Dealing in Symbols

I know: two questions in one post:

a. Is there a fast way to combine two symbol fonts into one? I have Part A and Part 2 and want to cojoin them into one long font. Each has 255 glyphs.

b. If I am creating a symbol font from gif images would my font look better if the Units Per Em were (4096*4=) 16384? (Currently, they’re set at a default 2048). Would I end up with a more detailed font?

This has arisen from trying to create a font from very small images which has to be exploded to be readable. I can enlarge the images before importing them or I can have FC Scale them in Glyph Transformer. In either of these actions the resulting character is “squared off” - losing rounded-ness.

The FC solution looks better; but I’m wondering if there is something better?

a. As far as I know, symbols fonts are limited to 255 glyphs. That leaves you with few options if you wish to combine them in one font.

i. Start a new font and paste the glyphs from each symbol font into that. However, which codepoints will you use and how will you type these symbols?

ii. Try my non-standard trick of making one symbol font a bold variant of the other font with the same name. Use the bold attribute in applications to access the second range of symbols. I’m not sure if this works with symbol fonts, but it might.

b. I think you would probably get the best results by enlarging the images before importing them into FontCreator. Corel Draw or Adobe Illustrator might be better for autotracing and smoothing the images than FontCreator.

I don’t think increasing the units/em will help. Higher values would help if the symbols are intended for use at poster sizes and contain fine details not visible at normal sizes. If your originals were tiny, then you have a lot of work to do if you want to use them at poster sizes. More funits/em just means a finer grid. If your symbols are pixelated, that means you already have too many points. See my sig.

a. Yes, you are right - 255 seems the limit.

MS Word only shows 255 when you use “Insert/Symbol” and the font is a Symbol font. Now I don’t understand why the series of numbers 61473+ doesn’t end 255 points later? Why would they be needed? Must be some method to their madness.

Are there no symbol fonts with more than 255 glyphs? Am surprised at that too.

So I created new font as Unicode character set. Copied and pasted 500 characters to the new font – the solution to my combining question. Since the default New Font is mapped to 653 glyphs, other than having postscript names (which can be corrected-see below), each of my 500 had a valid mapping id to assume.

MSWord showed all 500 as a regular text font of course. Tried to Insert/Symbol using the same font and THAT worked correctly. (Had to display Character codes from Unicode (hex) point of view, not ASCII.)

In conclusion, generate them as text fonts and everything works correctly, generate them as symbol fonts and some things don’t work correctly. … Never did like Symbol fonts anyway!

b. I don’t have Corel or Adobe but used Irfanview and Dopus to make the characters larger. Neither did as well as FC. My symbols are hand drawn characters so don’t really have fine details; but FC made any pixel edges into triangles whereas the other two made them into rectangles. The rectangles look like missing pixels whereas the triangles look like rough edges.

c. New question: is there any easy way to get Post Script Names into all 500 without going through each one Properties/Postscript Name, type, Enter, Select next one, and repeat? (which could be enhanced if the cursor started at Postscript Name when you bring up Properties… but it’s still 500 somethings. Maybe I should learn to use Macros…)

Thanks for the good thoughts. Dick

c. Using a macro recorder would certainly make it a little easier, but you can do it all from the Format, Post dialogue with a few keystrokes.

Default focus is in the list of glyphs.

  • Down arrow selects the next glyph
  • Tab gets the cursor to the postscript name field where you can type the next postscript name. This assumes that you’re able to recognise the symbols easily in the dialogue box and know the names that you wish to use.
  • Shift tab gets the cursor back to the list of glyphs
  • Down selects the next glyph
  • Tab gets the cursor to the postscript name field, etc.

The sequence: Shift tab, down, tab, can be assigned to a single keystroke such as Shift Enter, which speeds up input considerably.

Existing names, if there are any, can be stripped by deselecting the “Include postscript names” checkbox, closing the dialogue, then reopening the dialogue and reselecting it. Generate will add back the .notdef .null and nonmarkingreturn names. Default postscript names can be selected from the droplist.

a. Yes, you are right - 255 seems the limit.

I read somewhere that it is 224 of which the first might need to be a space. These are mapped into the Unicode Private Use Area at U+F020 to U+F0FF.

Now I don’t understand why the series of numbers 61473+ doesn’t end 255 points later?

Decimal 61473 is hexadecimal F021. It seems that the Symbol Font way of doing things allows a font to be treated as if being from U+0020 to U+00FF within applications, so that one may access symbols using ordinary keyboard characters and some of the accented characters of some of the languages of Western Europe, whilst maintaining a situation that the glyphs are, within the font, mapped to the Unicode Private Use Area.

Some readers might like to know that the Microsoft Calculator program which arrives with Windows is, in the View | Scientific mode, useful for converting from decimal to hexadecimal and from hexadecimal to decimal. Microsoft Calculator also has capabilities far greater than its keypad suggests. For example, it can calculate the arctangent of a value by checking the Inv checkox then clicking tan. Right-clicking on a key and then left-clicking on the What’s This? panel which is displayed provides details of the facilities which the key can be used to produce.

So I created new font as Unicode character set. Copied and pasted 500 characters to the new font – the solution to my combining question.

Yes. However, please remember that this technique is limited by the fact that some (many?) applications will assume things about the characters based upon the Unicode code point value, so whereas overlaying glyphs for symbols into code point positions used for Latin letters may well be fine, overlaying glyphs for symbols into code point positions used for Arabic letters might lead to an application assuming that they are from right to left. Also, if a code point which is in regular Unicode (as contrasted with a Private Use Area code point position) is unallocated, some applications may refuse to accept that that code point is available for use.

Whilst recognizing that having symbols keyable as ordinary letters of the alphabet is helpful, and I have done it myself, I tend to use the Unicode Private Use Area quite a lot. For example, the symbols in my Quest text font.

http://www.users.globalnet.co.uk/~ngo/QUESTTXT.TTF

I have been experimenting with an idea of mine to produce a pdf using Open Office 2.0 (it’s a free package from the http://www.openoffice.org/ webspace) such that the pdf is a type tray for symbols from a font. This would mean that an end user of a font could copy symbols from such a pdf and paste them into an application, as a sort of simulation of hand setting with metal type and having to get out a type tray to find a particular sort. I am thinking of producing such a pdf for my Chronicle Text font, a type tray of the ligature characters which are located in the Private Use Area.

http://www.users.globalnet.co.uk/~ngo/CHRONTXT.TTF

William Overington

Yes, for Symbol fonts there is a 224 character limit.
http://www.high-logic.com/fontcreator/manual/charactertoglyphindexmappi.html