Which letter of the English alphabet takes up most

Question

I am trying to do some dynamic programming based on the number of characters in a sentence. Which letter of the English alphabet takes up the most pixels on the screen?

Answer 1

It depends on the font. Crossed zero for example takes up considerably more than a regular one.

But if one could put up a guess, I'd go with X or B.

Answer 2

How about a programmatic solution?

var capsIndex = 65;
var smallIndex = 97
var div = document.createElement('div');
div.style.float = 'left';
document.body.appendChild(div);
var highestWidth = 0;
var elem;

for(var i = capsIndex; i < capsIndex + 26; i++) {
    div.innerText = String.fromCharCode(i);
    var computedWidth = window.getComputedStyle(div, null).getPropertyValue("width");
    if(highestWidth < parseFloat(computedWidth)) {
        highestWidth = parseFloat(computedWidth);
        elem = String.fromCharCode(i);
    }
}
for(var i = smallIndex; i < smallIndex + 26; i++) {
    div.innerText = String.fromCharCode(i);
    var computedWidth = window.getComputedStyle(div, null).getPropertyValue("width");
    if(highestWidth < parseFloat(computedWidth)) {
        highestWidth = parseFloat(computedWidth);
        elem = String.fromCharCode(i);
    }
}
div.innerHTML = '<b>' + elem + '</b>' + ' won';

Answer 3

Arial 30px in Chrome - W wins.

Answer 4

It also depends on the font. I did this 1 or 2 years ago with Processing and Helvetica and it is ILJTYFVCPAXUZKHSEDORGNBQMW in order of increasing pixels. The idea is to draw the text on a canvas with the font you are looking at, count the pixels, then sort with a HashMap or Dictionary.

Of course, this might not be directly relavant to your use as this calculates pixel area rather than just width. Might be a little overkill too.

void setup() { 
 size(30,30);
 HashMap hm = new HashMap();
 fill(255);
 PFont font = loadFont("Helvetica-20.vlw");
 textFont(font,20);
 textAlign(CENTER);

 for (int i=65; i<91; i++) {
    background(0);
    text(char(i),width/2,height-(textDescent()+textAscent())/2); 
    loadPixels();
    int white=0;
    for (int k=0; k<pixels.length; k++) {
       white+=red(pixels[k]);
    }
    hm.put(char(i),white);
  }

  HashMap sorted = getSortedMap(hm);

  String asciiString = new String();

  for (Iterator<Map.Entry> i = sorted.entrySet().iterator(); i.hasNext();) { 
    Map.Entry me = (Map.Entry)i.next();
    asciiString += me.getKey();
  }

  println(asciiString); //the string in ascending pixel order

}

public HashMap getSortedMap(HashMap hmap) {
  HashMap map = new LinkedHashMap();
  List mapKeys = new ArrayList(hmap.keySet());
  List mapValues = new ArrayList(hmap.values());

  TreeSet sortedSet = new TreeSet(mapValues);
  Object[] sortedArray = sortedSet.toArray();
  int size = sortedArray.length;

  // a) Ascending sort

  for (int i=0; i<size; i++) {
    map.put(mapKeys.get(mapValues.indexOf(sortedArray[i])), sortedArray[i]);
  }
  return map;
}

Answer 5

A solution to calculate the widths of fonts a bit like the solution posted by xxx was posted by Alex Michael on his blog (which funnily enough linked me here).

Summary:

• For Helvetica, the top three letters are: M (2493 pixels), W (2414) and B (1909).
• For a set of fonts that shipped with his Mac, the results are more or less the same: M (2217.51 ± 945.19), W (2139.06 ± 945.29) and B (1841.38 ± 685.26).

Original Post: http://alexmic.net/letter-pixel-count/

Code:

# -*- coding: utf-8 -*-
from __future__ import division
import os
from collections import defaultdict
from math import sqrt
from PIL import Image, ImageDraw, ImageFont


# Make a lowercase + uppercase alphabet.
alphabet = 'abcdefghijklmnopqrstuvwxyz'
alphabet += ''.join(map(str.upper, alphabet))


def draw_letter(letter, font, save=True):
    img = Image.new('RGB', (100, 100), 'white')

    draw = ImageDraw.Draw(img)
    draw.text((0,0), letter, font=font, fill='#000000')

    if save:
        img.save("imgs/{}.png".format(letter), 'PNG')

    return img


def count_black_pixels(img):
    pixels = list(img.getdata())
    return len(filter(lambda rgb: sum(rgb) == 0, pixels))


def available_fonts():
    fontdir = '/Users/alex/Desktop/English'
    for root, dirs, filenames in os.walk(fontdir):
        for name in filenames:
            path = os.path.join(root, name)
            try:
                yield ImageFont.truetype(path, 100)
            except IOError:
                pass


def letter_statistics(counts):
    for letter, counts in sorted(counts.iteritems()):
        n = len(counts)
        mean = sum(counts) / n
        sd = sqrt(sum((x - mean) ** 2 for x in counts) / n)
        yield letter, mean, sd


def main():
    counts = defaultdict(list)

    for letter in alphabet:
        for font in available_fonts():
            img = draw_letter(letter, font, save=False)
            count = count_black_pixels(img)
            counts[letter].append(count)

        for letter, mean, sd in letter_statistics(counts):
            print u"{0}: {1:.2f} ± {2:.2f}".format(letter, mean, sd)


    if __name__ == '__main__':
        main()

Answer 6

I know the accepted answer here is W, W is for WIN.

However, in this case, W is also for Width. The case study used employed a simple width test to examine pixels, but it was only the width, not the total pixel count. As an easy counter example, the accepted answer assumes that O and Q take up the same amount of pixels, yet they only take up the same amount of space.

Thus, W takes up the most space. But, is it all the pixels it's cracked up to be?

Let's get some empirical data. I created imgur images from the following B, M and W. I then analyzed their pixel count (see below), here are the results:

B : 114 pixels

M : 150 pixels

W : 157 pixels

Here is how I fed them into canvas and analyzed the raw pixel data from the images.

var imgs = {
 B : "//i.imgur.com/YOuEPOn.png",
 M : "//i.imgur.com/Aev3ZKQ.png",
 W : "//i.imgur.com/xSUwE7w.png"
};
window.onload = function(){
  for(var key in imgs){(function(img,key){
    var Out = document.querySelector("#"+key+"Out");
    img.crossOrigin = "Anonymous";
    img.src=imgs[key];
    img.onload = function() {
      var canvas = document.querySelector('#'+key);
      (canvas.width = img.width,canvas.height = img.height);
      var context = canvas.getContext('2d');
      context.drawImage(img, 0, 0);
      var data = context.getImageData(0, 0, img.width, img.height).data;
      Out.innerHTML = "Total Pixels: " + data.length/4 + "<br>";
      var pixelObject = {};
      for(var i = 0; i < data.length; i += 4){
        var rgba = "rgba("+data[i]+","+data[i+1]+","+data[i+2]+","+data[i+3]+")";
       pixelObject[rgba] = pixelObject[rgba] ? pixelObject[rgba]+1 : 1;
      }
      Out.innerHTML += "Total Whitespace: " + pixelObject["rgba(255,255,255,255)"] + "<br>";
      Out.innerHTML += "Total Pixels In "+ key +": " + ((data.length/4)-pixelObject["rgba(255,255,255,255)"]) + "<br>";
    };
  })(new Image(),key)}
};

<table>
<tr>
  <td>
    <canvas id="B" width="100%" height="100%"></canvas>
  </td>
  <td id="BOut">
  </td>
</tr>
<tr>
  <td>
    <canvas id="M" width="100%" height="100%"></canvas>
  </td>
  <td id="MOut">
  </td>
</tr>
<tr>
  <td>
    <canvas id="W" width="100%" height="100%"></canvas>
  </td>
  <td id="WOut">
  </td>
</tr>
</table>