"Why will Johnny no longer want to read?"

dberlow's picture

This is simple. Tell me, if you please, which you'd like to read, given no ither choice... "forever", the font on the right, or the font on the left.

If you don't know how to see this at "100%" please, ask. If you think that the artifacts of conversion to a bitmap file have unevenly effected these two specimens, please don't. Both specimens were made from the same outline font, both specimens are hinted for and displayed with CT, and are shown at the same size and as much as possible, in the same composition. In addition, the same effective difference in typographic quality can be shown for any extant CT font published so far.

Then the question of course, is: how could any trustworthy readability research be based on a font like the one on the right? Unfortunately, this is no longer my problem, or Kevin's problem, or the researchers problem as much as it had become, quite quickly and massively, a users problem, as evidenced by the fact that I can barely hear you all over the din.


L16OR16W.JPG228.46 KB
dezcom's picture

Something, shall we say, odd about the metrics on the right sample. I am particularly intrigued by the odd pattern in the string of "i"s which looks like pairings instead of a single uniform line of singles as in the left sample? It feels like I am reading while riding a horse--galumph, galumph, galumph,.


Jackie Frant's picture

Immediate upon seeing the two samples side by side - my eyes preferred the right hand side - for its larger x-height.

After separating them and trying to read them - the left hand sample seemed very choppy - and the capital P was bothersome...

However, not to give the right hand side that much credit, because who ever typeset it, did not know what they were doing. The widow stew isn't happy, and on a Flush Left, Rag Right - there was no reason for wonderful to become a hyphenated word. And in the second paragraph - two hyphens in a row - OMG...

dezcom's picture

Oh, and Johnny won't read it because it isn't about babes, beer, and fast cars.


dezcom's picture

I should have mentioned that I prefer the text on the left for a less jarring feeling. The weight of the text at right is fine but straight to straight sidebearings are rudely tight or just lumpy looking.


William Berkson's picture

The left is way better. The letter spacing on the right is, as Chris has noted, very messed up, and makes it less readable.

What are these?

mike_duggan's picture

David, can you let us know what you did to the image on the right? I doubt very much that we ever displayed such weird font spacing in any of our studies. If you take a screen shot and save as a png, you also dont get all of the backround dithering, so it might be helpful to just that, nothing worse than looking at an approximation of the real thing.


Nick Shinn's picture

MUCH prefer the one on the right.
Not a big fan of goosestepping men of lead.
And it heartens me to see the young business dudes with their just-got-out-of-bed hairstyles, two-day stubble, and shirts that aren't tucked in. Surely they will reject antiseptic mechanistic typography -- and have someone else read it 3% faster.

But I don't think it's just the irregular spacing that makes the one on the right better for me, it is also a bolder image, with more substantial "hairlines", at least on my present monitor.



Sharon Van Lieu's picture

At first, I preferred the left. When I took the time to read each side, I preferred the right.


Paul Cutler's picture

Like pearlstreet - at first glance as a graphic entity, not reading material - left - but after reading both - right.

And I find that pretty darn interesting…


pattyfab's picture

Agree with Paul and Sharon completely. Left looks more "pleasing" to the eye, but the right reads better. Isn't there a middle ground though?

Goran Soderstrom's picture

The right one I feel is more readable for screen reading – due to more black and sharp appearence. I dont know exactly what you did, but it looks as if they are rendered with different anti-aliasing "systems" when rasterising the fonts in some way.
Perhaps the different choises that are available in Photoshop or Imageready? Sharp, smooth etc. Or maybe one is from a PDF and one is a screen shot or something like that?

dberlow's picture


domdib's picture

Text on the right, although I agree about something being slightly off with the metrics. In the left sample, the bowl of the lc a is breaking up, there's also break up in the uc G, the top of the stem of the lc t is virtually invisible, the dots on the lc i & j are tiny and the lower bowl(?) of the lc g begins exactly on the baseline, while on the right it tucks above it. All of these interfere with legibility.

eliason's picture

What dezcom and William Berkson said about the spacing of straights. E.g. the "li" in "Pliny" and "flipped." "flipped" looks like a bomb detonated halfway through the word.

At first glance I assumed the right would be more readable, I guess because of its bolder quality, and bigger x-height? But left was considerably more comfortable when I actually read.

Miss Tiffany's picture

I'm on a crappy old CRT. I prefer the one on the left because it doesn't vibrate as much, but I prefer the one on the right because of the apparent weight. If I had to choose I'd read the one on the left. The one on the right seems to have odd spots of blackness where letters meet or where the stroke is thicker. I find that to be distracting.

Linda Cunningham's picture

I'm way older than my CRT, and the left one is much more readable -- the right one started to give me a migraine....

Nick Shinn's picture

This is all terribly anecdotal and unscientific.
But perhaps that's the point, that people outside of a laboratory test aren't compelled to read what's best for them. And in a real life situation, people won't even bother reading something with superb "readability", if a glance leaves them cold. So good typography has to be both showhorse and workhorse.

Bruce's picture

David, I find myself stuck between two poles. L feels very much in the mold of Jenson's 1470 pages, serene, calm, and even in color. But the letters feel just a squeak too letterspaced and open. R feels very bunchy, blobby, hiccupy. I do not feel as drawn to it for its darkness as others have been. While there have been some comments about individual letter structures and side-bearings -- things that might concern us technically -- I am concentrating on the experience of the reader, who (if I understand this correctly) reads in gestalt clumps and does not perceive individual letters.

I think I end up reading R more quickly, taking in the gestalts with more sureness, but it is fatiguing underneath because of the clashes of light and dark, the bunching of things.

L feels a lot calmer, but sometimes in the sameness of color (goose-stepping as Nick said above . . . or I thought of a picket fence) I cannot find my way as quickly in processing the words; I see the individual letters one after another and that reduces my ability to take them aboard in clumps.

What if I said I would not like either one as a "forever" companion, but something in between?

Also, are these meant to be read only on screen? (which I have done at 100%.)

If I had to pick one to read forever I would pick L.

Was this helpful? Hope so.

ben_archer's picture


I'm reading these at 100% on a G5 iMac LCD screen @ 1440x900. I find it easier to read the sample on the left.

What Chris says about the straight-to-straight sidebearings being overly tight is true. The line length of the right hand sample appears slightly longer than the left hand sample, and the right edge rag is softer, so I guess that overall there are more characters per line in the right hand sample. I notice that the leading values are different between the two samples – with the right hand one being leaded more generously – although that doesn't improve its readability IMO.

Are we actually looking at two different fonts here? or the same font with different spacing applied?

Goran Soderstrom's picture

No I got it – one is Mac and one is PC!


dberlow's picture

Thank you all. I lied as in "This is simple."© ;)

The most surprising thing to me: blur never came up, in word or root. Whew!

The x-hts are identical. Both samples use the same rendering, are from the same page in the same application on the same OS at the same time... only the font software has changed...and then, only TT instructions within the glyph data table.

If you like the sample on the right, you are all set. Many of the CT rendered fonts in Windows Vista will appeal to you as will type on the Mac and TV. Unless you return to print, or upgrade to a higher resolution LCD screen (or sample one), you will not know the difference.

IF CRT: This is not a good test for CRT-users, apparently. Sorry, when are your birthdays and how wide is your CRT's desk space?

If you like the one on the left, you might soon be able to get up to the same speed and comprehension as you read in print.

The primary difference between the two samples is that one is prioritized for the rhythm and reading on screen, and the other to match the widths when printed.

If you are not sure, I will work to improve both sides, and I hope you read again without thinking any more about this. Now watch the roving center period...you are getting sleeeeeppy...your eyelids are heavy...you hear a yawn...

Insomniatic Typophiles: An "abundance" of pixels are used in rendering digital outline fonts on the Mac (and in Cool type) because those mechanisms don't "really" use hints and anti-aliasing simultaneously for screen font rendering. Both designs in this study were made just light enough to avoid being too dark on the Mac. In Vista, (where hints and anti-aliasing are encouraged to work together), rendering is very light. So, some of the hairlines of the Mac version "burnt out" in Vista rendering. When I first saw this I thought "I scanned some overused metal by accident?", and I liked reading it very much. Though I still do, so much for personal opinion.

Thank you all again, I've learned so much.


William Berkson's picture

I looked at the samples on windows xp on a flat screen. On my screen the one on the left is a bit light or loosely spaced, and the darker weight of the right is better. But as I said the spacing on the right is seriously irregular, to the point of hurting readability.

hrant's picture

David, before I chime in, could you explain how you're accounting for the gamma difference between Windows and MacOS? Because that's pivotal, and frankly it's pretty shocking that nobody has brought that up yet.


mili's picture

Well, I'm on CRT/Mac, eyes about 60 cm from the 19 inch Panasonic Panaflat screen (located below eye level), born in the mid 1960's, wear glasses for computer, driving and tv, otherwise not at all. I might prefer the one on the right, even though the first impression was that the left one was better. Ideally something inbetween?

Bruce's picture

IF CRT: This is not a good test for CRT-users, apparently. Sorry, when are your birthdays and how wide is your CRT’s desk space?

Aha. Yes, I am using a 19" ElectronBlue CRT, 1024 x 768. And I was born in 19 and 49.

charles ellertson's picture

I'm using a Sony Trintron CRT with a 14-inch "deskspace" (horizontal, black-to-black on the screen). It is calibrated, I prepare a lot of B&W images for halftone printing. Oh. I was born in 1945.

If the left & right samples are my only two reading choices, I'll give up reading, except for printed books, probably soon only to be found in dark basements, hidden away from the police.

dezcom's picture

Farenheit 451


PS: born in January of 1944; wear trifocals

Sharon Van Lieu's picture

I have a couple of NEC diamondtron CRT's 19" and 17". I wear computer glasses (trifocal/bifocal).


Paul Cutler's picture

Apple 23" Cinema Display - no glasses - yet…


Beat Stamm's picture

So sad David spilled the beans already---this was sort of entertaining while it lasted. The goose-stepping vs galloping eyes was the hint that gave it away for me from the beginning. But now I'm left wondering: don't advance widths matter? Or did the sacred spacing bovines wander back into the 20th c?


dberlow's picture

"how you’re accounting for the gamma difference between Windows and MacOS?"

Subtle variations (hinted or, contoured), are all that can care of any of the differences between display technologies, arent they?

"don’t advance widths matter?"

They are almost all that matters, besides the letterform, are they not? I think those two things are dead even in the importance of being readable (at low resolutions). :) Don't you?

Syndicate content Syndicate content