2 ms·
I think it's partly because we recognize letters, and whole words, by glyph shape more than specific identity. Obviously a 2x2 grid can only depict 16 different
by HarHarVeryFunny 1y ago
I think it's partly because we recognize letters, and whole words, by glyph shape more than specific identity. Obviously a 2x2 grid can only depict 16 different patterns, but we're trying to recognize whole words, not arbitrary letter sequences, and the sequence of shapes (hence letter possibilities) is evidentially enough, a bit like reading crappy handwriting.
It's interesting how we can do this with this 2x2 font immediately without any training, but I suppose reading in general has provided enough training, and ability to read this 2x2 font just provides some insight as to how word perception works.
- tremon 1y agoMost letters are 2x3 px, the letter m is even 2x5. And I wouldn't say that I could comfortably read this, it was closer to deciphering than reading.
- HarHarVeryFunny 1y agoTrue, but it's interesting that we can quickly decipher/read it at all. It seems to be a typical case of human perception where top-down prediction (maybe of visual word forms?) meets bottom-up sensory input, and we've gained enough experience of this (reading different fonts, handwriting, various lighting, etc) that this particular type of impoverished input doesn't pose much challenge.
- NBJack 1y agoI believe it's in part because of our experience reading things at angles. In this case, it looks to me like letters tilted backwards on a table, and I'm peering at them just above their horizon. Legible, but not comfortable.