at night. That is how I deal with society, and I know that is what brings me happiness. Although, if I were to fight I wouldn't lose to anyone.
That is a better lifestyle than mine, Yoshikage 👍
Faffing about with orthography I devised a sci-fi font made only of straight borders, corner-to-corner diagonals, and two dots, one per center of the top and bottom halves.
The whole uppercase alphabet represented by 8 morphemes; the idea being they can be scanned instantly by space ships at a great distance like a barcode, while still being readable enough for a human viewing it directly. There would be some wierd compromises though, while A would look like a tilted Atlantis A without the crossbar, other letters like P would be an upside-down L, and R would be that but with a semi-centered dot hovering by its base as a reference to its missing half-diagonal that the font doesn't support.
0:30 idk if this was you or just the camera acting up, but that brief lighting change was golden
it was an accident but I liked it so I kept it, lol
@@AgmaSchwa I'm not going to lie, the lighting change happens throughout the entire video, and I'm watching in a relatively low-lit room... I thought you were intentionally directing me to look at the computer
:)
Reminds me of what would happen during certain (now fortunately resolved) severe anxiety spikes several years ago during which letters and would start to warp/contort in demented ways due to my pattern recognition going haywire from subconsciously trying to visually locate non-existant threats. Reading any text I wasn't already familiar with was next to impossible and even sentences and phrases I encountered regularly remained only barely decipherable. All that to say that from experience actively having to read this kind of thing in real life sucks horribly (especially when your job requires reading important disclosures verbatim on a recorded line) and is monumentally cursed. 0/10 then but 10/10 here, can't wait to see what other unholy abominations people are inspired to make from this haha.
Oof anxiety is no joke, it can really bring the brain to some extremes. I know under enough anxiety I can't physically make myself talk, and even lost motor controls and had to be practically picked up off the ground, I wish more people would get how much severe anxiety can impair a person. I hope you never have to be in that state again
7:41 didn't expect to hear The Licc in a linguistics video
The minecraft font's dimensions are 7x5, not 6x5. If you look at the "af" on the screen, you can count the 5 rows that make the a which reach up to the dash on the f, after which another 2 rows are needed to make the upper arch on the f.
I use Atkinson Hyperlegible for everything! It's very good, makes it easier to skim my own writing without getting confused.
The first message was pretty clear up to the last word. "Hey everybody, welcome to n???" the last two letters both LOOK like N but can't be, because N was already used.
It took a bit of puzzling to figure out that your username was InsaneInbox, but once I had that I could figure out the rest more easily.
The second message starts, obviously, with "My name is Yos" but is more or less illegible from there.
The third message, after staring at it for a bit, I realized, was "The square root of 49 is 7" which gave me a lot more letters to work with.
Now that I have those letters, I can go back to message two and see if I can puzzle that out. "My name is Yoshikage Kira" I think?
The first message, now that the second one gave me a G and the third one gave me the U and H, says "Hey everybody, welcome to nguh"
That was fun to figure out! I'm liking this whole alternate alphabets thing.
20:09 Now that I've been staring at this alphabet for a while, it's immediately obvious that the carrot's name is "Karotte". You switched to German.
Looking at the alphabet at 20:42, the last word of the first line appears to be “nguh”, yes, and the second line ends in “Yoshikage Kira”. Also, you didn’t do the 4th or 5th messages, so I’ll decode those too. The 4th is “do you have any idea what i am capable of?” And the 5th is “thanks for watching!” The only difference between “!” and “?” is that “?” is shifted one pixel to the right compared to “!”. Someone else can decode the text on the screen at 21:53, I won’t do it yet, I’ll do it tomorrow if it isn’t done by then.
... a bang, but with a whimper
light fades on this feable attempt at... (spelling not mine)
...s the next one will be kinder to
... way, we lived
...at is beautiful
... be not, and it will all play out the game ...
x dee return to monke
...ush me on the cursed conlang circus
...ing on it, I swear. Just give me and ...
I think this is all right
@@sheep4483 what is this seemingly random block of text you have presented?
@@CCM278 the text on the screen at 21:53
13:15 you've made cyrillic minecraft font б
the objectively best -eme is the one that best performs your purpose in communicating to your audience. If you want minimum ambiguity that means one thing (sans-serif on a screen and maybe serif fonts in print but who knows), and if you want to stack up entendres maybe you want to commit horrible crimes like this video.
Can't believe Danish just got a written form
Ooohhh ideasss!
Idea a. Make the pixel font as small as possible, so that each letter is naturally only one pixel off from a bunch of others. You would need at least 7 pixels (2^6 < 90 < 2^7), but since 7 is a prime number and wouldn't form a rectangular grid, either a 2x4 grid or a 3x3 grid should work.
Idea α. Reshape the 7-segment display so all characters can be shown. (Relatedly, I remember watching a video trying to reshape the segments so all 10 digits could be displayed with only 4 bits of information.)
Idea א. Break up each letter of the alphabet into "strokes", like how Chinese characters can be broken into radicals. I tried this for the lowercase letters in my cursive style, and got it down to a list of 13 strokes.
These have been swimming around in my brain since I first watched the video a week ago, and by now I've filled one page of physical paper, one page on the Notes app, and two pages of spreadsheets exploring ideas ay and aleph.
Anyways, yeah. Cool video.
Edited because I forgot starting a line with a Hebrew character makes the punctuation in the whole line go right-to-left. Oop.
Oh I love making alphabets like this
I’ve made one where I take away one line from each letter
I’ve made one where they are variations on a weird spiral shape (a couple letters didn’t work like that)
3:53 I don't think anyone argued about the number in the comments yet, but there's actually 94 under this definition - there are 32 punctuation symbols on the US keyboard.
10:45 the Minecraft font is 7 pixels tall, not counting descenders, and others have pointed out this makes the number of combinations 2^35.
Since orthos means straight in Greek I would guess the antonym of orthography should be campsography ("καμψός" meaning crooked)
10:10 If your version of the “perfect grapheme” doesn’t give capital “i” serifs, you’re simply wrong.
what's funny to me is ive never actually been able to register that uppercase cursive G as actually being a g, i always write it differently because it looks unrecognizable as g to me
It is more like a heavily stylized lowercase g, which is what makes it confusing, because that's the last thing you expect to see when you start off with basic script, where uppercase G looks like this
lowercase r and lowercase u were just ludicrous. even in alphabetical order I think you were just trolling with those two.
what that strange article puts forward is actually pretty intuitive and freeing. essentially, there are no rules to break; everything, like all things it seems, more and more, is an infinite matter of unknowable prerequisites only seeming to enact, or perhaps truly enacting themselves in the present, in the becoming, through life and conscious memory
insaneinbox: Hey everybody, welcome to nguh
insaneinbox: My name is Yoshikage Kira
insaneinbox: The square root of 49 is 7
insaneinbox: do you have any idea wgat i am capable of?
Seems like you picked, a very weird language to Try and figure out what it is.. It's some form of Swabian.
Having the lick in the slide for Part 3 reminds me of Adam Neely ^^
I really liked this, subjects like this are fascinating to me! I'm subbing.
One piece of constructive criticism I have is that it was pretty difficult to focus on you speaking when the video brightness kept shifting every time you moved. You seem to have auto-exposure locked On on your camera, and it might be better to set the exposure once each time you set it up, then keep it on that setting while you film.
0:25 Some may even say ' ' ' ' 'wrong' ' ' ' '
The text in the thumbnail reads “so this is how it all ends”
Oh, btw ChatGPT has gotten a lot better a conlanging
Oh boy, I'll have to give it another try then, haha
@@AgmaSchwastream. pls. not a video :)
Can't wait for the Homography Minecraft font that replaces every letter with "L"
Am I the only one that could pick up many words?
But yeah, I loved the idea. I'd do as such: I would challenge myself to only change x pixels per character, increasing x in every iteration. What is the minimal x that screws everything real bad?
Id love the whole thought process for merging alphabet
If only I knew every channel that gets 10K subscribers because it provides massive content (squeegee cleaning every corner of a topic so much tht if it were a black hole, every fold in the blanket of space time would be spotless), but the World and Himanity is not ready for - which, consequently, causes the YT algorithm to disfavor such releases… yeah there needs to be a specific home to YT for social intellectuals
Things that I didn't expect to not be great to watch before bed:
It's like the uncanny valley for text, its like I lost the ability to read my native language, I can imagine having a weirdly specific nightmare of this.
17:56
“Hey everybody, welcome to ___”
“My name is Y____ ___”
“The square root of 44 is 7” (I know it’s ‘49,’ but I originally read it as 44)
“Do you have any load what I am capable of?” (again, ‘idea,’ but my brain originally read it as load. in hindsight this sentence was pretty easy once I glanced at ‘capable’)
kinda crazy how some of the common words the brain instantly fills, then immediately fizzles to a stop on some of the others because it just couldn’t find one that made sense with the ‘letters’ being presented. (‘course, I sat through the sped up footage of the editing process, so it’s possible some of that my brain held on to enough to remember for later.)
as a jojo fan i was still immediately able to tell "My name is Yoshikage Kira.", its just baked into our souls
Yeah
just because something can be done does not mean it should be
dear gods
you have earned a subscriber
My favorite allographic variation has to be lower case Greek sigma, which is commonly written as , but turns into when word-final. Example: γνῶσις, gnosis; 'knowledge'.
What I find so interesting about it is that unlike the orthographic allographs you mentioned, which display allography based on the lexeme itself, lower case sigma comes in a complementary distribution based upon its position in a lexeme.
You could write down a rule for sigma (using || for graphemes and for graphs, something like |σ| -> / _#), which is something I'd find quite impossible to do for the orthographic allographs. Capitalization at the beginning of a word seems to also go into this direction, but is also different, as it is both positional and based on the lexical category of the word (proper nouns in English or all nouns in German) in which it occurs. Not to mention sentence initial capitalization.
Furthermore, in contrast to the allographs in free variation (like the different fonts), in this case it seems to me that there is an underlying grapheme |σ|, which is realized as the graph when at the end of the word, as the graph at the beginning of proper nouns and sentence initial words, and as the graph otherwise.
this randomly popped in my reccomended then i noticed you're the dude that made that hunger games sim that everyone puts characters or their friends in so thank you for this information and that wonderful website
i think i have dyslexia
so this is now it all ends
You're so funny! What a cool video-- its been recommended to me for a while, and I'm glad i finally watched it!
I just want to show my appreciation towards the different images you have showing on the desktop behind you, it really adds to the perceived quality of the video, which feels like that of a million subscriber channel. Especially the one at 7:45, I found that particularly funny, LMAO.
I used to use the Minecraft font so much that I would memorize it so I could draw it in MS Paint without having a font installed
also aren’t the letters actually 7×5? the ‹o› is 5px tall, and the ‹b› extends 2px higher?
Still haven’t beat Propeller Power
In a game Axiom Verge there are notes written in an "alien" language that is actually just english with very deformed letters. When I played I spent some time learning to read this deformed alphabet by recognizing familiar words. Later in the game you find an item that translate those notes to the standard english...
Always wanted to do this, can’t wait
Honestly its surprisingly fun attempting to decrypt this. Cool stuff :)
"so this is how it all ends"
Do you know that -7 is also a solution to sqrt(49)? Yeah, now you have an idea what I am capable of!
When a entire country learn cursive, but few write it well, you end up with most population to "decode" what was the meaning of the sentence. Plus, including grammatical, conjugation and other errors, you really have to be deductive. A lot of teacher in my country would not struggle for too long to play with your font. Heck, give me 20 minutes to be used to it, and I pretty sure I could play as if I was just a slow-reading child.
(im late but)
-7 is not equal to sqrt(49) as the "sqrt()" function/operation only takes the principal root (the positive one)
however, x=-7 is a solution to x^2=49 along with x=7 but not "-7 is a solution to sqrt(49)"
i feel like i couldve explained this better but im supposed to be asleep rn lmao
edit: if i missed that this was supposed to be a joke i apologize in advanced as i am very sleepy zzz
"My name is Yoshikage Kira"
I will NEVER miss that line. My friend has the entire copypasta with ONLY the vowels, so I've trained.
I could miss it depending on whether the vowels are English or Japanese, but I'm not going to lose any sleep over it.
SCP-CN-1001 moment
TFW Agma does a premier at 2pm
(Also the thumbnail concerns me for what eldritch thoughts Agma has birthed)
i've had to change the way i write two letters over the years, and done a third unintentionally. the first one i did around 5th grade, and it was lowercase y. i wrote it like a sans serif font with two strokes and it was hard to tell apart from lowercase x, so i decided to adopt the hook of a lowercase g like i saw some other people doing. then in 10th grade i didn't like that my lowercase t was hard to distinguish from a plus sign and that the first stroke frequently trailed into the second, so i started writing it like a sans serif font with a hook at the bottom. when i was quite young, i wrote lowercase a like a sans serif font, but i started writing it without the top hook when i got into school. and recently i've also started occasionally dropping the cross bars from capital i for writing speed and when i don't have much room
Is it bad that I knew the splash text at 16:02 was Absolutely no memes?
This one it would be great for someone to make a program that changes each letters, minimizing both the count of bit flips and the confidence level when shoved into OCR. That way the surprise is on you.
I would like to do that, but I don't think I got any free OCRs that can reliably detect pixel glyphs...
The auto adjustment of white balance or exposure or w/e on your camera is going nuts in this video
17:58
ゴゴゴゴゴゴ...
this would be fun to do with the pico8 font, it's only 3 by 5 so there's even less room for error
A doctor once called me ‘hyperlexic’ and I’ve never felt that more than being able to read these words instantly even through the pirate filter lol
As someone who loves seeing pixel fonts that attempt to infuse a recognizable style while keeping legibility all at really low resolutions this was painful, but I like it, it's definitely interesting to see where is the breaking point. It reminded me of when I struggled really badly at creating a functional font while only giving 4 by 4 pixels to every character, specially cos I wanted to cover some diacritics.
i think i caught something wrong with the math you did. In the minecraft font every pixel can either be on or off, since there are 30 total pixels that means there is 2³⁰ different combinations of pixels, a number way smaller than 30!
Thank god someone else noticed, I thought I was going insane
Your comments on the connection between the shape of letters being directly associated with the sounds they make (around 9 mins) cause me to think about how current AI works to understand images as an input, and make me wonder if this sort of thinking could how AI training models might be able to interact with each other to make it easier to get what you want out of the system.