Fontlab 5.0.2 warning message.

Primary tabs

23 posts / 0 new
Last post
Jess Latham's picture
Offline
Joined: 23 Jan 2004 - 6:11pm
Fontlab 5.0.2 warning message.
0

I just moved the font I'm working on into fontlab 5 and after generating an .otf file I get this warning in the output box:
[WARNING] The feature file OS/2 overrides TypoAscender and TypoDescender do not sum to the font bbox size!
I did a google search for the warning message and came up with a few pages but nothing that explained what was happening in a way that I could understand. Can anyone explain this in a way that I can understand? Should I be concerned about the warning?

paul d hunt's picture
Offline
Joined: 5 May 2005 - 8:44pm
0

it means your vertical metrics are not set optimally. for some tips, see:
Vertical Metrics How-To

Jess Latham's picture
Offline
Joined: 23 Jan 2004 - 6:11pm
0

It seems like fontlab could just make all this happen without me having to even look at metrics, hee.

Jess Latham's picture
Offline
Joined: 23 Jan 2004 - 6:11pm
0

The fontlab manual actually explains it and it works now! Yay!

Miguel Sousa's picture
Offline
Joined: 18 May 2003 - 8:30pm
0

Jess, I'll have to disagree with you. I think people rely too much on FontLab's default options and algorithms that fill-in the settings that were not provided by the font developer. In fact, I think FL should have a 'Release mode' that would warn everything that didn't comply with the standards/specifications, or was considered 'suspect'. That would be the only way to make a close-to-rock-solid font file.

Jess Latham's picture
Offline
Joined: 23 Jan 2004 - 6:11pm
0

True, it's just that all those numbery things scare me. After I read the manual on pg 303 I feel a little more confident to change those numbers. It's a fear of messing things up, "don't touch anything or this thing could come crashing down!" hehe

Mark Simonson's picture
Offline
Joined: 3 Dec 2001 - 11:00am
0

One thing that I wish is that export options be saved with the font. As it is now, if you change the settings, the change is global.

Even better, some kind of import/export facility for the settings would be nice. (Perhaps this is possible with Python...?)

John Hudson's picture
Offline
Joined: 21 Dec 2002 - 11:00am
0

Mark, you can save option settings (at least, you can in the Windows version), not with the font VFB but as independent registry file. I have several of these, used for different kinds of font generation, and import the one I need before generating a font.

Chris Lozos's picture
Offline
Joined: 25 Feb 2004 - 11:00am
0

Jess,
I had the same problem with fonts started in 4.6 and moved to FL5. I actually the FL5 version makes more sense even though it is less forgiving than the old 4,6 If you look at the OS2 numbers, check to be sure that they jive with the actual font metrics--and your typo ascender/desender sums to 1000. You will probably see some pair of numbers not matching and prompting the error.

ChrisL

Mark Simonson's picture
Offline
Joined: 3 Dec 2001 - 11:00am
0

Mark, you can save option settings

Silly me. A little voice in the back of my head said that such a feature indeed exists and that I should do a reality check before posting, but did I listen to that little voice? Obviously not. Thanks for the reminder, John. It's right there in the corner of the Preferences dialog.

The shear number of features in FontLab is almost too much to keep track of sometimes, not that I'm complaining.

Er... you wouldn't happen to know if there is a little green diamond "Autocomplete Font" button buried somewhere?

Thomas Phinney's picture
Offline
Joined: 3 Sep 2002 - 11:00am
0

I'd also recommend running MS Font Validator and Adobe's CompareFamily (from the AFDKO) on the completed font.

Cheers,

T

Adam Twardoch's picture
Offline
Joined: 3 Dec 2002 - 7:36pm
0

> I’d also recommend running MS Font Validator
> and Adobe’s CompareFamily (from the AFDKO)
> on the completed font.

...and before shipping the font, running http://www.fontqa.com/

And if you're on a Mac, also running ftxvalidator on the final fonts from http://developer.apple.com/textfonts/Fonttools/

And it also won't hurt if you try to decompile your fonts using TTX from http://www.font.org/software/ttx/

:)

A.

Adam Twardoch's picture
Offline
Joined: 3 Dec 2002 - 7:36pm
0

...and, of course, do *NOT* blindly believe what all the different validators tell you. Take their output with a reasonable dosis of skepticism. Many validators raise a flag just to tell you "hey, there MIGHT be a problem here", and it's still you who needs to decide whether the problem is a real one or not.

A.

Chris Lozos's picture
Offline
Joined: 25 Feb 2004 - 11:00am
0

Thanks for the links and advice Thomas and Adam!

ChrisL

Göran Söderström's picture
Joined: 15 Feb 2006 - 2:53am
0

Sorry, to pick an old topic here, but I want to “get down” and learn all this — for real now.

These are my impression after reading different articles and asking people:

• Regarding the vertical metrics:
Knowledge that has to do with the TypoAscender/Descender/hhea.Ascender/Descender/OS/2.winAscender/Descender etc, is not really any “final bank” of knowledge.

Why? As an example, I read the "vertical metrics how-to" and it seems that a little bit of that and a little bit of this, actually works. And if not set optimal – it doesn’t mean that the font will be corrupt or will not work, am I right? The how-to article is updated a lot of times, changing it’s mind so to speak :)

Is it that these things are more of a very advanced expert discussion going on, similar to which sparkling water has the most optimal bubbles :)

QUESTION:

Will, or will not, an OpenType font work fine (for standard users – not extreme typohpiles) in both Windows and OSX even though these OS/2/hhea/win values do not exactly match? Maybe it’s OK if it differs a little bit from plattform to plattform in the total height when used, since the designer for sure anyway will set his optimal values.

But will it work? What could actually happen if these settings are not "optimal"? Will the font be unstabile? What problems will the user have?

The reason I ask is that I think it's rather strange that the leading program FontLab’s recommended settings (Calculate Values Automatically) actually _creates_ a problem like this, and also that it doesn’t seem as if there is a straight answer anywhere to be found.

Finally – take the font Zapfino with it’s long swashes going way above the ascender – if I understand correctly: The Bbox size in this font, is WAY to small for the glyphs. So Zapfino is in other words not an optimal font?

OR… are these values “just numbers” that have to be set correctly — and then we don’t think more about it? :)

(Phew…)

PS. Sorry for a long post, but I’m having problems understanding all this, especially the more I read.

John Hudson's picture
Offline
Joined: 21 Dec 2002 - 11:00am
0

But will it work? What could actually happen if these settings are not “optimal”? Will the font be unstabile? What problems will the user have?

Lets say I have a carefully formatted Word document (or an Excel spreadsheet, or a Powerpoint presentation...) that I have created on Windows and which uses default linespacing as determined by the font metrics. I send it to a colleague who will add a couple of graphics in the spaces I have left for him to do so, and then print the document for distribution. He receives the file, opens it in Word on the Mac, and because the font has incompatible OS/2 and hhea table metrics, the linespacing is completely different from my original. All the formatting is messed up. The spaces for the graphics are no longer the correct size. The page breaks are in the wrong places, etc.

Now a smart user will control the formatting in such a way that it will not be screwed up by an incompatibility in font metrics, but as you specified, we're talking about 'standard users'.

That's the compatibility issue.

The other issue is clipping. If the OS/2 Win values are too small, any part of a glyph that extends beyond those values will be clipped in many apps and on some printers.

Finally – take the font Zapfino with it’s long swashes going way above the ascender – if I understand correctly: The Bbox size in this font, is WAY to small for the glyphs. So Zapfino is in other words not an optimal font?

The Apple version of Zapfino is an unusual font in a number of ways, but those long extenders actually do fit within the vertical metrics. I can't remember the exact values, and Apple fiddled with the font after delivery (setting a low upm value to increase the relative size of the font).

Göran Söderström's picture
Joined: 15 Feb 2006 - 2:53am
0

Once again, John – you are very pedagogic – thanks for your reply.

You see, the hard part for me when reading manuals and such, is also the language- and terminology, as I am swedish.

I think I begin to understand however — or at least touch it on the surface :)

Karsten Luecke's picture
Offline
Joined: 6 Aug 2005 - 8:41am
0

As to possibly different line distances in different applications, or squashing, or clipping of glyphs -- sometimes it's not possible to avoid all of them. E.g., if your font has a few glyphs with very large ascenders, then you may accept that these get squashed or clipped in some applications, to make sure that default line distance will not be too much. In some way, it is you who has to set priorities. (In some applications, clipping behavior even seems to differs depending if the font is PS-OTF or TT-OTF ...)

not really any “final bank” of knowledge

This is true for vertical metrics, but also for font family naming and a couple of other issues. There are too many factors involved, like, will you make PS-OTFs or TT-OTFs, will your family follow Microsoft's example and have exactly four styles, or will it have more styles (or less, which can turn out to be as problematic), or do you want to use your PS-OTFs down to OS9 (even more restrictions)? Unfortunately, this contradicts the compatibility claim made for the OpenType font format.
That OpenType is still 'under development' (see the Vertical Metrics How-to: 'Update to the update' near the bottom) doesn't make things easier too. Be prepared to make updates regularly.
Never expect definitive answers, or a "final bank of knowledge". They don't exist.  :)

Karsten

Göran Söderström's picture
Joined: 15 Feb 2006 - 2:53am
0

I think there is time for a “Font Dictator” to enter the scene, and set the rules – once and for all! ;)

Like som sort of God, saying; “from now on we do like this. For both Mac and Windows, these are the standards” etc.
Then one could concentrate on the design instead of mumbling with technicalities.

Thomas Phinney's picture
Offline
Joined: 3 Sep 2002 - 11:00am
0

So, who would you want as dictator?

T

Göran Söderström's picture
Joined: 15 Feb 2006 - 2:53am
0

Well, it was just a metaphore for me thinking all this seems a little bit too complicated for my me – but hey, You seem like a nice guy ;-)

Chris Lozos's picture
Offline
Joined: 25 Feb 2004 - 11:00am
0

Absolute power corrupts absolutely.

ChrisL

Nick Shinn's picture
Offline
Joined: 8 Jul 2003 - 11:00am
0

So, who would you want as dictator?

David Berlow.