Typophile - Adobe InDesign vs. QuarkXPress - Comments

hrant — Fri, 14 May 2004 11:18:34 -0700

> I thought you wanted categoric proof

Nah, I'm not that kind of guy. Just "enough evidence" is generally plenty for me.

> .... fails your criteria?

I think it could be better, and that it messes up sometimes. But from what I've seen (previously and in this thread), it looks like it improves the overall spacing of most fonts. Maybe 90% is too much, but it still feels way above 75% to me.

Plus the good news is that:
1) You can somewhat predict when you shouldn't use it (based on the source of the font at hand).
2) For point sizes outside the original "intended range" it does much better than brute tracking.

hhp

Nick Shinn — Fri, 14 May 2004 11:10:07 -0700

>I'd be curious to find out which.

Take your pick. It does it for them all.

>Hey, weren't you the "it's a matter of taste" guy?

Originally, and I continue to have my preferences.

However, I thought you wanted categoric proof, according to your own criteria (from the Microsite) "By ensuring

hrant — Fri, 14 May 2004 10:58:39 -0700

> I checked out a few more good old serifed faces

I'd be curious to find out which.

> QED

Hey, weren't you the "it's a matter of taste" guy?

hhp

Nick Shinn — Fri, 14 May 2004 10:54:04 -0700

>"my theories"

Hey, you're the bouma guy.

>Two words (one of them a non-word) from one font is more than enough testing

One word is not comprehensive enough to make a solid generalisation, but if optical kerning can't get "and" in Times Roman right....

Anyway, I checked out a few more good old serifed faces, and optical kerning does the same thing in all of them for the "a" -- adding extra space after, when a vertical stem is next.

That is really dysfunctional, because all the faces were designed so that the tail serif and sidebearing on the "a" provides both the appropriate "breathing space" and "connectivity" with following letters.

The right sidebearing on trad serifed "a"s is less than for other characters with lower right serifs, because the top of the right side of the a is curved back, away from the following letter. this is why the serif curves up, to compensate, to make the letters appear more connected, yet without touching.

The right vertical stem of the "a" is designed to be a certain distance from following vertical stems; and the serif of the "a" is designed to be very close to serifs that follow, yet Optical Kerning looks at that minimum distance and won't allow it, thereby pushing the adjacent stems too far apart, creating uneven color.
QED.

hrant — Fri, 14 May 2004 10:06:34 -0700

> I meant work for this thread.

No, because you brought "my theories" into it.
As you should: everything is related.

> QED.

Right. Two words (one of them a non-word) from one font is more than enough testing.
And nevermind that the default spacing in that "and" is too tight for 9 point.

hhp

Nick Shinn — Fri, 14 May 2004 09:58:19 -0700

>I've been doing the work for 6 years!

I meant work for this thread. I'd like to see examples of your empirical testing, not just take your word for the conclusions.

>more rigorous, like by working with frequent words

OK, how about this. Again, it's 9 pt type, Times Roman, above unkerned, below Indy optical kerning.

Indy adds kerning between a and n, making for more uneven text color. QED.

hrant — Fri, 14 May 2004 09:15:09 -0700

I've been doing the work for 6 years! And not just in theory (derived not from my dreams, but empirical and anecdotal evidence), but in InDesign evaluation too : Where do you think I got the "bi-linear" thing?

> Let's see some visual proof, in one-on-one testing, of the superiority of Optical Kerning.

You're doing a fine job! :-)
Although you need to be more rigorous, like by working with frequent words.
http://www.themicrofoundry.com/s_quality.html
(Far right.)

hhp

Nick Shinn — Fri, 14 May 2004 08:38:35 -0700

Hrant, I'm doing all the work here!

My argument is made with visual examples and testing, rather than theories about immersive reading and the way optical kerning works.

Let's see some visual proof, in one-on-one testing, of the superiority of Optical Kerning.

I've made some refinements to refute your criticism. I've made this sample from 9 pt. type. Then I blurred it, to highlight the rhythm of dominant strokes, and the broad shapes of inter-character spaces.

By bunching up "lbow", optical kerning has worsened the spacing here. Perhaps it's an anomaly, but this was the first character sequence I chose, off the top of my head, because it has some difficult combinations that I often use in the type design process.

Times: Above, no kerning. Below, Indy Optical Kerning.

piticu — Thu, 13 May 2004 23:34:32 -0700

Christian Robertson — Thu, 13 May 2004 11:39:18 -0700

>

In my experience, InD optical is best used sparingly. Aside from the=20 performance issues, I agree with Nick that it more often mucks things=20 up when you start with a well spaced font. Even with a poorly spaced=20 font, InD can only help so much. I used to set preliminary text samples=20=

in a font, to get an idea about which characters needed to change=20 before I spaced it properly. I found that InD optical only *influences*=20=

the spacing, as opposed to replacing it as Hrant suggests. So if you=20 have a poorly spaced font to begin with (e.g. ScanFont's auto-spacing),=20=

InD can only go so far to make things better--though it can improve=20 things. I've found that it's more useful for titling since so many=20 fonts are spaced for 12pt: even titling fonts. The optical size=20 compensations are definitely cool.

>

William Berkson — Thu, 13 May 2004 11:17:06 -0700

I have used optical spacing with Poetica and, especially with alternate caps, it seems to be noticeably better than the original spacing.

hrant — Thu, 13 May 2004 11:11:41 -0700

> that is probably because I try to use only top-quality, well-spaced fonts.

Bingo.
But one day if you want to use the ravishing (but tongue-tied) Mrs Eaves, Indy's optical spacing will save the day.

hhp

kentlew — Thu, 13 May 2004 11:05:42 -0700

Nick -- It looks more like Indy has adjusted both sides of the 'o', more than moving the 'w'. I think Indy's solution in this specific triplet is marginally better. Basically the Times 'w' spacing needs a little help. I don't like how Indy has opened up 'ni', especially since it closed up 'lb'.

Hrant -- it's not so much that 'lb' is worse; *alb* is worse.

My opinion, based only on relatively few opportunities that I've used Indy, is that I prefer not to use Indy's optical kerning -- but that is probably because I try to use only top-quality, well-spaced fonts. The one exception to this is that I have applied it selectively to things like drop caps or other mixed font/size situations. This can indeed be useful.

-- K.

hrant — Thu, 13 May 2004 10:35:45 -0700

> "kerning preference" varies with individual taste

Only during deliberative reading; immersion precludes "taste".

When you try to evaluate something like this, you have to look "behind" it so to speak, to see what it would be seen as in the parafovea for example.

BTW, "rhythm" is in your imagination. :-) It's "pattern".

> taken from thin air

Don't be such a lawyer, dude.

> Just try it on a cursive script

I'm talking about text fonts.

> .... the kerning seems to be the same

Then there's something wrong with your test. I've done close measurements: Indy's optical spacing is "bi-linear": a steep line from 4 to 12 point, a shallow line from 12 to 72, and flat on either side.

BTW, about the tracking: messing with the tracking will skew the results, since Indy's optical tracking doesn't (and if you think about it maybe shouldn't) take that into account (look at the behavior of "boundary conditions" like the right of the "r"). So it's best to leave the tracking zero, certainly for testing.

hhp

Nick Shinn — Thu, 13 May 2004 10:20:00 -0700

>but every other pair is better.

Anybody else have opinions?

Anyway, I'm not talking about pairs, but about the overall color of the type.

And even if my opinion is in the minority, it is still an opinion, and that suggests that "kerning preference" varies with individual taste, and that one-size-fits-all standardization is not a good idea.

You figure of Optical Kerning improving 95% of typefaces is taken from thin air, and is completely absurd. (Just try it on a cursive script ;-)

I tested it on a handful of typefaces to start with, and found that it is just as likely to make fonts worse!

>optical spacing is sensitive to point size

That sample was at 96pt, but I tried it at 9.6, and the kerning seems to be the same, although the tracking is different. (Rather like the marvelous Quark Tracking Utility, but not user-modifiable.)

Adobe InDesign vs. QuarkXPress

eekamouse — Tue, 4 May 2004 03:28:14 -0700

My company is sending myself and other Mac users on a two day course next month to learn about Adobe InDesign. Specifically, whether it is worth changing the whole (advertising) company from Quark 4 to InDesign as opposed to upgrading Quark (the more expensive option).

Although I am a bit wary about changing over, I’ll be going along with a (relatively) open mind. But I would like to know what thoughts do you people have about the differences between Quark and InDesign? Things to look out for? Basic boxes or tricky typographic twiddling? From an overall design perspective to a more particular typographic viewpoint - which is best?

Apologies if this has been covered before, but it would be really useful to have a few opinions and perhaps things to look out for from people who know their stuff.

Thanks