Angles Of View

If you are reading these words on a printed page, you are seeing them in 11 point Times New Roman Italic. If you are using your browser to read them at www.da-lite.com, it’s hard to say exactly what typeface you’re seeing. And if by chance you’re projecting the contents of your computer screen up onto (or through) a projection screen, then what will you see? And, more importantly, how well can you see it? This series has devoted many, many words to the visual display of information. Now it is time to consider the shape of those words which means looking at

Fonts - A Case in Point

Fonts, of course, have been around for more than 500 years - ever since Herr Gutenberg invented movable type in 1437. And, as anyone who owns a computer knows, there are lots and lots of them to choose from.



Figure 1

All fonts have faces and all faces belong to a family. Thus this font appears here in its standard face, but there is also Italic and Bold as additional faces. To see the elements which distinguish one font from another, consider Figure 1.

These are the principal parts of the characters and symbols which make up any font family. Additionally, ever since fonts became scalable, we can describe their printed size by specifying the number of points a letter is to take up. There are approximately 72 points to the inch and 12 of these points comprise a pica. Precise as those measurements are, however, they regrettably are of little or no use when the font is not to be printed but projected. Approximately one seventh of an inch high letters are fine for this page but, obviously, would be hopeless on a 100-inch diagonal projection screen.

Some interesting things may be observed about the legibility of fonts in general. One of them is that mixed case text is easier to read than TEXT WHICH IS PRINTED ALL IN CAPITAL LETTERS or even Text Which Has Just The First Letter Of Every Word Capitalized.

Another is that legibility depends on the tops of words.




Now see how much more difficult it is to decipher the lower half of the same sentence:



Figures 2&3

An enormously important attribute of fonts is whether any particular face is serif or sans serif. Serifs are small, usually horizontal cross strokes that are added to the ends of a letter’s main strokes. This typeface has serifs. This typeface does not have serifs. Because of their ability to coax the eye along the line of type, fonts with serifs are generally acknowledged to be more readable than fonts that are sans serif.

Becoming sensitive to these observations is helpful, of course, but it still doesn’t really answer the question, how can we think about fonts for projected displays and what guidelines can we follow for their manipulation? Fortunately there are some valuable answers and once again they emerge from research undertaken by Dr. Joy Ebben of JME International in Alta Loma, CA.

Dr. Ebben breaks the design of projected characters into five principal categories: Shape, Height-to-Width Ratio, Pixel Matrix, Stroke Width, and Character, Word and Line Spacing.

With respect to the shape of the possibly ambiguous characters, she offers the following case sensitive criteria:

A needs clearly delineated space above its horizontal stroke.
B needs approximately equal loops.
C & G are easily confused with each other and with O if the C break is not clearly discernable or if the horizontal stroke of the G is not long enough.
D & O can be confused if the O appears to square.
E requires clearly delineated spaces.
M & W need sufficiently long center sections.
P requires a large enough loop.
S & 5 are easily confused if the S is too square and/or the horizontal top of 5 is not long enough.
1 & l [sic] must be made to look different.
U & V will be confused unless the uprightness of vertical strokes of the U is maintained.
Y needs a long tail to differentiate it from V and it needs a distinctly v-shaped top to differentiate it from a T.
6 & 9 need apparent (but not too large) loops and fairly straight tails.

Other likely confusions are that X can be called K (and vice versa), H can be thought to be M or N, J or T can be called l, and K can look like R. Additionally, B can look like R, S, or 8; 0 (zero) can seem to be O or both can look like Q; and, to get to the end of the alphabet, a Z can often look like a 2.

Based on the typical width of a set of capital letters, the Height-to-Width ratio should be not less than 70% and not more than 90%.

The pixel matrix available for any character or symbol is absolutely critical to adequate legibility. Ironically, however, here is a case where the increased and increasing resolution from better and better visual displays does not help. When you upgrade either your computer or projector from the 600 x 800 device you had last year to the fancy new 1024 x 768 you’ve been promising yourself for this year, the first thing you’ll notice is how much harder it is to read.

Why is that so? Because, although your new machine will display 306,432 more pixels than your old, not one of those extra pixels is devoted to the improvement of your font. Thus, if your SVGA machine reserved, say, a 7 x 5 pixel array to write each of its lowercase characters, then your new machine also will allocate only the same 35 pixels to the same task, even though each of those pixels is only 3/5 the size of its predecessor. That’s (roughly) the difference between reading this phrase and trying from the same viewing distance to decipher this one.

Dr. Ebben’s rule of thumb for visual tasks that require the continuous reading of projected text is that the minium matrix be 9 pixels high by 7 wide. Clearly, however, even more pixels will always result in even better character definition.

Next we come to stroke width for which the historical ANSI (American National Standards Institute) standard has been that it be greater than 1/12 of character height. Dr. Ebben and others believe, however, a stroke width of 1/4 to 1/8 of character height is preferable for projected text. Furthermore, Dr. Ebben has discovered that there is marked discrepancy between what she calls forward video – light characters on a dark background – and its inverse.

The rule, then, is to use at least a double pixel wide stroke for large screen displays which are forward video. Interestingly, this is true because double stroke width in reversed video (dark on light) appears to be narrower than single stroke width in forward video.

The ANSI standard for mixed case horizontal character spacing has historically been given as being not less than 10% of character height. Because of even moderate off-axis viewing angles, however, Dr. Ebben believes that the minimum space should be increased to 25%. In terms of a 9 x 7 pixel matrix, this corresponds to two dot elements. However, this spacing should be increased to 50% whenever viewing angles become large.

With regard to the appropriate spacing between words, Dr. Ebben reports that the distance between one word and another should not ever be less than the width of one character. Spacing between lines – leading (rhymes with bedding) – has been established by ANSI to be a minimum of 2 stroke widths or 15% of character height, whichever shall be greater. Dr. Ebben believes that there is good reason to increase this minimum to 50% of character height for projected text.

Are all of these typography criteria really worth attending? Yes. Because we must never lose sight of the enormous difficulties inherent to the task of reading information that has been projected.


In addition to Dr. Ebben’s original research, other sources used in the preparation of this article are:

Shurtleff, Donald A., How to Make Displays Legible, Human Interface Design, 1980.
Cavanaugh, Sean, Digital Type Design Guide, Hayden Books, 1995
Carter, Rob, Working With Computer Type, Rotovision, 1997