Benson Linear (Beta)

Jared Benson's picture

In the spirit of dusting off work in progress, here's a specimen of a face I'm working on. This is an early port I made to the Mac several months ago to see how it looks on screen- The actual font is being designed on my Palm Vx, and I can already notice changes from this specimen.

How small can you go and maintain legibility? By allowing the ascenders and descenders to define the word shape, the 4px x-height is surprisingly readable. I'm curious to hear your thoughts.

benson linear

hrant's picture

Jared, you're a freak for making us read that difficult text in that supertiny type... But what can I say - it works! The only thing I can spot is the exclamation point: too short.

BTW, it's cool how you're getting such differing color just by rounding the corners in one style but not the other. I guess the lighter one could serve as an "italics" of sorts?

hhp

jonathan's picture

I would have to disagree with David, I find the linear version easier to read....I think the linear letters coalesce better than their nonlinear counterparts.

All in all I would have to say that I like this font very much!

hrant's picture

> It would seem odd to have the exclamation mark ascend higher [than the question mark].

It would, but I don't think it is! :-)
And if it applies to bitmap fonts (where the unit is so huge), it certainly applies to outline fonts too! In my current design (my first Latin text face, I admit) I have the exclamation higher than the question mark (even though the reversed Spanish ones have the same span/location).

It's tempting to align the excalamation (and question) marks with the UC and/or ascender height, but in actual text setting, where does the mark usually end up? After a string of mostly x-height glyphs. In text setting apparent size is much more important than alignment (which I feel is generally a very superficial merit in a text face). Since the question mark has a much stronger top than the exclamation, it only makes sense for the latter to go taller.

Note: Thanks to OpenType you can now make exclamation marks of varying height, and specify which one to use when.

hhp

mart's picture

My only question is why, why, why?

hrant's picture

Why, why, why, what, what, what?

Why not align things as much as possible?
Conscious appreciation of display typography does admittedly give alignment some merit, but in the largely subconscious act of reading, alignment (of letterform components) is about as useful as compass to a dancer.

I even doubt the value of the baseline itself (the Ancient Greeks for one didn't use it nearly as much as those regularity-freak Romans). But: This does not mean I doubt the value of interlinear space - it's crucial for guiding saccades.

hhp

hrant's picture

> In Web use, it's better suited for things like copyright lines, ....

I think it's too readable for that stuff...

hhp

hrant's picture

> shouldn't a true italic at least have some slant to it?

Terminologically, I guess. Although Majoor's wonderful Seria Italics is a killer exception:
http://www.fontfont.de/packages/seria11178/serita29960/serita29960_xx.html
BTW, is it true that Seria has a *second* italics, somewhere?

But functionally, italics is about emphasis, and in this case slanting would only emphasize why slanting on-screen is a bad idea...

hhp

mart's picture

I was asking why design this face, but I can imagine it could well work on LCD devices like cell phones.
I totally disagree about it being good for web pages though. I have exceptionally good eyesight, something like 40-20, (or should that be 20-40?), and I have difficulty with it on screen. I think that has nothing to do with the traditional reasons that a font is not legible, but to do with the normal tendency of the brain to seek out familiar pattern - and the fact that at this size that just isn't happening. At least not for my brain! The type is so small that a kind of anti-anti-aliasing seems to taking effect.

hrant's picture

> 20-40

I think the first number has to be 20, and the second number is better lower. At my most recent eye test (about 3 years ago) I was 20/15. Maybe 40/20 is [somewhat] like saying 20/10, which is actually suspiciously exceptional! :-)

Here's something via Google:
http://www.milfordeye.com/vtest.htm

But note that those numbers refer to *feet*, and people have very different ranges. My range for example is extremely close: freeway signs are a little blurry to me, but during my eye test I undershot the equipment's minimum range! Probably from using computers for 23+ years now... Or maybe it was painting D&D miniatures by candle-light: the bombs would cut our electricity... :-/

Back to the font:
On the other hand, I think you're right that the font is too small for any normal definition of "comfort". But I also think that sometimes space is expensive enough to warrant using it (like the things Joe mentioned). It's a niche thing, which *does* make it valuable.

hhp

beejay's picture

we did some eye-chart style t-shirts last year and ran across this link:

http://www.cascadilla.com/eyechart.html

interesting letterforms

hrant's picture

I think it looks like a mistake... sorry.
You'd need to go ClearType for an oblique at that size.

hhp

hrant's picture

David, you sparked an idea:
Instead of slant, maybe this design should use width as the emphasis method for its "italics"?

hhp

hrant's picture

> italic is typically more narrow than a roman

Sure, but in this case the roman is already as tight as possible, so the "italic" *has* to go the other way, no?

> This might make it closer to a "bold", though.

Actually, making it wider [without increasing the stem weight] would make it [slightly] lighter.

hhp

fonthausen's picture

Hi jonathan,
really neat!!

Jacques

hrant's picture

Pretty cool!
I like the caps being upright, and the lc "a" solution is very interesting.

hhp

Jared Benson's picture

Here's a new version, refined and improved. Your thoughts?

benson linear

Jared Benson's picture

Here's a rough comparison of Benson Nonlinear and Benson Linear:

benson_nonlin

Jared Benson's picture

I was wondering the same thing about the exclamation mark. It disappears when seen in context here- however, when compared to the question mark, they are the same size- It would seem odd to have the exclamation mark ascend higher. I'll have to play with it. When working with such tiny bitmaps, one pixel makes a huge difference.

And as for the italic, that was an idea I had as well - it definately would serve as emphasis, but shouldn't a true italic at least have some slant to it?

anonymous's picture

Is it just me, or is the Nonlinear form considerably more readable than the Linear form?

Both are great designs, but the Nonlinear characters seem to coalesce into words much more readily than the heavier, blockier Linear characters.

I suppose they could be used as companions, but if it were me, I would use the Nonlinear much more than the Linear, I think.

Regarding the punctuation, I think I'd like to see both the exclamation mark and the question mark extended by a pixel, even though it sacrifices the elegeance of the caps-height punctuation. (Perhaps making two sets -- a font with caps-height punctuation, and one with ascender-height punctuation?)

Joe Pemberton's picture

Mart, are you asking why would Jared create this?

Apart from being an interesting experiment, it's
actually quite functional. Kudos Jared.

I think it's one thing to see this font set in a
block of text on a computer monitor but quite
another to see it perform on a Palm device. It's
very efficient and surprisingly easy to read on
a Palm.

In Web use, it's better suited for things like
copyright lines, legal-ese and other tiny
acoutrements.

//joe

anonymous's picture

By Nonlinear, I mean the sample shown on the left, and by Linear, I mean the sample shown on the right. I hope I have the correct names with the correct samples.

Having said that, and with respect to Jonathan Armstrong's opinion, I still find the sample on the left (Nonlinear, I presume) to be more readabloe than the sample on the right (Linear?).

That bit of clarification aside, kudos indeed to Jared for a supremely interesting experiment well executed. I would be proud to have either of these fonts appearing on my own website. (Someday. :) )

David

Jared Benson's picture

Just playing around, but here's an "oblique" version, which I'm starting to think might have some potential. Any thoughts?

oblique

anonymous's picture

Try opening it (the italic) up by expanding it horizontally by a pixel or two... That might help. As is, it isn't reading terribly well.

David

anonymous's picture

That's an interesting thought.... Which should be wider, though, the unemphasized "roman" or the emphasized "italic"?

The reason I ask is because an italic is typically more narrow than a roman, but to achieve emphasis in this font, perhaps wider would be more effective? This might make it closer to a "bold", though.

Perhaps, on the other hand, the emphasized text should be set in the narrower width for consistency with commonly recognized characteristics of italics, and not to be confused with a bold.

(Musing out loud... or whatever one calls musing while typing... :) )

David

hrant's picture

For the design constraints, it's pretty good!
At that size you'd normally use caps, but you never know what could be useful, so...

Very smart to sacrifice some x-height discrimination in favor of boumas (word shapes). But the clogging in "e" and "g" (and in "a", "s" and "z" to a lesser extent) is a problem.

Even in a 4pix x-height there are ways to avoid clogging, although it leads to some unconventional forms. For example, a 4pix bicameral "a" is one where the bottom row of a conventional 5-pix bicameral "a" is dumped (and the same could be done for the "e", but at the top). For the "g", I'd go with the Koch form (again).

It's tempting to fall back on mono forms for "a" and "g", but the bi forms are what would make Benson Linear unique, and I'm glad to see you going for them.

As for the caps, I think there are some legibility issues (like the "B"). BTW, since you have 2 pixels of descender, have you considered taking the caps one pixel lower than the lc baseline? I think they can afford to be bigger anyway.

Anyways, keep it up!

hhp

Syndicate content Syndicate content