ATF rights?

pablohoney77's picture

who would own the rights to old ATF/Mackellar designs? anyone? i don't wanna step on any toes...

Thomas Phinney's picture

Legally, any design patents are long ago expired (ATF had a lot of these, and many can be seen in the ATF archive at Columbia).

In the USA, I understand there is no copyright on abstract typeface designs per se. But even if there were, the copyrights on much of the ATF stuff would have expired years ago. All works created prior to 1923 are in the public domain. Additionally, for works published between 1923 and 1963, the copyright owner would have to have filed an extension with the copyright office, else they would have expired by now. (Most of this info is taken from: ; note that I am not a lawyer and this should not be taken as legal advice.)

So, what other rights are there? Well, there are the typeface names, which may be trademarked. I gather trademarks can last indefinitely if maintained. It's my recollection that some non-typographic people in California ended up owning the Kingsley/ATF trademarks after ATF died off. Sorry I don't have more info.

Again, I'm not a lawyer, and there's no good substitute for real legal advice, which this is not.



pablohoney77's picture

thanks for the info Thom. So do i need to find a specialized lawer? how much does this kinda work usually cost?

raph's picture

I'm interested in this topic, as I've been tracing some of the classic ATF designs, as well as trying to reverse engineer, if you will, the way ATF did optical scaling in the metal days. I'll be posting some samples soon, but there's a bit of info in my blog.

hrant's picture

> trying to reverse engineer, if you will, the way ATF did optical scaling in the metal days.

Huh, that's something I've been working on too! And I'm a CompSci guy as well...

Based on actual character set waterfalls of ATF Garamond (which I printed at the ArcheType press last year), I've taken measurements of 11 parameters across sizes from 6 point to 48 point, and one day :-/ I'll write a paper explaining the findings from my dissection.

BTW, I think ATF is the best target to "reverse engineer". ATF Garamond in particular makes for superb study since it's a design that pushes the limits of optical scaling versus "design integrity". One thing you should note though: width and weight changes are only two of the many things that need to be accounted for, and the Benton boys (engineers as well, tellingly) certainly knew that.

As for figuring out any "algorithm" that was used to automate the scaling, that's a very worthwhile goal. The Cost article I didn't find very enlightening though. So I've collected the names of half a dozen "old-timers" who might know more, and eventually I have to track them down and see if they'll spill the beans.



raph's picture

Thanks for the response and links to the Adobe Garamond and trapping threads, Hrant. [do you monitor and respond to these boards 24 hours per day, perhaps?] We're kinda going off the original topic of the ATF legal rights, but I did want to briefly respond to a couple of scaling points.

I absolutely agree that ATF is the best historical model for scaling - I seriously doubt that any numerically inclined font designer has put more thought into the question before or since. However, while the scaling in ATF Garamond is quite pretty, I've pretty much settled on the Bodoni family, for a few reasons:

1. The Bodonis are present in the 1912, 1923, 1934, and 1941 books. Of these, the 1912 arguably has the highest reproduction quality. ATF didn't introduce their Garamond until after the 1912 book came out.

2. From what I can tell, the Bodonis use a single pattern plate for all sizes from 6pt to 72pt, while Garamond has two patterns: one for 6pt through 14pt, another for 18pt through 72pt (although, in the original 1923 book, 14pt is made from the display patterns). Thus, we see the machine scaling get much more of a workout with Bodoni than the fonts with multiple masters.

3. The shapes of Bodoni are easier to analyze because they're more geometric.

Actually, if I'm going to analyze the scaling of Garamond, I'll look at the originals, as we know that Garamond wasn't working around the inherent mathematical (as opposed to optical) scaling nature of the pantographic engraving process. I have a nice reproduction of Berner's 1592 specimen sheet, and there's a good range of scales of authentic Garamond types from Canon (roughly 45pts) to Garamond (roughly 9pts).

I absolutely agree that width and weight aren't everything, but for some of the ATF fonts (I've looked most carefully at the Bodonis, Century Oldstyle, and the News/Franklin Gothic family) those two parameters are the lion's share of the adjustments. Many fonts also have shrinking descenders in small sizes, but I don't really consider that part of optical scaling per se. Other than that, the main thing I see is that serifs get a mite shorter in the small sizes, after everything else is normalized away. A few letters have additional adjustments. For example, in the Bodoni Book lowercase 'k', the "less than sign" moves farther away from the vertical stem in smaller sizes, so as to preserve the visual distinction.

But, from what I've seen so far of the Benton's work, fully automatable techniques should capture 90% or so of the effect of optical scaling. Of course, people will always want to hand-tweak beyond that, but in my opinion even a mediocre interpretation of optical scaling is better than the current status quo.

hrant's picture

> do you monitor and respond to these boards 24 hours per day, perhaps?

I used to. But then I got back from a 2-week vacation and got inundated with Typophile emails, and decided that's it. Now I mostly skim (especially the logo crit stuff - for me it's: just the fonts, maam), but still try to post when it feels useful.

I think the 1923 specimen has great repro, but you're saying the '12 is even better? I could believe that. The Enschede 1908 specimen for example is visibly better than the -still masterful- 1978. What was the print run for the 1912? The '23's was 60K+, and maybe that put a dent in the quality.

But anyway a carefully printed one-off (with the help of a master printer) on high-quality paper is even better. The only problem with that though is you can't be 100% it's the real thing. For example the 7pt in my waterfall is clearly faulty. And the 14 is funny too - in fact Kent has noted that there have been two cuts of the 14: one texty, one displayy. And I see you've noticed this too! Good job.

The fact that the Garamond came later gives it yet another advantage, no? I mean, you're saying that the Bodoni had only one pattern plate (a very interesting observation), and to me that raises a warning flag, because -as you might have read elsewhere on Typophile- I believe there's a qualitative difference between text and display fonts. On the other hand, since Bodoni is so "constructed" maybe it only really needed one? But then that plays into my other point: because Garamond is more "organic" it pushes the limits of the automation more - so it yields more insight (both into the nature of optical scaling and the nature of the ATF automation) in the end.

The original Garamond:
1) The ATF Garamond is one of the many "fake" Garmonds. It's actually based on fonts made my Jean Jannon.
2) Yes, Claude Garamont was not as good as the Bentons in optical scaling. But that's a reason to avoid him, right?

Interesting here is a comparison of the "Garamond" of the Imprimerie Nationale versus the ATF's, in terms of optical scaling. A lot of it is comparable, except when you get to the really small sizes: at the very low end the ATF boys kicked the optical scaling way up, while the Frenchies tried to maintain more of the character of the design (which however I consider pretty hopeless at that size, and you lose too much readability from the small apparent size).

Here's some stuff I did a while back:

> I don't really consider [shrinking descenders] part of optical scaling per se.

No? Although some of the jumps in extender length are due to integer discretization*, increasing apparent size (through an increase in x-height) to me is a very big part of optical scaling. And there's something else: at the very smallest sizes all-caps setting becomes more important, so the baseline is brought [even] lower.


> even a mediocre interpretation of optical
> scaling is better than the current status quo.

Indeed. And I think we can get even higher than 90%.


raph's picture

Thanks for the additional followup.

I think the ATF Garamond and Bodoni are both worth studying, perhaps having slightly different lessons to teach. I still feel that the Bodoni is more useful for careful analysis of exactly what the Bentons did and how they did it. You can make the argument that the results of the Garamond are more aesthetically appealing, but that is not exactly the focus of my quest right now. It's obvious that it's possible to achieve optical scaling by drawing lots and lots of masters, but that's just too labor-intensive, even in the Bentons' day.

As for the going back to the originals, let me try to state it more clearly. The original Claude Garamond fonts carry absolutely no risk of having artifacts introduced by the pantographic scaling process, as each punch is cut by hand so as to please the eye. For example, a waterfall of ATF Garamonds has a discontinuity as you switch from one master to another.

One possible reason why the ATF scaling may be more extreme than Garamond's is that they appealed to much more of a mass market, while I assume that Garamond mostly concerned himself with fine printing.

Let me put some scans together to illustrate my points, and I'll start a new scaling thread - this definitely isn't a "Release" issue any more.

paul d hunt's picture

back to this question... so wassup with Kingsley? Does anyone know anything about them?

kegler's picture

As with other Type companies with military sidelines:

billtroop's picture

For the record, Mark Batty told me some years ago that ITC was still paying 'a little old man' royalties on several ATF trademarks.

Miss Tiffany's picture

This thread has been moved to "RELEASE".

Miss Tiffany's picture


Syndicate content Syndicate content