Unsupported characters in the symbol font

[Randolph]

I'm working with an old PDF file which has a lot of mathematics in it, and I'm running into problems with the symbol font. For instance, the following equation:

Is being imported as:

As you can see, most of it comes through properly but for some reason the the parentheses in the Symbol font do not. (The other parentheses are in Times, and they're fine.) Is there a fix for this other than manual correction of each formula? At least a way to search for these unsupported characters?

[lacerto]

(...)

[Randolph]

In fact it seems these can be copied and located with "find," so I now have a tedious but plausible way of dealing with them. I would much prefer not converting the whole thing to curves - the book is 700 pages long!

Affinity? I think this may be a bug, or at least an obscure feature.

[Joachim_L]

Could you upload a sample page PDF, so that we all can have a look?

[Randolph]

Sure, here's the whole page with that equation I dragged out of the document with Preview (I am using a Mac.) However, the page by itself looks correct when I open it in Affinity Publisher. I wonder what goes wrong when I import the whole book. I suspect that part of the problem is how very old the original file is; it probably dates from the early 2000s.

RwR, p495.pdf

[Joachim_L]

On Windows the page looks even worse, perhaps due to the fact, that I do not have the original fonts installed. But there seems to be a problem, because the Font Manager is telling me that Unsupported characters are used with SymbolPS. Additionally weird tracking values are used on this equation. So I fear you have to do some extra work.

[v_kyr]

• Mathematical Symbols in Unicode

[Randolph]

Joachim L: I think so. Thanks for looking into it.

v_kyr: The book's mathematics is set in a mix of Times and Symbol; the book's text is set in Adobe Garamond Pro. I am hoping to replace Times with Garamond, but I am not sure I can and wonder if I even ought to try; even with modern tools mathematical typesetting has many pitfalls.

The Symbol font is 35 years old, long predating Unicode. It has its own character encoding. Apple has its own Symbol font which uses a slightly different encoding. So, confusion for the developers of a layout program which came along much later.

[lacerto]

(...)

 

[speters33w]

Have you checked TeX math packages? It is very possible the original document is not Unicode compliant.

It is also probable if it's very old it was generated with a version of TeX.

An older TeX font might fix the problem.

Or try one of these.

https://ctan.org/pkg/unicode-math?lang=en

You might also try the old Adobe 256 glyph Symbol font. If you can find an old Windows Acrobat 5 Reader installation package, it includes a type 1 postscript version of the original symbol with its original mapping and glyph names.

Edited June 2, 2020 by speters33w
Additional stuff

[Randolph]

The original equations were probably set with eqn and troff, though I'm not sure which versions of those decades-old programs were used. The equations are mostly set in a combination of Times and Symbol. I am hoping to substitute Garamond for Times to save on fonts, but we'll see. Right now I am just clearing up the problems with Symbol.

As you say, resetting the whole document would be a lot of work!

The PDF import tends to scatter the mathematics through multiple text frames. One thing that would help would be operations that allowed one to combine and split text frames, but I have yet to find them. Have I missed them?

[speters33w]

@Randolph Check your messages on this forum

[lacerto]

(...)

[Randolph]

I have worked out some answers. I'm going to give them here, and in separate posts. I would like to see these incorporated in the FAQ.

1. The "locate" function of the font manager will locate unsupported characters, which may then be replaced. Unsupported characters may not be what you expect, they may look normal, but the font manager will find them for you.
2. When importing PDF files with significant mathematical content, turn off  "Favor editable text over fidelity" and "Group lines of text into text frames." This much reduces your chances of getting scrambled mathematics. If you have a document that contains extensive text you need to edit as well as mathematics, consider doing both imports, and then moving pages from the less scrambled one into the one with editable text.
3. Also, when importing mathematics, it sometimes helps to import single pages by themselves. I am not sure why. It may be an obscurity in the PDF conversion process, or an actual bug.

[Randolph]

Followup: the macOS "preview" app, which is among other things a PDF viewer, will correctly read and export the mathematics in my document, and Publisher imports the resulting document without fuss. Unfortunately, its export leaves me with the pages all 8½ × 11 inches in size, and these would have to be cut down in Publisher. So I'm doing individual pages when I need to.

So with this, and the information in my previous comments, my problem is solved. Thank you all for the advice and encouragement.