How to put IPA on the Web --
a simple hack

Dafydd Gibbon, U Bielefeld

1 October 1997

Contents

 
• The problem
• Solution Part I - an eggbox for carrying IPA symbols
• Solution Part II - automatic insertion into the eggbox
• Conclusion
• PostScript version
• versions: original; eggboxes
• Summary of Prosody List correspondence on this article.

The problem

Web browsers which support the standard symbols of the International Phonetic Association (IPA), for example with Unicode characters, are not in general use, but the need for rendering IPA symbols is rather urgent for web publications and teaching materials. A simple procedure, which is better than the scanning process which is sometimes used, was developed in order to render IPA symbols in the hypertext version of the recently published Handbook of Standards and Resources for Spoken Language Systems, edited by Dafydd Gibbon, Roger Moore and Richard Winski, published by Mouton de Gruyter, Berlin (1997). The procedure was required for inserting a large number of IPA macros into the HTML version of a formatted book, since the widely used latex2html hyperdocument generator software, in the version and configuration available, was not able to handle the problem. A sample is contained in Table 1.

 

IPA SAMPA ASCII Open- Front- Rounded- Comment

Close Back Unrounded

a a 97 open front unrounded

A 65 open back unrounded

æ { 123 near-open front unrounded English bad

6 54 near-open central unrounded German Butter

Q 81 open back rounded

O 79 open-mid back rounded

e e 101 close-mid front unrounded

E 69 open-mid front unrounded

@ 64 mid central unrounded schwa

3 51 mid central unrounded

i i 105 close front unrounded

I 73 near-close front unrounded lax

o o 111 close-mid back rounded

ø 2 50 close-mid front rounded

9 57 open-mid front rounded

& 38 open front rounded

u u 117 close back rounded

U U 85 near-close back rounded lax

} 125 close central rounded

V 86 open-mid back unrounded

y y 121 close front rounded

Y 89 near-close front rounded lax

Solution Part I - an eggbox for carrying IPA symbols

An obvious solution for rendering IPA symbols in HTML documents is to use in-line transparent GIF images for each symbol or combination of symbols; conversion of single symbols was targeted in the present application. What is not so obvious is how to generate the GIFS automatically and insert them into complex texts formatted in . The GIFs generated by the procedure described here can be imported into proprietary web assistants and HTML editors if required, and of course converted into other graphics formats.

The IPA macros used in the EAGLES application are those contained in the wsuipa macro set, but the procedure operates with any special character macros (e.g. the \LaTeX macro, which is rendered as ).

The method uses a property of the latex2html script which passes certain objects which it cannot handle, such as figure and math environments, to other programmes, effectively using the following cascade:

1. Call latex to generate DVI format.
2. Call dvips to generate PS format.
3. Call pstogif to generate GIF format.
4. Call giftrans to generate transparent GIF format.
5. Generate HTML inline image element.

The size of the image is controlled by the .latex2html-init configuration file; this file provides the following scaling parameters which can be modified to produce an acceptable text/figure size ratio:

# This number will determine the size of the equations, special characters,
# and anything which will be converted into an inlined image
# *except* "image generating environments" such as "figure", "table"
# or "minipage".
# Effective values are those greater than 0.
# Sensible values are between 0.1 - 4.
$MATH_SCALE_FACTOR = 1.4;

# This number will determine the size of
# image generating environments such as "figure", "table" or "minipage".
# Effective values are those greater than 0.
# Sensible values are between 0.1 - 4.
$FIGURE_SCALE_FACTOR = 1;

The IPA rendering procedure uses these features of latex2html by inserting the IPA macros into a math environment with a simple macro which I call eggbox. Packed into an eggbox, the symbols are then safely transported through the latex2html conversion cascade described above, and emerge as inline GIFs of the correct size:

Macro definition: \newcommand{\eggbox}[1]{$\mbox{#1}$}
Macro example: \eggbox{\eggbox{\schwa}}

The latex2html call for converting this document into a zero page depth hyperdocument (with minor manual post-editing) was:

latex2html -html_version 3.0 -t "IPA on the Web" -split 0 ipa2eggbox.tex

Solution Part II - automatic insertion into the eggbox

The \eggbox macros were automatically inserted into the original file using the standard UNIX script language sed:

#!/bin/sh
# ipa2eggbox, Dafydd Gibbon, 5 May 1997
# IPA macro set is wsuipa
# Convention: macro may be followed by one or more spaces.
# Caution: check that macro name is not a prefix of another macro name.
# Macro \eggbox definition: \newcommand{\eggbox}[1]{$\mbox{#1}$}

cat $1 |
sed "s/\\\schwa */\\\eggbox{\\\schwa}/g
     s/\\\scu */\\\eggbox{\\\scu}/g
     s/\\\sci */\\\eggbox{\\\sci}/g
     s/\\\inva */\\\eggbox{\\\inva}/g
     s/\\\invv */\\\eggbox{\\\invv}/g
     s/\\\niepsilon */\\\eggbox{\\\niepsilon}/g
     s/\\\scy */\\\eggbox{\\\scy}/g
     s/\\\stress */\\\eggbox{\\\stress}/g
     s/\\\eng */\\\eggbox{\\\eng}/g
     s/\\\esh */\\\eggbox{\\\esh}/g
     s/\\\eth */\\\eggbox{\\\eth}/g
     s/\\\openo */\\\eggbox{\\\openo}/g
     s/\\\nitheta */\\\eggbox{\\\nitheta}/g
     s/\\\dz */\\\eggbox{\\\dz}/g
     s/{\\\ipa{\\\symbol{[^}]*}}}/\\\eggbox{&}/g
     s/\\\diaunder\[\\\syllabic|m\]/\\\eggbox{\\\diaunder\[\\\syllabic|m\]}/g
     s/\\\revepsilon */\\\eggbox{\\\revepsilon}/g
     s/\\\downp */\\\eggbox{\\\downp}/g
     s/\\\curlyc */\\\eggbox{\\\curlyc}/g
     s/\\\invh */\\\eggbox{\\\invh}/g
     s/\\\invscr */\\\eggbox{\\\invscr}/g
     s/\\\nj */\\\eggbox{\\\nj}/g
     s/\\\invscripta */\\\eggbox{\\\invscripta}/g
     s/\\\baro */\\\eggbox{\\\baro}/g
     s/\\\invy */\\\eggbox{\\\invy}/g
     s/\\\scr[^i] */\\\eggbox{\\\scr}/g
     s/\\\curlyz */\\\eggbox{\\\curlyz}/g
     s/\\\baru */\\\eggbox{\\\baru}/g
     s/\\\scripta */\\\eggbox{\\\scripta}/g
     s/\\\yogh */\\\eggbox{\\\yogh}/g
     s/\\\nigamma */\\\eggbox{\\\nigamma}/g
     s/\\\glotstop */\\\eggbox{\\\glotstop}/g
     s/\\\secstress */\\\eggbox{\\\secstress}/g"

Conclusion

Until browsers appear which provide renderings for IPA Unicode, the automatic inline GIF here can fill the gap reasonably satisfactorily. If and when such browsers appear will depend, among other things, on the efforts and influence of the spoken language community.

Footnote:

...procedure

This document can be freely distributed in any format. Acknowledgment would be appreciated.