BTW, they also redesigned their web site at a cost of $28 million. That's what the first link is.

Jonathan,

Can the Typophiles have an exclusive sneak peek today to build some buzz?



TYPOPHILE
Membership Has Its Privileges


bj

Brian,
Thanks for the heads-up! It's also helpful to own a pre-redesign issue, for comparison. Along with the press release I'd give up a few pounds for a large digital setting - mostly to measure ink gain against a hi-res scan of the actual output. Please?...

BTW, in terms of content my fav daily has gotta be The Herald Tribune.

hhp

Some of us lucky bastardos that were able to attend Jonathan and Tobias's lecture at the Walker last month got a little sneak preview ;)

Minneapolis has it's priviledges.

err privileges.

Who the hell has time to read a newspaper nowadays? And who wants all those old mountains of newsprint lying around in untidy piles? I don't even have steady work, but I don't have the time or patience for newspapers. Sorry - I love print, but I've never understood the appeal of newspapers. Give me a weekly mag like The New Yorker or something like The Atlantic any day. Perhaps if I'd ever seen a beatiful paper in my life I might feel different and have some time for them.

The *type*, my man - where's the *type*?! :-)

hhp

Re: Lacava.
Thanks for the reference, but The National Post online:
http://www.nationalpost.com/
is busy, busy, busy with information and design elements so I'll have to try to find it at a big newsstand.
While Lacava Design at:
http://lacavadesign.ca/
is Flash-based and it wrongly detects that my plug-in is not up to date, not letting me in. So I don't think much of her web talent.

The IHT is one of the only sites I actually read
on screen. It's very thoughtfully done.

Stephen

the typeface in question would be in the stock listings, fyi.
i guess jonathan hasn't gotten around to that press release yet. didn't see anything on typography.com.

WARNING: I wouldn't want to turn this thread into a discussion of press gain (at least not entirely so), but please allow me this digression of sorts:

> the typeface in question would be in the stock listings

Oh, good thing you told me - I almost scanned the wrong section. "Scanned?!", you ask? Well, I've been obsessed with measuring press gain in newspapers for a while now. This is partly because Nour (Patria's Armenian counterpart) is supposed to end up as the body face for a daily newspaper. Another reason is my fascination with trapping - something not usually needed but potentially very beneficial in a newspaper, *especially* for tiny type like in stock listing...

So what I would love to be able to do with TFJ's new font is to superimpose a small snippet of hi-fi outline data over a hi-res scan of actual output. A big favor to ask, maybe, but something I will do once I have the scan for you guys.

In the meantime, you can check out the following images:

A scan from my "target" newspaper, next to a rendering, with measurements:
http://www.themicrofoundry.com/other/armnews2.gif
Point size: 8

This one is not my own comparison; it's from an article by Gurtler & Mengelt in Visible Language:
http://www.themicrofoundry.com/other/latnewsp.gif
Point size: they don't say.

Now, the Armenian sample clearly exhibits much more gain than any western newspaper would find itself stuck with, and even the second image is probably too drastic these days, but as long there's *some* gain, I'd like to measure it.

BTW, in case you'd like to see how bad it can get, here's a staggering overlay of Mandel's masterful Colorado font (used for phone books):
http://www.themicrofoundry.com/other/armnews2.gif
Point size: ~4

hhp

Sorry, that last link was supposed to be:
http://www.themicrofoundry.com/other/mandel.gif

hhp

> Well, I've been obsessed with measuring press gain in newspapers for a while now.

Then I think you will be very interested in HTF's solutions to this very problem! I hope he posts a sample with the press release.

By the way Jonathan, if you are reading, I would like to thank you for a very inspiring, if not "insightful" lecture that you gave at the Walker. You could hear the audience squirming in their seats when you said that you used fontographer's auto-hinting rather than manually hinting your fonts.

cheers.

-delete-

- delete - *

Jonathan and Tobias. Very nice. An economical font for a financial page.
And saves the paper money, too. Thanks for the posts.


bj

* the above was a belated April Fools thing.

What a nice press release. And what a nice font. Those fractions are genial - and phenomenally legible at that size - as are the alphabetics. And the font has 10 "uniwidth" weights too, wow. But I'm a sucker for technical wizardry... On the other hand, I even like the name!

BTW, am I correct in noting that the paper quality is a bit lower now than before, or is that just within expected variance? If it is lower, then that makes Retina even more impressive. Another coolness is the trapping. Look at the bold "N" in the zoom view - it's clearly superior to the old Helvetica results. And the characters with small bowls (like the "R" - not shown there) have that distinct Bell Centennial thing going on.

Anyway, if nobody at THF minds, I'd still like to scan and post a couple of more views of Retina: one showing a short paragraph, the other at very hi res. Tomorrow.

TFJ + THF: Keep it up!

hhp

I was going to give away the comparison to Bell Centennial, but i thought that would be telling.

OK, first a caveat: my scanner is only 600dpi optical, so anything greater has been interpolated - but it *is* an Agfa. The thing is, 600 is beyond the absolute resolution-fidelity of the reproduction anyway. *However*, gain can still be measured.

The first is at 1200dpi, and it shows the good effects of the killer trapping, in some key letters. It also shows how the lc "g" has been reduced in the x-height to make more room for its descender.
(~250K) http://www.themicrofoundry.com/other/retina_1.gif
Note: I've applied some gentle Levels, Unsharp Masking, and Posterize.

The second is at 2400dpi, and this is the one I would like very much to be able to superimpose the outlines. If anybody could pass this request to TFJ for his consideration, I'd be very grateful.
(~300K) http://www.themicrofoundry.com/other/retina_2.gif
Note: This is a raw scan, just posterized down to 128 shades.

The third shows a rare text block in Retina (600dpi). WSJ is till using Helvetica for most of their tiny text blocks - but maybe this was a test to see if Retina could replace it wholesale.
(~200K) http://www.themicrofoundry.com/other/retina_3.gif
Note: Tweaked as #1 above.

----

What's really interesting about Retina is this:
If you consider that optimization for small sizes and press gain goes hand-in-hand with a certain "funkiness" in letterforms (like severe trapping, and exagerated counters - look at the lc "e"), it looks to me like Retina actually has a lot more room "down below". In effect, among the three lower limits (reproduction fidelity, font features and reader acuity), Retina removes the font issue: it could be used even much smaller than in the WSJ, were it not for limit #1, but especially #3.

hhp

JH:
>Despite Mr. Stalin's nanofont goofiness above

(Actually, it's Comrade Joe, we don't go in for that "Mr" stuff)

You mean it's not true!? But that's what my source says, and he's never wrong...

JH:
> I do think that these kinds of atomic
>metaphors are useful for agate type...

For sure, it's similar to some of the trickery you do with type or logos for lowres bitmaps for screen/web design where you don't have the pixels to really do what you need, so you perhaps create shapes/counters "outside" the shapes you're actually "shaping" to trick the eye into seeing something that isn't really there.

And of course the kind of tricks played with sub-pixel positioning.

The issue with small type like this is that the distortion to type isn't relative to it's size, it's pretty much a constant. Where 0.05mm ink spread makes no difference at 72pt, it's sure going to have an effect down there at 6pt.

Is this the one? It's Bell Centennial, "Bold Listing"

In the BC that I find the cent isn't so special:
http://www.myfonts.com/CharacterInStyle00a2-12177.html

But the *per*cent certainly is!
http://www.myfonts.com/CharacterInStyle0025-12177.html

And the dollar sign is extra special too:
http://www.myfonts.com/CharacterInStyle0024-12177.html

(BTW, I'm still digesting your previous post...)

hhp

Forgot to add:
The "structural" stuff is indeed the true challenge. A comparable challenge on the "local"/trapping side is the great benefit of being able to *automate* it. I have this device called the Trapping Flower*, and a third-party is actually working on a FontLab script for it now, so...

* http://www.themicrofoundry.com/ss_trapping1.html

BTW, I know I'm sick, becuase I dream of MM fonts with a "gain" axis, where the degree of trapping could be varied continuously.

hhp

Right, that's it with the cent sign there. the key thing is you need to open up the closed counter, so Carter's done that by trimming down the vertical bar. I have to say though that I'm a bit surprised that there's no taper to the ends of the vertical bar into the body of the character.


Joe

That's only the most obvious kind of structural "sensitivity". Retina goes far beyond that.

hhp

> www.typography.com/press/wsj_after_retina.gif

First, thanks for that scan. It's much much better than mine (no surprise).

I've taken the liberty to do two things with it:


1. Crop out this string:

http://www.themicrofoundry.com/other/retina_4.gif

If TFJ is indeed "listening", I'd like to revise/downgrade my request for outlines to just those three letters (at that weight, of course). If it doesn't happen, no biggie - if it happens, super nice of him.

BTW, it's interesting to note that some of the trapping is still visible - a sure indication that Retina has more room "below".


2. Splice together a comparison of pre and post numerals:

http://www.themicrofoundry.com/other/retina_5.gif

I'll note again that the quality of the paper/reproduction seems lower in the Retina sample: look at the decimal points (which I'm presuming are rigid rectangles in the outlines).

But the point of this comparison is the *structures* of the numerals, and Retina's are so superior to Helvetica's (while remaining stylistically homogenous), it's not even funny.

----

Thinking some more about what I previously wrote concerning the three "lower limits" of legibility (with respect to size), I realized that the untapped potential of Retina (I mean untapped by WSJ) isn't really in terms of size (especially considering that the average WSJ reader is more likely to have faltering vision, being more "senior"), but more in terms of fidelity (or actually, lack of it): TFJ has made a font which can work well in much poorer reproduction conditions than we're seeing here. Now, a trapping axis would be *so* nice, but at least its [apparent] absence gives us something to look forward to!

> Tobias likened these strangenesses to particle theory

Not to be too literal, but let's not forget that we're physical beings in a physical world!

I in turn might draw a parallel between the states of matter (solid, liquid, gaseous, plasma) and the states of reading (here with respect to size). I once sketched this rough graph:

http://www.themicrofoundry.com/other/optic_01.gif

It tries to show a correspondence between size and optical compensation (here I mean things like more open counters). Above a certain point size (nominally 24 here) there is no need for compensation at all. Below a certain point size (like 5), the extenders [have to] become so short that you'd break down and use all-caps - at that size immersive reading is virtually impossible anyway. And the segment of least slope (here between 8 and 12) is the most delicate part, where boumas (word shapes) have to be just right. I could elaborate more, but you might be bored senseless already, and it's getting late over here...

So let me try to wrap this up:
The reason for the non-linearity of the graph is due to the "fighting-sisters" nature of Legibility versus Readability: they come into play differently depending on type size (among other things, of course), and together they give up three states of reading:
1. Below comfortable immersive reading: here legibility is king (queen?), and x-heights have to swell, for example.
2. Immersive reading: readability is more important than legibility - and for example ITC-style x-heights ruin everything.
3. Above comfortable reading: legibility is a non-issue (although what might be called "decipherability" is key), and readability is moot. What counts most here? What I call "drama" (avoiding the potentially misleading "beauty").

hhp

Were you thinking I forgot? :-)

So I scanned the Retina specimen from "Language, Culture, Type"*, and did the compositing myself:
http://www.themicrofoundry.com/other/retina_4a.gif

* http://www.atypi.org/lct/ Worth every penny.

It's not the best quality (and I hope I chose the correct weight - the Agate Five), but hopefully it gives you some idea.

Thinking about this some more, I feel more comfortable in restating that Retina goes further than it needs to in this specific commission/application (the WSJ stock listings), but that *does* mean it can survive even worse repro conditions and/or smaller sizes.

hhp

The current issue of Print (LVI:III - aren't Roman numerals just grand?...) has a superb treatment of the WSJ redesign, including before/after comparisons of Helvomita/Retina. It also has an interesting article by Paul Shaw about type design that only the designer uses (including the work of Gerrit Noorzij) - meaning on purpose. On the other hand, it also has an article about some awards competition set in... Comic Sans. Bejeezus.

hhp

Ooops. Print flip-flopped the hedcuts of Hoefler and Tobias. At least in the mag. we rec'd on the Left Coast.

More on hedcuts.

http://www.google.com/search?hl=en&ie=ISO-8859-1&q=%22hedcuts%22+%22Wall+Street+Journal%22&btnG=Google+Search

http://www.npg.si.edu/exh/journal/

bj

Print flops ALOT eh. whats with that. why why why? Comic Sans. hahahah can't wait to see.

great link on the journal bj

Yeah BJ, the one I saw was also correct.

hhp

hmmm. Very odd. I will get a scan up soon.

*

bj

Also BTW, the redesign will feature a new font from Tobias Frere-Jones, developed at a cost of less than $28 million. There'll be a press release about it on typography.com tomorrow.

If you're in the market for a beautiful paper, check out the National Post, designed by Lucie Lacava. It's quite something.

I'm surprised no one has posted this yet.



See the new WSJ at: http://tour.wsj.com

Though I admit to not remembering the old
Wall Street Journal Site, I'm not all that
impressed with the new one. I'm assuming they
spent 27.9 M on the behind-the-scenes stuff.
(http://online.wsj.com/)

The International Herald Tribune site is much
more visually appealing. (http://www.iht.com/)

Getting to it presently. Hang in there.

-deleted-

We're up:

http://www.typography.com/press/pr_04_09_02.html




ps. And, thanks, Kevin!

Hrant:
> it could be used even much smaller than in
>the WSJ

Right, I heard that Jonathan and Tobias were approached by Motorola not long ago (who'd got a tip-off from the focus group sessions for the WSJ), to make a font suitable for use on microprocessors - ie at the 0.n micron level.

Basically it's a real challenge because you're effectively dealing with fairly wobbly molecules at that level, rather than nice sharp pixels. The issues of ink spread pale into insignificance.


Joe

Retina does work successfully even smaller -- we've proofed the fonts down to two point (!), and they still read comfortably. Of course, at this size you're really depending upon reliable printing. Sheet-fed offset on coated stock is a little more hospitable than the maelstrom of the WSJ print facility.

Here are some additional high resolution scans of Retina and its predecessor. These are true 1,200 dpi scans, unadjusted. (They're also about 2mb each, so download them at your own peril.)

www.typography.com/press/wsj_after_retina.gif
www.typography.com/press/wsj_before_retina.gif

At some point I'd like to rescan these in color, but these should do for now.

Despite Mr. Stalin's nanofont goofiness above, I do think that these kinds of atomic metaphors are useful for agate type. Like Bell Centennial, Retina has a lot of burlesque moves that are incomprehensible in large sizes but essential in small ones. In explaining this to Joe and David at the Journal, Tobias likened these strangenesses to particle theory, noting that when you reduce things down to an atomic scale, the laws of physics break down. I thought it a good comparison.

True, true. There are also a lot of structural things you can do that go beyond ink traps and other "local" treatments -- perhaps someone here can post an illustration of the cent sign from the boldest of the Bell Centennial fonts. It's diabolically clever, and once again a reminder that Matthew Carter has all the good ideas. The fink.

That's not the cent sign I was thinking of; either it's different in one of the other weights, or I'm thinking of the wrong font altogether. When I get to the office tomorrow, I'll see if I can figure this out.

BJ, are you very sure of that? In my copy, the "hedcuts" for JH and TFJ are correct.