Decode Faux Braille

Question

⢣⠃⢎⠆⣇⡇⡯⡂⠈⡏⢰⢵⢐⡭⢸⠪⡀⢸⢐⡭⠀⢹⠁⢎⠆⢸⣱⢸⡃⢎⠰⡱⢸⣱⢸⡃⠈⡏⢸⡃⡱⡁⢹⠁⢸⡀⡇⡗⢅⢸⡃⠈⡏⢸⢼⢸⢐⡭⠀

⣇⢸⡃⢹⠁⢹⠁⣟⢸⢕⢐⡭⠀⡮⡆⡯⡂⣟⠀⡯⠰⡱⢸⣸⢸⢕⠀⣏⡆⢎⠆⢹⠁⣪⠅⢸⢼⢸⠰⣩⢸⢼⠀⡮⡆⡗⢼⢸⣱⠀⢎⠆⡯⠀⢇⠇⡮⡆⡯⡂⡇⡮⡆⣟⡆⣇⢸⡃⠸⡰⡸⢸⢸⣱⠈⡏⢸⢼⠀

⢎⠆⡗⢼⢸⡃⢸⡃⡗⠔⡇⡯⠂⢹⠁⢣⠃⠸⡸⢸⡃⡯⡂⢹⠁⡇⢎⢰⢵⢸⡀⢸⡀⡇⡗⢼⢸⡃⢐⡭⢸⡃⡯⠂⡮⡆⡯⡂⡮⡆⢹⠁⣟⢐⡭⠀⢎⢸⢼⢰⢵⢸⢕⢰⢵⠰⡁⢹⠁⣟⢸⢕⢐⡭⠀

⡮⡆⢐⡭⢸⠕⢰⢵⠰⡁⣟⠀⡇⣪⠅⢈⣝⢸⡃⡯⡂⢎⠆⠸⡰⡸⢸⢸⣱⠈⡏⢸⢼⠀

⣪⠅⢎⠆⢸⠈⡏⠀⣇⠰⡱⠰⡱⢸⠪⡀⣪⠅⢸⡀⡇⡗⢅⢸⡃⠸⡰⡸⠰⡱⢸⢕⢸⣱⢐⡭⠀⡮⡆⡯⡂⣟⠀⣪⠅⣟⢸⠕⢰⢵⢸⢕⢰⢵⠈⡏⢸⡃⣏⡆⢸⣳⠘⡜⠀⢹⠁⢇⢆⠇⢎⠆⢸⡀⡇⡗⢼⢸⡃⣪⠅

⡇⡗⢼⢸⠕⢸⣸⠈⡏⠀⡇⣪⠅⢰⢵⠀⣪⠅⢹⠁⡯⡂⡇⡗⢼⠰⣩⠀⢎⠰⡱⢸⠢⡇⢹⠁⡮⡆⡇⡗⢼⢸⢸⠢⡇⢎⡅⢸⠅⡮⡆⣇⡇⡱⡁⢸⣳⢸⢕⢰⢵⢸⢸⡀⣇⢸⡃⠰⡱⢸⠅

⢎⠆⡗⢼⢸⡀⢣⠃⢸⡃⡗⢼⠰⣩⢸⡀⡇⣪⠅⡧⡇⢸⣸⢸⠕⢸⠕⢸⡃⡯⡂⢎⢰⢵⢐⡭⢸⡃⢸⡀⣟⠈⡏⠈⡏⢸⡃⡯⡂⣪⠅⢰⢵⢸⠢⡇⣏⡆⢐⡭⢸⠕⢰⢵⠰⡁⣟⢐⡭⠀

⡮⡆⣟⡆⢎⢸⣱⢸⡃⡯⠰⣩⢸⢼⢸⢀⠇⡗⢅⢸⡀⡗⠔⡇⡗⢼⠰⡱⢸⠕⠰⣩⡆⡯⡂⣪⠅⢹⠁⣇⡇⢇⠇⢇⢆⠇⡱⡁⢣⠃⣩⡃

⢎⠆⣇⡇⢹⠁⡯⠂⣇⡇⢹⠁⢸⠢⢺⢰⢵⠘⡜⠀⣟⡆⣟⠀⣇⡇⡯⠂⡯⠂⣟⢸⢕⠀⢎⠆⡯⡂⢸⡀⢎⠆⢇⢆⠇⣟⢸⢕⠰⡁⡮⡆⣪⠅⣟⠀

⣪⠅⡧⡇⢎⠆⡯⡂⢹⠁⣟⢐⡭⠈⡏⠀⢇⢆⠇⡇⡗⢼⢐⡭⠀

⡗⢼⠰⡱⠀⣇⠰⡱⠰⡱⢸⠕⢸⢼⠰⡱⢸⡀⣟⢐⡭⠀

ASCII version of the above

⡯⡂⣟⢸⡀⡮⡆⢹⠁⣟⢸⣱⠀

About Braille characters

A Braille character packs a 4 by 2 rectangle of dots, which can be viewed as a Boolean matrix.

The concatenation of all such matrices is a 4 by 2*n boolean matrix, where n is the length of the input string.

You should look for vertical lines without any dots in them, and use those as separators to split the big matrix into smaller matrices for each character.

Then, look for patterns to convert them to letters of the English alphabet or spaces. Note that after removing the separators (vertical empty lines), a space is a 4 by 0 matrix.

Below is a description of the alphabet in ASCII:

A   | B   | C  | D   | E  | F  | G   | H   | I | J  | K    | L  | M     | N    | O   | P   | Q    | R   | S   | T   | U   | V   | W     | X   | Y   | Z
----+-----+----+-----+----+----+-----+-----+---+----+------+----+-------+------+-----+-----+------+-----+-----+-----+-----+-----+-------+-----+-----+----
.#. | ##. | .# | ##. | ## | ## | .## | #.# | # | .# | #.#. | #. | #...# | #..# | .#. | ##. | .##. | ##. | .## | ### | #.# | #.# | #...# | #.# | #.# | ###
#.# | ### | #. | #.# | ## | #. | #.. | #.# | # | .# | ##.. | #. | ##.## | ##.# | #.# | #.# | #..# | #.# | #.. | .#. | #.# | #.# | #.#.# | .#. | #.# | ..#
### | #.# | #. | #.# | #. | ## | #.# | ### | # | .# | #.#. | #. | #.#.# | #.## | #.# | ##. | #.## | ##. | .## | .#. | #.# | #.# | #.#.# | .#. | .#. | .#.
#.# | ### | .# | ### | ## | #. | .## | #.# | # | #. | #..# | ## | #...# | #..# | .#. | #.. | .### | #.# | ##. | .#. | ### | .#. | .#.#. | #.# | .#. | ###

Specification

• The input is a sequence of Unicode code points in the range U+2800..U+28FF represented as the usual string type in your language (e.g. char array, char pointer) in any popular encoding supported (UTF-8, UCS-2, etc).

• Trailing spaces in the output are fine.

---

EDIT: apologies to those whose browsers misrender the dots, it's supposed to look like this (image):

Answer 1

Python 3, 181 179 171 167 161 159 bytes

Input by UTF-16 little-endian bytes without BOM. First decompose into columns using bit shifts, split by empty column, then hash them into a lookup table.

-2 bytes thanks to ngn.
-5 bytes thanks to Mr. Xcoder.

lambda h,j=''.join:j(' ZAQV;JWP;MBOS;YRKCGXDF;ILHUENT'[int('0'+i,27)%544%135%32]for i in j(chr(64|i&7|i>>3&8)+chr(64|i>>3&7|i>>4&8)for i in h[::2]).split('@'))

Try it online!

Answer 2

JavaScript (ES6),  148 146 143  142 bytes

Saved 1 byte thanks to @ngn

s=>[...s].map(c=>g((k=c.charCodeAt()/8)&8|k*8&7)&g(k&7|k/2&8),o=x='',g=n=>x=n?x*27+n:!(o+=' DZQGYWXNHJ.CSTIO.AFB.LPVE..KUMR'[x%854%89%35]))&&o

Try it online!

Answer 3

Python 3, 305 302 301 286 251 198 182 bytes

def f(s,A=''):
 for c in s:l=bin(ord(c))[-8:];A+='7'+l[5:]+l[1]+'7'+l[2:5]+l[0]
 print(''.join('K.L.SXC PRU.NYEOGZVJIW..HFBTAQDM'[int('7'+c,22)%141%109%35]for c in A.split('70000')))

Try it online!