InDesign Nerdery

[ñͤͣ̄ͦ̌̑͗͊͛͂͗ ̸̨̨̣̺̼̣̜͙͈͕̮̊̈́̈͂͛̽͊ͭ̓͆ͅé ̰̓̓́ͯ́́͞]

Posted here to avoid spagging up QG's thread.

Alternatively, whether a font is embeddable or not, is a matter of ONE BIT in the .ttf or .otf or whatnot file. I once had some commandline ttf tweaking tool that could flip it. I forgot what it's called. It did mention that you should not do it if you were doing it in order to embed a font. But IMO it's kind of ridiculous how the "font" can not be embedded (because it's technically a piece of software for purposes of copyright) while its outlines can be embedded (because they're merely the output of this software), I know they can contain all sorts of complex kerning rules (though a LUT is hardly software IMO) but then it's pretty obvious most of QG's fonts do not

This is where typography nerds will make a seemingly retarded distinction between a "font" and a "typeface". "Font" refers to the software or complete set of physical objects (for example, in letterpress those backwards, cast metal letters are called "sorts", the complete set of sorts is the font) which are used to create a design. A "typeface" refers to the output, whether that's a print, image on a computer screen, or what have you. Fonts are tools, typefaces are the finished visuals.

Here's why it's a meaningful distinction: it's almost never practical to take a typeface and backwards engineer it into a usable font. First of all because it's pretty rare to find designs that use every little glyph (trying to design missing glyphs is surprisingly difficult). Second, you'd have to figure out the side bearings, assuming you could even find something with every letter, number, punctuation mark and so on. A lot of people assume this is done mathematically, but it's not. Side bearing information is included in the font software and is designed just as painstakingly as the typeface.

Basically, converting fonts to outlines makes it more trouble than it's worth to try to steal. I've heard and read conflicting accounts as to whether this is still relevant these days, and my impression is that it's getting phased out (that said, I'd still convert to outlines if I were doing work for a company with enough money to attract lawyers). Outlining is generally reserved as a work-around for technical issues such as buggy TrueType fonts and outdated RIP software. All of the font information is included when you embed it, so it's possible to hack it out and have a usable font, however, this information is destroyed by converting them to outlines.

Usually it's best not to convert to outlines as this increases your file size and your text is treated like an image by your computer—so no searching through the document for a word or phrase, no selecting quotes to copy and paste. Also, it destroys the hinting information for your computer screen—so if the PDF isn't going to be printed and you aren't having any technical issues, you've degraded the quality of your PDF for no reason. The only time you should outline your fonts is if your printer asks you to do so, to get around software bugs or restrictions that are preventing your PDF from exporting or printing properly, or to do some artsy shit where the letters are all at funky angles or sliced apart or whatnot.

[Triple Zero]

Very interesting! Thanks for posting this.

Alternatively, whether a font is embeddable or not, is a matter of ONE BIT in the .ttf or .otf or whatnot file. I once had some commandline ttf tweaking tool that could flip it. I forgot what it's called. It did mention that you should not do it if you were doing it in order to embed a font. But IMO it's kind of ridiculous how the "font" can not be embedded (because it's technically a piece of software for purposes of copyright) while its outlines can be embedded (because they're merely the output of this software), I know they can contain all sorts of complex kerning rules (though a LUT is hardly software IMO) but then it's pretty obvious most of QG's fonts do not

This is where typography nerds will make a seemingly retarded distinction between a "font" and a "typeface". "Font" refers to the software or complete set of physical objects (for example, in letterpress those backwards, cast metal letters are called "sorts", the complete set of sorts is the font) which are used to create a design. A "typeface" refers to the output, whether that's a print, image on a computer screen, or what have you. Fonts are tools, typefaces are the finished visuals.

Here's why it's a meaningful distinction: it's almost never practical to take a typeface and backwards engineer it into a usable font. First of all because it's pretty rare to find designs that use every little glyph (trying to design missing glyphs is surprisingly difficult). Second, you'd have to figure out the side bearings, assuming you could even find something with every letter, number, punctuation mark and so on. A lot of people assume this is done mathematically, but it's not. Side bearing information is included in the font software and is designed just as painstakingly as the typeface.

What's "side bearings"? Is that like the kerning information? As in, where a glyph should be rendered relative to the other glyphs?

If so, I'd assume part of it could be done mathematically, to get most of it mostly right, and then they'd need to be tweaked for visual appeal manually by a human.

Except for real complex fonts with big swashes and such, I can imagine they'd need to be done manually entirely cause an algorithm would get hella confused.

Also, afaik, you don't just kern individual glyphs, but you need to kern pairs of them, since the spacing depends on what they're next to. Hence me talking about a LUT, coding speek for Lookup-Table.

Basically, converting fonts to outlines makes it more trouble than it's worth to try to steal. I've heard and read conflicting accounts as to whether this is still relevant these days, and my impression is that it's getting phased out (that said, I'd still convert to outlines if I were doing work for a company with enough money to attract lawyers).

Oh, totally. If I were a gfx designer, I'd do it even for smaller companies, just out of sense of professionalism.

Outlining is generally reserved as a work-around for technical issues such as buggy TrueType fonts and outdated RIP software. All of the font information is included when you embed it, so it's possible to hack it out and have a usable font, however, this information is destroyed by converting them to outlines.

Got it.

Usually it's best not to convert to outlines as this increases your file size and your text is treated like an image by your computer—so no searching through the document for a word or phrase, no selecting quotes to copy and paste.

That's odd because I think PDF has the ability to contain both. I've seen scanned ebooks that are obviously bitmaps, low res even, and you could select and search the text. Maybe they put the text in some random simple font "behind" the scans, or PDF has the option to say "this block contains this text" regardless of whether it's rendered from a font.

BTW even if the PDF text *is* made from an embedded font, selecting/copying text and searching for phrases (or converting the whole PDF to plaintext) is a really really tricky job. This is because a block of text in PDF is represented as a separate box object for each line. And if a character in that line is kerned differently, this line object box is broken up in pieces and the kern character is a single box. And since each box contains its coordinates relative to the page, they could theoretically appear in random order in the file, and still get rendered correctly. Add to that the fact that you can have columns, captions and pull-quotes.

A program that wants to allow you to select text from a PDF as a continuous piece of data (while it's actually just boxes that happen to be near eachother) is going to need a shitload of heuristics and guessing to determine which boxes belong together and which just happen to be close. That is why you often get parts of pull-quotes and captions when selecting text from a PDF. This really could have been designed better, but yeah, standards eh!

Also, it destroys the hinting information for your computer screen—so if the PDF isn't going to be printed and you aren't having any technical issues, you've degraded the quality of your PDF for no reason. The only time you should outline your fonts is if your printer asks you to do so, to get around software bugs or restrictions that are preventing your PDF from exporting or printing properly, or to do some artsy shit where the letters are all at funky angles or sliced apart or whatnot.

Well, hinting is getting less important these days. It's great to have it for system fonts in your applications and browser, but with monitors being higher resolution and font pixel sizes getting larger, naive anti-aliasing does the trick 8 times out of 10. In fact I get the idea that most PDF viewers implement their own graphical implementation of the hinting information, and they don't always get it right. The fonts rendered natively by the OS are nearly always better hinted than those in a PDF, in my experience.

My experience does not include Macs btw. Just Linux and Windows. Since I'm working from Windows today, I can tell you it has slightly more hinting than Linux and I'm not sure I like it because the letters appear kind of flimsy and the ClearType configuration thing was a wizard that would just ask me "what looks better?" without letting me tweak the actual settings

I'll be glad to be back on Linux soon :P

[ñͤͣ̄ͦ̌̑͗͊͛͂͗ ̸̨̨̣̺̼̣̜͙͈͕̮̊̈́̈͂͛̽͊ͭ̓͆ͅé ̰̓̓́ͯ́́͞]

I just wanted to drop you a line saying I enjoy geeking out with you and will continue this nerdfest when I have a moment to post at full spag. My brain is starting to come off the rails, and I think it's best I go enjoy it where people can't see me.

(Why do animals have to sleep? This is bullshit.)

[Q. G. Pennyworth]

So, if this is the place for InDesign nerdery should I be dropping my booklet template here? I dunno how many people it would be useful for, but I need the damn thing to print anything out right.

[Triple Zero]

Sure, feel free I need to acquire InDesign before I could check it though, but maybe it's useful for others.

[ñͤͣ̄ͦ̌̑͗͊͛͂͗ ̸̨̨̣̺̼̣̜͙͈͕̮̊̈́̈͂͛̽͊ͭ̓͆ͅé ̰̓̓́ͯ́́͞]

What's "side bearings"? Is that like the kerning information? As in, where a glyph should be rendered relative to the other glyphs?

If so, I'd assume part of it could be done mathematically, to get most of it mostly right, and then they'd need to be tweaked for visual appeal manually by a human.

Except for real complex fonts with big swashes and such, I can imagine they'd need to be done manually entirely cause an algorithm would get hella confused.

Also, afaik, you don't just kern individual glyphs, but you need to kern pairs of them, since the spacing depends on what they're next to. Hence me talking about a LUT, coding speek for Lookup-Table.

Side bearings essentially are the default, built-in spaces on either side of a glyph. Ilovetypography.com has a great entry on it here. When you adjust the tracking you're adding or subtracting space from the side bearings.

Correct, side bearings are approximated by an algorithm and finessed by a human, but they are distinct from kerning pairs. IIRC, kerning tables only correct specific letter combinations that would otherwise result in awkward spaces, while side bearings are designed more globally—with the overall letterspacing in mind. They're a lot more tricky to settle upon since you're trying to take into account so many variables.

Usually it's best not to convert to outlines as this increases your file size and your text is treated like an image by your computer—so no searching through the document for a word or phrase, no selecting quotes to copy and paste.

That's odd because I think PDF has the ability to contain both. I've seen scanned ebooks that are obviously bitmaps, low res even, and you could select and search the text. Maybe they put the text in some random simple font "behind" the scans, or PDF has the option to say "this block contains this text" regardless of whether it's rendered from a font.

Some scanners have a particular button you can press for scanning text, but this is just a scan that gives you an image and then it's converted into text with optical character recognition software. You could do the same thing with a regular scan (which results in an image with no text information) and then using Acrobat's OCR function. It always fucks up though, even on nice clean scans.

When you embed the font, you don't have to dick around with OCR to get it back into a selectable, searchable form. Most people only have Acrobat Reader which I'm pretty sure does not have OCR. So when you convert to outlines you make the PDF less useful to most people as a digital file.

Also, it destroys the hinting information for your computer screen—so if the PDF isn't going to be printed and you aren't having any technical issues, you've degraded the quality of your PDF for no reason. The only time you should outline your fonts is if your printer asks you to do so, to get around software bugs or restrictions that are preventing your PDF from exporting or printing properly, or to do some artsy shit where the letters are all at funky angles or sliced apart or whatnot.

Well, hinting is getting less important these days. It's great to have it for system fonts in your applications and browser, but with monitors being higher resolution and font pixel sizes getting larger, naive anti-aliasing does the trick 8 times out of 10. In fact I get the idea that most PDF viewers implement their own graphical implementation of the hinting information, and they don't always get it right. The fonts rendered natively by the OS are nearly always better hinted than those in a PDF, in my experience.

My experience does not include Macs btw. Just Linux and Windows. Since I'm working from Windows today, I can tell you it has slightly more hinting than Linux and I'm not sure I like it because the letters appear kind of flimsy and the ClearType configuration thing was a wizard that would just ask me "what looks better?" without letting me tweak the actual settings

I'll be glad to be back on Linux soon :P

It's kind of a best practices thing to make it accessible to as many people and devices as possible. If you compare the way a PDF renders outlined type versus embedded fonts on a few different PDF viewers I think you'll still notice a difference, even in this age of higher quality computer screens.

[ñͤͣ̄ͦ̌̑͗͊͛͂͗ ̸̨̨̣̺̼̣̜͙͈͕̮̊̈́̈͂͛̽͊ͭ̓͆ͅé ̰̓̓́ͯ́́͞]

So, if this is the place for InDesign nerdery should I be dropping my booklet template here? I dunno how many people it would be useful for, but I need the damn thing to print anything out right.

By all means.