=Paper=
{{Paper
|id=Vol-1452/paper21
|storemode=property
|title=Irregularity as a Quantitative Assessment of Font Drawing and Its Effect on the Reading Speed
|pdfUrl=https://ceur-ws.org/Vol-1452/paper21.pdf
|volume=Vol-1452
|dblpUrl=https://dblp.org/rec/conf/aist/TarasovS15
}}
==Irregularity as a Quantitative Assessment of Font Drawing and Its Effect on the Reading Speed==
https://ceur-ws.org/Vol-1452/paper21.pdf
Irregularity as a Quantitative Assessment of Font
Drawing and Its Effect on the Reading Speed
Dmitry Tarasov1, Alexander Sergeev1,2
1Ural Federal University, Ekaterinburg, RUSSIA
2Institute of Industrial Ecology, UB of RAS, Ekaterinburg, RUSSIA
datarasov@yandex.ru
Abstract. It is proposed to use irregularity, the scale invariant index based on
the ideas of fractal geometry to assess the spatial features of font drawings. The
index is sensitive to the shape of characters in the font, which affects text legi-
bility. Preliminary results have shown promising application of the proposed
index for classifying fonts by reading speeds.
Keywords. Font; Fractal; Scale invariance; Legibility; Typeface
1 Introduction
Research in the fields of legibility and readability are maintained for over a hundred
years. They are particularly important for the development of textual materials in-
tended for readers with emerging reading skills. A significant place in these studies
takes fonts. Many researchers have investigated the clarity, legibility, readability of
different fonts, the influence of serifs, the influence of the pattern and spatial charac-
teristics of the font on the understanding and memorising the content of the text and
some other factors. The obtained results are contradictory. So far there is no consen-
sus on what fonts features and how affect the reading process. This is largely due to
the lack of an objective index, which could describe the typeface, and allows compar-
ing different fonts.
Artemov [1] proposed to divide the concepts of visibility and readability of the font.
Readability is influenced by reader’s physiological characteristics. Visibility depends
on the quality of font drawing and vision features of the person. Differences in type-
face readability investigated in [2-4]. Some fonts are marked as the most readable.
The superiority of some small book fonts connected to their shapes and drawings is
demonstrated. Thick font reads faster. At the same time, respondents preferred the
other fonts. Similar results were obtained in [5]. Studies have shown the presence of
subjective preferences of readers, as well as an objective difference in readability of
fonts with different shapes. The review [6] analysed the various features of fonts with
respect to their readability, but also contains a large number of different, often con-
flicting, views on the impact of serifs, size and font style for readability. Results of
study [7] compares the readability of some common fonts by testing the reading speed
177
of texts in Russian. A higher reading speed for serif fonts is demonstrated. However,
no explicit font characteristics affecting readability are identified. The work [8] pro-
vides an overview of the situation of modern typography of textbooks and considers
contradictions of the current state with the font design to the rules of the current tech-
nical regulations. Lots of researchers consider serif fonts more legible and it is be-
cause of their serifs which add more information to the eyes [9] and enhance the legi-
bility of a text by helping the readers to distinguish the letters and words more easily
[10]. Results in [11], [12] indicated serif fonts are believed to be read faster due to
their invisible horizontal line made by serifs. Results of study [13] is against the
prominence of serif fonts. The space between letters in serif fonts is slightly reduced
due to the ornaments that they have. Consequently, as mentioned in [14], serifs act as
visual noise when the readers’ eyes attempt to detect the letters and words. The reduc-
tion of the space leads to other problems: One is a problem of crowding which is hin-
dering of letter recognition when a letter is flanked by other letters (cited in [15]) and
the other is that letter position coding may be hindered which decreases the ability of
word recognition [13]. The results of studies [15], [16] showed out equal legibility
and perception between serif typefaces and sans serif ones.
Thus, almost equal numbers of studies showed advantages and disadvantages of ser-
ifs, as well as a preference of other features of text. The preferences of specific font
features and font size are highly dispersed, too. It can be suggested that legibility is
more sensitive to some combinations of spatial features of text. No special type font is
suggested to use. The point to pay attention to is the familiarity of the subjects with
special typefaces and subjects’ preferences. The aim of this work is to find the way to
assess the spatial features of font drawings by using an objective scale invariant in-
dex.
2 Approach
An assessment of the visual characteristics of fonts represents certain difficulties as-
sociated with the difference in approaches to the understanding of what is a set of
visual characteristics and what criteria should be used in their assessment. The simi-
larity of some graphic elements of letters in font and the letters themselves, as well as
the font as a whole, suggests the possibility of using the ideas of fractal geometry to
make the assessment. A special case of the fractal dimension d is expressed by well-
known formula that combines the number of objects n, with which the measurement
is taken, and the geometric size of the object a:
d = log n : log a-1. (1)
Mandelbrot showed [17] that for fractal sets the expression relating the length of the
perimeter of the object P and its area S is performed:
P1/d : S1/2 = const, (2)
which implies that S ~ P2/d.
178
The fractal dimension can be understood as the degree of filling of the space by irreg-
ularly distributed substance. Thus, in either family of flat figures (like font), geomet-
rically similar but having different linear dimensions, the ratio of the length of the
shapes border to the square root of its area is a number that is completely determined
by the general form for the family. The equivalence of different linear extensions in
many cases is very useful [17]. The relation between abris’ length (perimeter P) of the
character or set of characters in the font and its area (S) can be used as a unique font
index. Considering the font as a coherent geometric set, by analogy with the way
proposed in [18], it is possible to apply the definition of compactness of the set C (3),
circularity coefficient Cc (4) and irregularity Cn (5) which is proposed to use as such
unique font index.
C = P2 : S. (3)
Cc = 4πS : P2, (4)
Cn = Cc-1 = C : 4π = P2 : 4πS (5)
Vector graphics software having an intrinsic macro language based on VBA can helps
to solve the task of index calculation. In the present work a public macro CurveInfo
for CorelDraw package is used. The macro calculates perimeter (in mm) and area (in
mm2) of a coherent vector object.
Fig. 1. Set of font letters and its division
As a representation of a font the full set of 66 uppercase and lowercase letters (for
Russian language) of each font is used (see Fig. 1). To obtain information about the
perimeter of a particular letter the perimeter of the external abris of a letter (Pout) must
be added to (if available) of the internal perimeter of the letter space (Pin). The perim-
eter of the full set of 66 letters (P) equals the sum of the perimeters of all letters (6).
To obtain an area of a letter it must subtract the internal area of the letter space (Sin)
from the general area bounded by the outer abris of the letter (Sout). The area of the
full set of 66 letters (S) is equal to the sum of the areas of all the letters (7).
P = Σ (Pout + Pin) (6)
S = Σ (Sout – Sin) (7)
3 Results and discussion
For the measurement 21 fonts (straight light drawing, sizes 12 and 18 pt) were select-
ed: 9 sans serif fonts, 11 serif fonts and a script font. For the selected set of fonts the
described procedures and formulas were applied. By the distribution of the values of
irregularity Cn for groups of serif and sans-serif fonts the mapping is undertaken.
179
Also, 5 fonts among the full set are used to test their reading speeds. Participants of
the experiment are 10 students. They read text samples with different font layout and
count the reading speed. The correlation between reading speeds and irregularities is
assessed. The results of calculations by formula (5) and reading speed measures for 5
fonts are given in Table 1. As it can be seen from Table 1, a irregularity has almost
constant values for each font. It indicates the objective nature of the scale invariant
index proposed that is useful for research. A small variation of the index for some
fonts can be explained by features of font scaling. Figure 2 shows the distribution of
irregularity for groups of serif and sans-serif fonts by peer review. Serif fonts are read
slightly faster than sans serif ones, on average. Script fonts have a low reading speed
but extremely high irregularity. Figure 3 shows the distribution of reading speeds and
its dependence from irregularity for 5 selected fonts, regression and confidence inter-
vals. Statistical analysis reveals strong negative correlation between reading speed
and irregularity (correlation coefficient –0,69, p<0,05).
Table 1. Irregularities and reading speeds.
No Font Feature Reading Cn, 12 pt Cn, 18 pt
speed, chars
per sec
1 sans-serif 305 305
2 sans-serif 418 418
3 sans-serif 439 416
4 sans-serif 459 464
5 sans-serif 470 470
6 sans-serif 43,2 481 469
7 sans-serif 575 575
8 sans-serif 35,5 605 605
9 sans-serif 825 825
10 serif 472 472
11 serif 585 585
12 serif 653 653
13 serif 33,6 675 655
14 serif 682 682
15 serif 703 704
16 serif 714 714
17 serif 778 792
18 serif 31,4 796 796
19 serif 875 875
20 serif 880 802
21 script 28,5 1717 1651
180
Fig. 2. Distribution of irregularity for groups (numbers) of serif (above) and sans-serif (below)
fonts by peer review
Fig. 3. Distribution of reading speeds and its dependence from irregularity for 5 fonts, regres-
sion and confidence intervals (drops).
4 Conclusion
Although there are massive bodies of analysis considering typography and font fea-
tures, there is no agreement among researchers regarding legibility factors in print.
One of the most complicated issue is accounting for the effect of font drawing on
legibility. The work offers a solution to this problem. The scale invariant index to
assess the spatial features of font drawing is proposed. Accounting for the index in
research of reading (e.g. [19]) might help to identify predictors of reading speed, as
well as quality of assimilation not only for paper, but also for electronic texts. In any
case, the scale invariance of the proposed index allows to accumulate experimental
results without any subsequent processing or conversion, which is convenient.
181
5 References
1. Artemov V.A. Technographic analisys of summarized letters of new alphabet. Writing and
revolution. 1933, 1, 58–76.
2. Ushakova M.N. The new typeface for newspapers. Polygraphic manufacturing. 1952, 4, 22–
23.
3. Ushakova M.N. The new typeface for narrative literature. Polygraphic manufacturing. 1952,
11, 26–28.
4. Ushakova M.N. The new typeface for large circulation narrative literature. Polygraphic man-
ufacturing. 1959, 11, 26–27.
5. Alexandrova N.A., Chiminova V.G. New typefaces for newspapers. Polygraphic manufac-
turing. 1962, 5, 57–61.
6. Tokar O.V., Zilbergleit M.A., Petrova L.I. Legibility of typefaces. Proceedings of the higher
educational institutions. Problems of printing and publishing. 2004, 2. 79–92..
7. Tokar O.V., Zilbergleit M.A., Litunov S.N. Font legibility assessment on the sample of offi-
cial document. Omsk scientific herald. 2009, 2(80). 246–249.
8. Tarbeev A.V. Font design of modern textbook. Proceedings of the higher educational institu-
tions. Problems of printing and publishing. 2004, 4. 93–102.
9. Geske J. (1996). Legibility of Sans Serif Type for Use as Body Copy in Computer Mediated
Communication. Presented at Annual Meeting of the Association for Education in Journal-
ism and Mass Communication 1996
10. McCarthy M.S. & Mothersbaugh D.L. (2002). Effects of typographic factors in advertising-
based persuasion: A general model and initial empirical tests. Psychology &Marketing,
19(7-8), 663–691.
11. Romney C. (2006). Improving the visual appeal of classroom handouts. JALT2005 Confer-
ence Proceeding. Tokyo: JALT. Retrieved May 21, 2012, from http://jalt-
publications.org/archive/proceedings/2005/E121.pdf
12. Arditi A. & Cho J. (2005). Serifs and font legibility. Vision Research, 45, 2926–2933.
13. Perea M. & Moret-Tatay C., Gómez P. (2011). The effects of inter letter spacing in visual
word recognition. Acta Pschologica, 137, 345–351.
14. Woods R. J., Davis K., Scharff L.V.F. (2005). Effects of typeface and font size on legibility
for children. American Journal of Psychological Research, 1(1), 86–102.
15. O’Brien B.A., Mansfield J.S., Legge G.E. (2005). The effect of print size on reading speed
in dyslexia. J Res Read, 28(3), 332–349.
16. Paterson D. G. & Tinker M. A. (1932). Studies of typographical factors influencing speed
of reading. X. Style of type face. Journal of Applied Psychology, 16(6), 605–613.
17. Mandelbrot B. Fractal geometry of nature. Moscow, Institute of computer studies, 2002,
p.656.
18. Alexanina M.G., Karnatsky A.Y. Comparison of the spatial characteristics of the fields of
sea ice and cloud according to the visible channel AVHRR / NOAA on the example of the
Okhotsk Sea / The Sixth All-Russian Open Annual Conference "Modern Problems of Re-
mote Sensing of the Earth from Space". Moscow, IKI RAS 10-14 nov 2008, 299–302.
19. Tarasov D.A. & Sergeev A.P. The leading as a factor of readability: development of the
methodology for educational use. Procedia – Social and Behavioral Sciences, Elsevier,
v.106, 2013. 2914–2920.
182