What is the benefit of variable fonts?

Question

Is there actually a typographical benefit to variable fonts? Or is it just the font companies trying to justify new purchases?

I remember Adobe Multiple Masters (and Apple's GX fonts too), which were the cause of all sorts of problems, but the utility of specifying 'slightly not quite as Bold'; 'slightly more compressed' than the supplied fonts seems questionable.

There's a risk of reducing contrasting between weights, making things more 'samey'; or exaggerating the contrast beyond aesthetic limits.

I can see there might be some occasional uses, like logo design, but for general use, might it leads to poor choices?

Answer 1

YMMV. I think there are benefits in uses such as video and animation, having letters you can easily morph is good there. I think a lot of the benefits in web design will be on crazy avant-garde stuff like some of the things Nick Sherman's done for Typographics, websites where it feels like the type is bubbling and twisting in front of your eyes, maybe uses that don't even exist yet.

If it's more conventional web, document and poster design, the particular benefit for my view is optical size options, you can have one file for all sizes. Google Fonts' article introducing Roboto Flex shows what you can do particularly well. They use a standard setting of Roboto for the body text and different variable font setting for the heading: by pushing up the x-height, stretching out the width and slimming down the weight it morphs into feeling sixties sci-fi, like the end titles of The Incredibles. It's really quite striking, it feels both distinctively different to the text and smoothly compatible with it.

For posters, I guess the benefit is being able to jam a big bold headline into exactly the right space, and maybe to crunch the width a bit if something changed at the last minute and you needed to scrunch in a bit of extra text. Wood type was sold in a huge range of widths for exactly that reason.

Often though designers, even ones who care about niceties like optical sizes, want a hierarchy with a separate typeface for each use: a display type for headings, a serif or neogrot for body text, a separate agate/caption font for small text, etc. These could all be catered for by one VF, but often you'd rather have different typefaces for them. We're not in the sixteenth century where the expectation is a Garamond-style serif at every single size and for every single use, although if you like that classical style you could do it with VF and make it look very very good.

For websites, compared to tracking garbage and images, the font size is a very small part of a website's payload, not very important.

The really irritating thing about some of the more complex variable fonts, at least ones I've seen, is that they haven't been marketed with a quick set of recommended styles that look good. That was my gripe with the big Google Fonts demos, Amstelvar and Roboto Flex: sure, everything is parametric, you can pull the ascenders longer or the descenders longer, but which combinations of settings look nice for what uses? It would be nice to see a quick highlights reel of nice combinations of settings.

Answer 2

the utility of specifying 'slightly not quite as Bold' […] than the supplied fonts seems questionable

Just to address this aspect in an example: Yesterday I created a small poster using a font coming in seven weights and yet I found myself wanting to fine-tune the weights on three occasions:

• Most of the text was dark on light background, but some was light on dark background. Since a line of the same width is perceived thicker if light on dark, I would have needed a slightly lighter font to match different parts of the text. Just going down one notch in the supplied weights was too much here.

• I needed to use some superscript letters. Since the font did not supply those natively, I had to use rescaled regular letters. Without any change, these were too thin to match the regular text (more on this), so I turned up their weight one notch, but this was too little, while two notches were too much. (Of course, with a font having native superscript for everything, this is not needed, but implementing this is arguably more work than creating a font with variable width.)

• I had a headline, which should have a given width (implied by the rest of the design). However, with this constraint none of the weights was harmonious with the rest of the text. Either the headline was too prominent or lacked prominence.