Tactile Graphics in Braille Texts

The aim of our paper is to present and describe the research project on tactile graphics at the University of Quebec at Montreal, in progress since 1990, which is part of a larger project concerned with the standardization of braille in french.


Tactile Graphics in Braille Texts¹

1  Introduction

The aim of our paper is to present and describe the research project on tactile graphics at the University of Quebec at Montreal, in progress since 1990, which is part of a larger project concerned with the standardization of braille in French. A brief history of the overall proposals for standardization will be followed by a detailed description of the project with special emphasis on a first experiment with visually handicapped students.  We will conclude by pointing out a certain number of parameters which will be treated in future experiments.

2  History

In the early 80’s the Quebec Ministry of Education began funding the production of braille material, first for university-level students, then gradually, for students at the secondary and elementary levels. The increased production of didactic material in braille made standardization imperative. This standardization included on the one hand the format of the system and on the other, specific symbols that were not available in the initial braille system. In 1985, the Ministry of Higher Education and Science became involved in the standardization process and began to fund research projects that led to the publication of the first volume of Code de transcription de l’imprimé en braille in 1990. The work was officially launched in Quebec in May 1990, and in France in October, 1991.

On the one hand, the Code de transcription de l’imprimé en braille adopted the methodology contained in the Code of Braille Textbook Formats and Techniques, and on the other hand, it contains innovations: since it was conceived in the context of the French language, it took into account abbreviations specific to that language as well as the symbols and typographical rules that govern the writing of texts in French.

In the first volume, the Code de transcription de l’imprimé en braille did not cover all the elements to be standardized in the braille transcription. That is why, since January 1990, thanks to new funding by the Ministry, we are working on Phase 2 of the process of standardization of the format of French language documents. Among the elements to be treated in this phase are: notes and references, color, and tactile graphics. At this point, we are concentrating all our efforts on tactile graphics.

Individuals and organizations have been making more and more insistent demands that tactile displays (when they are available) should be included in the braille volume. With techniques using aluminum foil and thermoforming[2], the tactile displays are found either at the end of the braille volume (if they are not too numerous), or in supplementary volumes. Users certainly recognize their usefulness, but they also find manipulation cumbersome and slow. Furthermore, the lack of norms[3] for aspects as basic as footnotes and references from the braille text to tactile displays and vice versa makes the process of consulting and working with these documents more difficult. Moreover, the no less serious lack of norms for determining the design and production of tactile displays gives the visually handicapped little autonomy in the use and correct interpretation of the message being communicated. Finally, the producers of tactile displays find the storing of this basic material (matrices) extremely cumbersome.

3  Objectives

Within the general context of standardization, and taking into account the needs of users, the research project on tactile graphics has established as its goals:

      • to determine norms for the production of graphic material amenable to tactile decoding;
      • to constitute a computerized data base of graphic elements consistent with the defined norms.

4  Research procedure

The research project on tactile graphics comprises the following steps:

      • the examination and evaluation of the equipment available for the production of tactile graphics;
      • a review of the literature on the subject;
      • an inventory of the types of graphics present in textbooks used at the elementary and secondary levels[4].  Textbooks in mathematics, biology, ecology, geography, nature studies, initiation to technology, etc., were examined.

5  Evaluation of available equipment

5.1  The braille printer

Since publications in braille are produced (in the case of the text itself) by braille printers, the most direct method for the production of tactile graphics to be included in the text seemed to be by the use of the braille printer. Therefore, it was the first piece of equipment examined and tested as to its graphic potential with regard to computerized image processing.

The results of these tests were generally negative. It should be noted, however, that the positve aspects of this printer are that it allows for the use of a very satisfactory paper quality and the inclusion of graphics and text, or simply alphabetical and numerical information. However, the mere attempt to reproduce straight lines, each having a different angle, was sufficient to demonstrate the impossibility for the braille printer (even when an excellent graphics program was used), to reproduce with precision basic geometric forms other than the square and the rectangle. In every case, there was distortion in the drawing of the line. Since precise rendering of shapes was impossible, it seemed useless to attempt the production of tactile graphics with the braille printer.

5.2  The Pixelmaster printer

After we examined a few samples of tactile illustrations produced by the Pixelmaster printer[5], we decided to acquire one due to the interesting potential we discerned in this product. The Pixelmaster is not (as such) intended as a printer of tactile information. It is first of all a color printer. However, the various overlays of ink coating produce reliefs that are perceptible to the touch and measurable. The minimum relief is around 0.045 mm (1.7 thousandth of an inch), and the maximum relief 0.225 mm (8.9 thousandth of an inch).  We worked with the hypothesis that this gradation of reliefs, even though limited and inferior to the relief of the standard braille point (13 to 15 thousandth of an inch) offers possibilities which merit consideration, especially with regard to the available alternatives.

Our experiments with the Pixelmaster and the braille printer went from the simple to the more complex. We therefore attempted the production of sets of straight lines with modified angles and basic geometric forms of various sizes. These initial tests produced satisfactory results compared to similar tests made on the braille printer. The Pixelmaster allows for a highly precise rendition of the tactile image. Furthermore, the parameters of vertical relief of the line, the width of the line and the surface textures each present individually a range of options, and a combination of these increases the possibilities available. However, the Pixelmaster presents certain disadvantages in its present version:

      • it only allows the use of 21.6 x 27.9 cm (8.5 x 11 in.) size paper, of standard thickness;
      • it modifies the texture of the line depending on its orientation, especially when maximum relief (8.9  thousandths) is used. Nevertheless, at this stage in our research, the advantages of the Pixelmaster outweigh its disadvantages. In its present version, the Pixelmaster allows for:
        • precise rendition of lines
        • three-levels of relief
        • a large range of textured surfaces
        • considerable flexibility at the level of line widths
        • the insertion of alphabetical and numerical notes and even lines of text in braille

To confirm our observations and to verify that we were on a course acceptable to users, we conducted an experiment with a group of learners in June 1991. This experiment was necessary to determine if the work on representation and standardization of tactile graphics should be pursued. It would be safe to assume that there cannot be a complete dichotomy between the tool being used to produce tactile material and the norms to be created, proposed and applied. Hence the importance of a well-adapted and sophisticated technology that will answer to the well-defined needs of the users of tactile graphics in diverse situations.

6  Experimentation

 6.1  Object of the experiment

To verify the capability of subjects to differentiate by touch differences in relief of printed lines.

Basic hypotheses:

(a) the minimum relief produced by the Pixelmaster printer (1.7 thousandth) is discernible by touch;

(b) the more the relief is emphasized, the faster the tactile recognition;

(c) the orientation of the lines does not interfere with the perception of the relief;

(d) it is possible to distinguish the three levels of relief of lines produced by the Pixelmaster printer;

(e) the quality of the paper influences the tactile perception of the relief of the lines : on satin-finish paper, tactile perception is easier, and consequently faster.

6.2  Subjects chosen for the experiment

Selection criteria

The subjects were chosen according to the following criteria:

      • They were at the time students at either the second level of elementary school, the first or second level of high school, or at college level;
      • They had no handicap other than blindness;
      • They had used braille as a regular working tool (reading and writing) for at least two years.

There were 25 subjects (16 girls and 9 boys) who agreed to take part in the experiment and who answered to these criteria. The minimum age was 10 years, the maximum 23, and the average was 16.7 years. With regard to schooling, they were distributed as follows:

1 attented second level in elementary school

20 attented in secondary school:

2 in level I
3 in level II
6 in level III
1 in level IV
3 in level V
3 in special classes
1 in continuing education
1 in an alternating school/job program

3 were in college

1 was in post-collegiate studies

The 25 subjects followed courses in 11 different institutions spread over 4 different regions of Quebec:

      • 2 were from Montreal
      • 2 were from the Laurentian region
      • 11 were from the South Shore of Montreal
      • 10 were from the Quebec City region

6.3  Procedure[6]

Two experimenters together met each subject, either at the school they attended or at their home (this choice was left entirely up to the student). The experimenters met with the student and identified themselves.  For each of the 42 blocks of questions and figures :

      • they slowly read, each in turn, the same question (the question was therefore heard twice);
      • they reminded the subject that he had to explore the figure using only one hand, while the other remained in place;
      • they placed the sheet on the table, facing the subject;
      • they placed the subject’s hands directly on the figure;
      • they timed the subject’s response.

When all the figures had been examined, they asked a number of questions concerning the subject’s identity (name, age, school level, school attended, and name of school commission).

In order to insure a rigorous and constant procedure during the entire experiment, which was spread over a two-week period, the experimenters agreed to the distribution of tasks between them before the first interview and maintained that arrangement for all meetings (reading the questions, handing out the figures, placing the subject’s hands on each figure, noting the answers using the chronometer, noting the answering time).

6.4  Material for the experiment

6.4.1  The figures

42 figures constitute the core of the experiment. Each figure is drawn on a different sheet and is made up of a group of three lines in relief. The sheets are size 8.5×11 in.

All the lines are exactly the same length (4.72in.), the same width (7.9 thousandth) and are spaced at equal distance from each other (0.39in.)[7].

The oblique lines are oriented at a 45 degree angle. The arrangement of the lines on the page is identical on all the sheets, irrespective of the orientation of the line. The number of the figure on each sheet is written without relief (so as not to interfere with the tactile decoding measured in the experiment) and corresponds to the number of the question. Two types of paper are used; matte finish (Roland Concorde bond 10M) and satin-finish (Roland 10M long-grain white LT 2000).

The experiment is constructed around 21 combinations of lines. But since in the fifth hypothesis tactile decoding is associated with the quality of the paper, the 21 sets of lines are presented on both matte-finish and satin-finish paper. The figures are presented randomly, since the presentation order is not meant to be systematic.

Of the 21 basic sets of lines, 9 each present a set of 3 lines of uniform relief:

      • 1 set of horizontal lines, 1 set of vertical lines, and 1 set of oblique lines, with a maximum relief (8.9  thousandth);
      • the same three sets with intermediate relief (3.3 thousandth);
      • the same three sets with minimum relief (1.7 thousandth).

The 12 other combinations of lines each present a set of three lines of mixed relief, that is, the three reliefs that correspond to the potential of the Pixelmaster printer.  These 12 sheets include :

      • 4 horizontal sets of lines;
      • 4 vertical sets of lines;
      • 4 oblique sets of lines.

These initial patterns printed on matte paper are printed in identical fashion on satin-finish paper.

6.4.2  The questionnaire

Parallel to the 42 tactile figures produced with the Pixelmaster printer, 42 blocks of questions are associated with the figures and successively read to the subject. Each block of questions concerning one figure is made up of three elements: one main question, a sub-question 1, and a sub-question 2. At the beginning of each main question, the figure is described. Here is an example of a main question:

Question 1: Here are 3 horizontal lines (that go from left to right).  Do these lines have the same relief, that is, do they all rise to the same height on the sheet?

a) yes
b) no

If the subject answers in the affirmative, the experimenter goes on to the following block of questions and to a new figure.

However, if the subject gives a negative answer, the experimenter continues with the same figure and reads the first sub-question. Here is an example:

Tell me, which line has the least relief?

c) the bottom line ?
d) the middle line ?
e) the top line ?

Then the experimenter goes on to the second sub-question. Here is an example:

Tell me, which line has the most relief?

f) the bottom line ?
g) the middle line ?
h) the top line ?

6.5  Results of the experiment

6.5.1  Data 

The data accumulated from the answers to the 42 blocks of three levels of questions by 25 subjects may be examined from a number of different angles. Within the framework of this article, we will limit our comments to those aspects that directly correspond to the five basic hypotheses of the experiment.

It is however interesting to point out, that in general, the average number of errors per subject is relatively low, and that the errors made were concentrated in only a few subjects.

The average number of errors for each category of questions is the following:

main question
sub question 1
sub question 2


Consequently, the success rates are:

95.1%, 95.5%, and 96%

Taking into account that

      • for the main question,
        2 subjects had errors totalling 23.8%
      • for sub-question 1,
        4 subjects had errors totalling 11.9%
      • for sub-question 2
        3 subjects had errors totalling 4.5%,

the average number of errors, excluding the marginal subjects, is 3.09%, 3.09%, and 2.6%. The rate of success, then, rises to 96.1%, 96.1% and 97.4%.

6.5.2  First hypothesis

There is no doubt that the minimum relief (1.7 thousandths) produced by the Pixelmaster printer is discernible to the touch. This is clearly demonstrated by the small number of errors. The average number of errors on the main question is 6.7%. Furthermore, if we associate the parameter of relief with the response time parameter, confirmation of the hypothesis is reinforced.  In fact, of the four sets of lines examined from the point of view of relief, the one that is most quickly recognized is the sets of lines with the minimum relief (1.7 thousandth).

intermediate relief
(3.3 thousandth)
11.48 seconds
maximum relief
(8.9 thousandth)
11.44 seconds
minimum relief
(1.7 thousandth)
09.46 seconds
multiple reliefs
(1.7 / 3.3 / 8.9)[8]
08.03 seconds


6.5.3  Second hypothesis

The preceding demonstration disconfirms the second hypothesis that predicted that the more the relief is emphasized, the faster the tactile recognition. The experiment indicates that the subjects took almost two seconds more to identify sets of lines with maximum relief than to identify similar sets of lines with minimum relief.  (See table in section 6.5.2.)

6.5.4  Third hypothesis

The third hypothesis stated that the orientation of the lines does not interfere with perception of relief. This hypothesis, which concerned two parameters: the orientation of the lines and relief, cannot be verified because of the limited number of figures having only one relief. However, the results of the data analysis concerning the orientation of the lines can be compared to what we sensed was a weak point of the Pixelmaster printer. This, namely, is its inability to keep constant the texture of line, whatever the orientation. The results indicate that the lines that seem the easiest to recognize are vertical lines, to which the Pixelmaster printer gives a rougher texture. A study of the average number of errors analyzed from the point of view of the orientation of the lines shows the following gradation :

vertical lines
0.79 errors
horizontal lines
1.14 errors
oblique line
1.71 errors


We are obliged then to note that the subjects made more than twice the number of errors with oblique lines than with vertical lines.

We must point out that the orientation of the lines does not affect response time. The average time for correct answers by questions category is as follows:

Main question

oblique lines
10.02 seconds
vertical lines
9.08 seconds
horizontal line
8.54 seconds


Sub-question 1

oblique lines
6.19 seconds
vertical lines
5.47 seconds
horizontal line
5.77 seconds


Sub-question 2

oblique lines
4.94 seconds
Vertical lines
4.20 seconds
horizontal line
3.90 seconds


6.5.5 Fourth hypothesis

The success rate of the subjects for the whole of the experiment a priori confirms the fourth hypothesis. Parallel to the first hypothesis, (see section 6.5.2), the results of the experiment indicate that it is clearly possible to distinguish the three levels of relief produced by the Pixelmaster printer. This can be verified not only on the basis of the rate of correct answers to the sets of lines with one level of relief, but also on the basis of the rate of correct answers to the sets of lines with multiple levels of relief. The maximum error percentage recorded is 10.7%.

6.5.6  Fifth hypothesis

The experiment does not confirm the hypothesis which stated that the quality of the paper influences the perception of the tactile relief of the lines. The rate of correct answers for the sets of lines presented on matte paper versus satin-finish paper does not show any significant differences despite a slight superiority observed with the use of satin-finish paper. By combining the two factors, response time/type of paper, the results obtained do not lead us to conclude that the paper used has a determining influence on the results.  The results are as follows:

Type of paper
Main question
9.67 seconds
8.76 seconds
Sub-question 1
5.85 seconds
5.47 seconds
Sub-question 2
4.46 seconds
4.23 seconds


6.5.7  Additional remarks

The amount of data accumulated in the course of this first experiment is considerable. There are numerous possible angles from which one may observe, analyse, and reflect on this data. Therefore, the interpretation of the information gathered from those subjects who took part in the experiment is being pursued and will help us to increase and widen our knowledge pertaining to the data obtained. No doubt we will be able to review these at a later date.

7  Future prospects

The deliberately restrictive parameters used in the framework of this first experiment with the Pixelmaster printer (levels of relief, type of paper, (straight) lines and their orientation) are the basic elements necessary to the development of tactile graphic displays. To the observations already gathered from this experiment, we will have to add parameters such as : line width, line spacing, the variable forms of broken and dotted lines, the tracing of forms and their sizes, surface textures, etc. Only the evaluation of such parameters taken individually, then combined, will allow for a precise estimate of the potential for tactile graphic production by the Pixelmaster printer.  Such an evaluation could perhaps influence, if not reorient, its development for the greater benefit of users.

Nicole Trudeau Ph.D. (UQAM)

Colette Dubuisson Ph.D. (UQAM)

Traduction en anglais: Florence Blouin



[1] We wish to thank Estelle Bérubé and Daniel Daigle who carried out the experimentation as well as Lucie Huberdeau who was responsible for the statistical manipulation of the data.

[2] For more details on these techniques, see Vermy (1969), Foulke (1990), James (1975), among others.

[3] For more details on the need for norms, see James (1972), Jansson (1984), (1987), and (1988), among others.

[4] We have deliberately excluded, at this stage, the examination of texts used at higher levels.

[5] For technical comments on this printer, see Miastkowski (1990).

[6] The procedure was tested by the experimenters on the person responsible for the project and on research center colleagues before being used on the subjects.

[7] Braille system spacing: from point 1 of one cell to point 1 of the cell directly below it (on the next line).

[8] Whatever the order of the reliefs.



Code of Braille Textbook Formats and Techniques / published by American Printing House for the Blind, Louisville, Kentucky, 1985.

FOULKE, E. / Methods for the Production of Tangible Graphic Displays / (Unpublished Paper presented at Kyoto in July 1990).

Gouvernement du Québec / Code de transcription de l’imprimé en braille, tome I / 1989, 374 p.

JAMES, G. /  « A Kit for Making Raised Maps » / in The New Beacon, vol. LIX, no 696, 1975, pp. 85-90.

JAMES, G. / « Problems in the Standardization of Disign and Symbolization in Tactile Route Maps for the Blind » / in The New Beacon, vol. LXVI, no 660, 1972, pp. 87-91.

JANSSON, G. / « Wich Are the Problems with Pictures for the Blind and what can be Done to Solve Them? » / in The Framework of the Concerted Action on Technology and Blindness (Paper read at a Workshop on Man‑Machine Interfaces, Graphics and Practical Applications, Maastricht, The Netherlands, November 14‑16, 1988).

JANSSON, G. / « An International Pool of Useful Symbols for Tactual Pictures » / in Special Education, May 11‑15, 1987, Wenner‑Gren Center, Stockholm (A proposal presented at the Unesco ‑ Wenner‑Gren Center, Regional seminar on New Technologies for Handicapped).

JANSSON, G. / « Research Needs and Production Concerns, Research Needed to Obtain more Useful Tactual Maps » / in Aids and Appliances Review, no 14, 1984, pp. 3-6.

Miastkowski, Stan / Printer Generates Tactile Graphics for the Blind /  in:  Byte, March 1990, pp. 24 et 28.

VERMY, G. J. / « Observations on Raised‑line Drawings » / in Education of the Visually Handicapped, 1, 1969, pp. 47-52.


Article publié dans: 

World Congress on Technology Conference Proceedings / 1-5 decembre, 1991, Arlington Virginia / vol. 4 pp. 382-400 / Nicole Trudeau Ph.D. et Colette Dubuisson Ph.D. / Tactile Graphics in Braille Texts.

Sur des sujets apparentés :

Tactile Graphics in Braille Texts (en français)

La normalisation du graphisme tactile 

Vers la normalisation du graphisme tactile

Envoyer un commentaire

Votre adresse e-mail ne sera pas publiée.