Converting Fonts to Outlines and Weight Gain

Diner's picture

This has always perplexed me, in Illustrator when a font is converted to outlines, it seems to get a significant weight gain.

Is this perception primarily just the on screen display of the face or does the font gain in the output as well?

I've done a few tests and found that this is more apparent at font sizes below 10. . .

Is there any way to counteract the affects of the font to outline process?

Stuart :D

Joe Pemberton's picture

I think it's all an optical Illustrator illusion but
I'm willing to be corrected.

sevenfingers's picture

It does gain weight in output also (even though it's more apparent onscreen). It's really strange, but I've noticed it's more significant on laser printouts than "real" printjobs. Anyway, I also want to know...

beejay's picture

I was in Stuart's camp because I'd see this effect with my own eyes and always wonder. But Joe might be right about the optical illusion.

I did a quick 'color stack' test at 8 pt. and 800 pt. with Helvetica UC I and found the differences to be imperceptible at either size, but I didn't try printing.


When you zoom in and out, Illustrator redraws quickly and that's when I've noticed the weight difference.

An another subject, when you stroke with dashed lines, Illustrator doesn't let you create dashed outlines. Has anyone figured out a way to create dashed outlines?

bj

hoefler's picture

Font outlines are rasterized by ATM, which uses a different algorithm for turning on pixels than the one used by Illustrator. Put differently, fonts are hinted but other outlines aren't. The effect is visible even at 1,200 dpi on paper, and more so on screen.

I think the effect has been mollified in Illustrator 10, where everything seems to be anti-aliased, but I haven't used the application enough to be sure.

hrant's picture

From my understanding:
The algorithms for rendering fonts versus outlines are different, with respect to deciding if a border dot should be turned on or off. This translates to an average of half a dot of difference on the edge. On the screen, this is highly visible (irrespective of the additional difference in rending method - b&w versus gray). On a laser printer, it's very visible at text sizes, and even perceptible at sizes upto and beyond a single-fixation of the fovea (at normal reading distance). On lino, it's generally perceptible at just below normal reading distance, which isn't so bad.

On lino, the only real problem is when you're making specimens of a text face for close scrutiny, and you don't want to give the service bureau the fonts (or they don't want to bother with them). An 8pt text block for example will be visibly darker. You can reduce this by going to the maximum resolution (in case the resultant lower lpi doesn't matter), which can be above 3000 dpi. On the other hand, since almost any reproduction method will introduce *some* gain, you could view the result as being closer to the eventual output! But of course that's not quantitively controllable.

hhp

Syndicate content Syndicate content