PANOSE Your Fonts

[Dave Crosby]

PANOSE your fonts

Now that you've built your new font, Why not give it some extra luster?

Do you ever get frustrated searching through your collection for just the right font?
Do you need/want a sorting or classification scheme for your font collection?
There have been many machinations proposed. It is easy to end up with too few, or too many font buckets in schemes, so
instead of searching through 10,000 fonts, you end up searching through 20,000 buckets.

Surprise! There is a sorting system already built into each of your TrueType (TTF) Fonts and Open Type (OTF) Fonts!
It is named PANOSE, and it is used by most commercial font foundrys.

Part A
Q. What is Panose?

A. A Panose Number consists of a ten digit number, each digit containing a value between 0 and 15. I'm using hexadecimal numbers (0-9, A-F) inserted into the OS/2 Table of the Rich Font Description (RFD) incorporated into each True Type font. It was invented to speed up printers by minimizing the number of fonts required in the printer memory. For more information see the NOTES at the end of this tutorial.

Examples:
[attachment=1]PanExamples.jpg[/attachment]
Topology????

Topology -Math. The study of those properties of geometric forms that remain invariant under certain transformations, as bending or stretching.

As far as I know, ALL Non Latin fonts have yet to be classified, so they should be placed in the “No Fit” category of 1111111111.

However, most fonts with other PANOSE numbers contain Non Latin characters in their Supplementary Multilingual planes. See: viewtopic.php?t=386

Q. Why should I care?

A. 1. Many of your applications use Panose to replace damaged or absent fonts in documents. If you don’t reset the numbers when you build a new font, they remain set at 0000000000, Which means the font contains no Panose information, so it can NOT be used to replace any other missing or damaged font.
Usually Arial, Courier, or Times New Roman will be selected instead of your font.

2. Most of your collected fonts will remain jumbled at the bottom of the pile when you do a Panose or Similar search for them.

With FC you can embed Panose Numbers to make them available to Explorer and MainType searches.

3. It gives that special Professional Touch to your fonts.

In their Introduction, Monotype states:

Hewlett-Packard does not restrict the use of PANOSE Classification Numbers in typeface products developed by third parties.

So if YOU want to use the Panose spaces to set up your own system, do it!

FONT COLLECTORS: PANOSE can be used to identify "holes" (missing font styles) and duplicates or knock-offs!

To see Panose in action: if you are using Windows XP, Select My computer (or press Start), Control Panel, Fonts.
You will be presented with four choices for looking at your font files. We are now interested in choice 3, Similar. Yeah, the icon with AB on it. Select Similar then pick a font from the box “List fonts by similarity to:” by pressing the down arrow. Look at the results. Select another font that is not similar to the one you were just examining. Interesting, huh!

The PANOSE numbers embedded in the fonts are used to make these comparisons.



Do you think Common Bullets looks anything like Arial? I think Common Bullets should have gone into section 5 instead of section 2.
Obviously “similar” is in the eye of the “Panoser.”

Most of my favorite fonts are always at the bottom of the list with the notation “No PANOSE information available.”
Hmmm, Why is that? Could it be that the font creators never bothered to insert the Panose Numbers?
Conclusion: If you get your PANOSE straight, your fonts will join all the other good little fonts. What a concept!

An excellent way to use the Panose Numbers to sort/compare/select your fonts is High-Logic's very nice program MainType.
Go to http://www.high-logic.com/download.html for a free trial.



MainType gives you your choice of 4 pre-selected Information layouts AND the option of moving things around any way you want and saving an additional two options.
I selected the Panose setting under Columns and moved it to the top of the list. Also, under Information, I opened the Classification subheading for this presentation.
Note that Ancestry SF was not given a Panose number. It should at the very least begin with a 3.
Also note that MainType allows many other ways to search for a particular font! Just click on ANY column head to resort by that information.

MainType allows you to easily move fonts into any group folder/file system you chose with or without Panose.

Panose doesn't work as a truly great classification system (after all, it is NOT in the order I would have chosen AND you would want something different than what I want) because that isn’t precisely what it was created to do.
Perfection is a matter of opinion!

It is hard work to perfectly encode Panose. Then, even when we go to the trouble, it doesn't give us the differentiation we need.
Most font collectors choose buckets labeled by themes such as Engraved, Stencils, Medieval, kid hand; or pictorial themes such as African Art, Valentines, Christmas, Flowers, et al. They prefer to put all their text fonts into just one or two buckets, perhaps dark or light, one bucket for Serif, another for Sans Serif is usually enough. They usually don't care about obtuse cove, square cove or obtuse square cove.

But then again, Panose is there, and it works quite well, so why not use it?

EDIT 07 Jan 2010:
How I use Panose: I created a Font Folder with a sub folder labeled Panose and sub-sub folders for the first two digits and just visually tossed in the fonts I really like. Feel free to copy and print the following image.

PanGlance8x11.jpg (46.97 KiB) Viewed 142000 times

A more complete single page version is at:
postedfiles/Panose.pdf
Boring Fonts are in alphabetical folders.
Perhaps some day I'll take time to correct the Panose numbers, but for now I can quickly and easily find the font I want for most projects.

Q. How do I “Panose?”

A. The easy way is to use FontCreator:

1. Open a font similar to yours that contains Panose information.

Select Format, Settings, Classification, and jot down the information you find on the Value Line.



NOTE: "Font-family" at the bottom of the window is IBM's Classification scheme.

2. Open your font, Format, Settings, Classification, and enter the information where your font is the same as the other.
You have two choices:
1. Type the numbers in directly on the Value Line in Hex or decimal, or
2. Select from the list below the Value Line and let Erwin do the work.

Remember, if you want to make up your own system, just type in whatever you want on the Value Line.

3. You may wish to check the first three digits with the information in Part B, below.

4. If you wish to do more, go to the appropriate Panose web page To figure out whatever adjustments you need to make to your font Panose Numbers.


Part B
A brief introduction to Assigning PANOSE Numbers

Caution on measurements: When measuring a design with a highly rounded or bowed inside stem, be certain to calculate the correct theoretical edge for the location of the stem edge. Curved stems can alter the measurements for classification significantly enough to alter the resulting category. A face such as Optima can classify quite differently if the W Stem is incorrectly measured at the narrowest or widest portion of the stem.

Beyond the first 3 digits, it can get complicated/confusing. But then, 3 correct digits is MUCH better than 0000000000, and Erwin made it easy for you to GUESS at the rest if you have to, and do quite well.


The first digit, Family Kind, is the Most Important Digit!

At present, only 6 (0-5) digits are used in the space for the first digit:

0 = Any (No PANOSE information!) The mapper will replace it with any Latin Font. Do NOT USE!

1 = No assigned Font is similar. The font has a valid PANOSE number, but the mapper will do name matching only, not font substitution. Use only as a last resort.
For more information on the first digit 1 see:
Note: http://www.panose.com/printer/pan1.asp has been redirected to:
http://www.panose.com/ProductsServices/pan2.aspx

From here on, there is a two step process to decide which group a font belongs in:

Step One.
Answer the following three questions.

1. Does the font belong to a family that includes italic versions?

2. Are the characters in the font made up of standard topologies constructed of standard parts?


3. Is some portion of the font suitable for composing a paragraph of text?


Step Two.
If the answer is yes to all three, then the font belongs in Latin Text - Section 2
(where Serif Style is most important).

If there are any no's, look at the Second PANOSE Digit of the
Handwritten (Script)-Section 3 (where Tool Kind is most important) and
Decorative -Section 4 (where Class - Topology and Decoration - are most important)
families to see if there is a better fit to the question.

IF the answer to question #3 is NO, the font belongs in Symbol -Section 5.

(Remember Common Bullets?)

Edit Note 21 Jan 2008
Since writing this tutorial almost three years ago I've found many people have a problem with this first step.

In plain words, you only need to consider 4 first Pan Digits:

• Pan 2 = alphanumeric fonts consisting of more or less standard serif and sans serif shapes.
  Usually a LOT of design work has gone into the mechanical angles, shapes, and spacing for recognition and legibility considerations.
  
  Pan 3 = alphanumeric fonts YOU think would have been made with specific tools (charcoal, paint brush, felt tip, engraving tool, etc) out in the "Real World" and the Designer Decision is final.
  These fonts are looser, free flowing and less exacting than 2's. All Hand Scripts belong here unless they are "overly ornate," whatever that means.
  
  Pan 4 = either Pan 2 or Pan 3 fancied up in some way.
  
  Pan 5 = those fonts that are not alphanumeric in character.

However, many "pictorial fonts" have images placed in alphanumeric edit windows and could be classified as Pan 4 as they were not classified as Symbol Fonts.
(rhetorical question) Does that help?

Also, some Web Locations have changed and are noted in RED.

2 = Latin Text. Most of my fonts fall into either this group or Latin Hand Written. Other collectors may have very different interests.

The succeeding digits convey:
2. Serif Style, 3. Weight, 4. Proportion, 5. Contrast, 6. Stroke Variation, 7. Arm Style, 8. Letterform, 9. Midline, 10. X-height


Latin Text: Second Digit (2.2) = Serif Style. “The most sophisticated digit in the PANOSE classification system”
includes both Serif and Sans Serif typefaces.

Dump your font into the hopper, turn the crank and see what slot your font falls through.
"Transitional Serifs" Flared and Rounded belong in either group depending on base to stem ratios.



Anything fancier than this belongs in Family 3 Script or Family 4 Decorative.


20 = Any (Don’t use.)
21 = No Fit (Last Resort.)
22 = Cove (default for Serif) Serif joins stem with a curve.
23 = Obtuse Cove Serif attaches at an obtuse angle
24 = Square Cove
25 = Obtuse Square Cove
26 = Square AKA Slab or Egyptian
27 = Thin AKA Hairline
28 = Oval Previously "Bone"
29 = Exaggerated
2A = Triangle AKA Wedge
2B = Normal Sans (default for Sans Serif) Bottom of "A" stem parallel to base line
2C = Obtuse Sans BOTH Obtuse and Acute are included.
2D = Perpendicular Sans Bottom of "A" stem strikes base line at angle greater than 15 Degrees.
2E = Flared Stem widens at base. Narrow = Sans, Wide = Serif
2F = Rounded Stem is rounded at base. Narrow = Sans, Wide = Serif


Latin Text Third Digit (2.3) = Weight.


png Courtesy Bhikkhu Pesala

FYI: Lucky for Panose Sections 2.3, 3.3, and 4.3, the calculations for weight are among the easiest to compute.

There are specific letters used for measuring weight:
CapH is the Height of the capital H.
WStemE is width of the stem of the capital E.
WeightRatio = CapH / WStem(E)

Thin Stems will return large Rat (Ratio) numbers.
Thick stems will return small Rat numbers.

Many designers appear to just throw a dart. Perhaps that is good enough.

Code: Select all

2_0 = Any (Don't use.)
2_1 = No Fit (Don't Use)
2_2-Very Light....................WeightRat ≥ 35 (35 or greater)
2-3-Light....................18 ≤ WeightRat < 35 (18 or greater, less than 35)
2_4-Thin.....................10 ≤ WeightRat < 18
2_5-Book....................7.5 ≤ WeightRat < 10
2_6-Medium..................5.5 ≤ WeightRat < 7.5
2_7-Demi....................4.5 ≤ WeightRat < 5.5
2_8-Bold....................3.5 ≤ WeightRat < 4.5
2_9-Heavy...................2.5 ≤ WeightRat < 3.5
2_A-Black...................2.0 ≤ WeightRat < 2.5
2_B-Extra Black...................WeightRat < 2

For more information on this family see:
NOTE: http://www.panose.com/printer/pan2.asp has been moved to:
http://www.monotypeimaging.com/Products ... /pan2.aspx


3 = Latin Hand Written (AKA Scripts!). Usually the letters appear to be joined (flowing), but the group includes all "handwritten" fonts and non-flowing calligraphy.


The succeeding digits convey:
2. Tool Kind, 3. Weight, 4. Spacing, 5. Aspect Ratio, 6. Contrast, 7. Topology
8. Form, 9. Finials, 10. X-ascent



Latin Hand Written (Script) Second Digit (3.2) = Tool kind
30 = Any (Don’t use.)
31 = No Fit (Last Resort)
32 = Flat Nib
33 = Pressure point
34 = Engraved
35 = Ball (Round Cap)
36 = Brush
37 = Rough
38 = Felt Pen/Brush Tip
39 = Wild Brush - Drips a lot

Latin Hand Written. Third Digit (3.3) = Weight. Same as calculating 2.3.
3_0-Any (Don’t use.)
3_1-No Fit (Last Resort)
Etc.

For more information on this family see:
NOTE: http://www.panose.com/printer/pan3.asp has been moved to:
http://www.monotypeimaging.com/Products ... /pan3.aspx


4 = Latin Decorative. Fancy Fonts.

The succeeding digits convey:
2. Class, 3. Weight, 4. Aspect, 5. Contrast, 6. Serif Variant, 7. Treatment,
8. Lining, 9. Topology, 10. Range of Characters

Latin Decorative.Second Digit (4.2) = Class. Classify faces by using the lowest reasonable digit.




40 = Any (Don’t use.)

41 = No Fit (Last Resort)
42 = Derivative Decorative fonts built on standard text forms
43 = Non-standard Topology Unusual forms for entire letters but still uses standard stems
44 = Non-standard Elements Usual forms but unusual treatments of parts of them, such as serifs or ascenders.
45 = Non-standard Aspect Usual letter forms but unusual proportions such as very high or very low waists.
46 = Initials Typefaces with majuscule (uppercase) characters only.
47 = Cartoon Each entire letter is made up of a single picture that forms the outline of the character.
48 = Picture Stems Each stem is made up of a picture or pictures and the letters made from groups of these elements.
49 = Ornamented Typefaces have additional flourishes and details added to the character.
4A = Text and Background faces have the characters displayed as the absence of pattern on a patterned background
4B = Collage faces have the characters made up of repeating nonstandard elements
4C = Montage faces have the characters made up of non-repeating nonstandard elements.

Latin Decorative. Third Digit (4.3) = Weight. Same as calculating 2.3.
4_0-Any (Don’t use.)
4_1-No Fit (Last Resort)
Etc.

For more information on this family see:
NOTE: http://www.panose.com/printer/pan4.asp has been moved to:
http://www.monotypeimaging.com/Products ... /pan4.aspx


5 = Latin Symbol. Pictorial. Borders, Dingbats, Drawings, Silhouettes.

The succeeding digits convey:

2. Kind, 3. Weight, 4. Spacing, 5. Aspect Ratio & Contrast, 6. Aspect Ratio of Character 94, 7. Aspect Ratio of Character 119,
8. Aspect Ratio of Character 157, 9. Aspect Ratio of Character 163, 10. Aspect Ratio of Character 211

5 = Latin Symbol. Second Digit (5.2) = Kind.
50 = Any (Don’t use.)
51 = No Fit (Last Resort)
52 = Montages A mixture. No single type of symbol is more that 50% of the total set.
53 = Pictures Religious symbols, hands, feet, hardware, flowers, buildings, clocks etc.

54 = Shapes Abstract shapes like arrows, boxes, squares, dots, stars, etc. (Common Bullets?)
55 = Scientific Math, Greek letters, differential segments, etc.
56 = Music Notes, clefs, sharps, flats, trills, etc.
57 = Expert Specialized ligatures, groups of small caps, etc
58 = Patterns Line fill symbols, textures, etc.
59 = Boarders Simple or fancy corners and boarder sections.
5A = Icons Block illustrations and symbolic shapes.
5B = Logos Copyrighted logos and registered artwork
5C = Industry specific Engineering, law, medicine, Road Signs, etc.

5 = Latin Symbol. Third Digit (5.3) = Weight. This one IS different!

5_0 (Don’t use.)
5_1 = No Fit The Weight digit is required by the PANOSE engine, but is not meaningful for symbol faces,
so the third digit is ALWAYS set to 1.

For more information on this family see:
NOTE: http://www.panose.com/printer/pan5.asp has been moved to:
http://www.monotypeimaging.com/Products ... /pan5.aspx

In Conclusion, You will find many of your collected fonts have not been "Panosed," or have been given the wrong Panose number. Here is your chance to correct them.
______________________________________________________________
Notes:
This tutorial would have been impossible without the assistance of Dick Pape, Erwin Denissen & Bhikkhu Pesala.

Major Sources of Information:
http://www.myfonts.com/person/bauermeister/benjamin/
http://www.panose.com/printer/pan1.asp through

http://www.panose.com/printer/pan6.asp
http://www.w3.org/Printing/stevahn.html
http://www.w3.org/Fonts/Panose/pan2.html
http://msdn.microsoft.com/library/defau ... t_48aa.asp
http://style.cleverchimp.com/corestyle/ ... -font.html

A Manual of Comparative Typography: The Panose System (Paperback)
by Benjamin Bauermeister
* Paperback: 257 pages
* Publisher: Van Nostrand Reinhold (October 1987)
* Language: English
* ISBN: 0442211872

It was developed by Benjamin Bauermeister, who was born in St. Louis, Missouri, in 1960. In 1990 he was the co-founder with Clyde McQueen of ElseWare Corporation in Seattle, where he first revealed his PANOSE1 Typeface Matching System which began as a 7 digit number. Each succeeding digit breaks the font collection down into ever smaller groups.
Hewlett Packard Co. Soon purchased Elseware Co. And expanded PANOSE to ten digits, each containing a value varying from 0-15 which works great in Hexadecimal. HP created a PANOSE ENGINE that compressed font information into 2 KB packets and incorporated the Panose Numbers into their Agfa Monotype typefaces to identify which packet should be used with which font. Then they designed their printers so that instead of using an entire font, they just sent the number. The printer memory did the math and reproduced a simulation of the font. In other words, the PANOSE numbers told the printers how to draw the typeface.
Some improvements were made and Panose1 became Panose2.

The Panose Numbers have even been used to generate NEW fonts.

HP began reaching out for Partners to make PANOSE an industry wide standard.
Panose Partners have included: AGFA (MonoType); Hewlett Packard Co. (FontSmart); Adobe Corp. (PageMaker); Bitstream Inc.; Caere Corp.; Corel Corp. (CorelDRAW); Lotus Development Corp.; No Hands Inc. (Common Ground); and Microsoft Corp (Word Pro).

We suspect the word PANOSE is an acronym for the full name of the process, but then again, perhaps it is just derived from:

pannose <botany> Similar in texture or appearance to felt or woolen cloth.
Etymology: 19c: from Latin pannosus, from pannus cloth.

Perhaps the word was chosen to imply that Panose Numbers describe the “full cloth” of typeface possibilities.

So far as I know, there is presently no PANOSE 3 under construction, but .. . “Who knows what evil lurks in the hearts of ...”

Section 6 states: The tolerances of the weight classification have been determined by testing a variety of fonts. While this has provided reasonable averages for the ranges of weights, these will not always directly correspond with a font’s external name. It is not uncommon to have a font that contains the word “Bold” in the name that actually classifies as 7-Demibold, etc.

In addition, certain families that have a surplus of font weights may not progress smoothly through the differing classification options. It is, however, rare that two members within the same family will have two weights that exist in the same classification category.

Caution on measurements: When measuring a design with a highly rounded or bowed inside stem, be certain to calculate the correct theoretical edge for the location of the stem edge. Curved stems can alter the measurements for classification significantly enough to alter the resulting category. A face such as Optima can classify quite differently if the W Stem is incorrectly measured at the narrowest or widest portion of the stem.

[Bhikkhu Pesala]

The tutorial is now much improved. Panosing fonts can be daunting, but to add the first few fields at least will certainly make one's fonts much easier to sort. MainType can use the Panose information in fonts to sort or filter them. I hope that later versions of Font Creator or MainType will offer some kind of Panose Wizard to make the process semi-automatic.

It would be fantastic if one could add Panose info to a font just by selecting check boxes on dialogue boxes with illustrations of the different font characteristics.

[Erwin Denissen]

I also tried www.panose.com without success.

That domain’s Registrant (Monotype Imaging, Inc.) is still online; probably just a temporary server/ router problem.

Dave,

I like the tutorial, it contains some valuable information. I've already added Panose improvements to the to-do list.

[Dick Pape]

Good effort Dave.

While you and I did discuss this Panose business over a cup of tea or two, you decided to go ahead and publish it anyway!! Hah.

I agree with Bhikkhu Pesala about having some mechanism to make it automatic or easy to assign valid Panose codes for those of us who "need help" or might forget. It is something Font Validator can easily support I feel.

Where it might become more useful is communicating with others who may not have the same fonts installed as you. While never perfect, and maybe not in the same "range" as 'helpful', you can aid the operating system select an alternate font if your Panose are proper and it might help your readers (especially when Common Bullets are similar to Arial!!)

I was surprised to find The Font Thing used Panose for selecting groups of fonts. I was not surprised that Microsoft's Font Validator program checks the Panose values to the weight assigned in some of the tables. I was not surprised to find very few font designers/vendors use Panose correctly. And I was surprised when I checked some of the operating system fonts to find gross errors. Support therefore is all over the ball park.

We Can Do Better, huh, Dave!

We don't know the eventual direction of Panose, but it's easy to set up the first 3 digits in all Latin font types which WILL help substituting and the grouping of fonts. Since Panose is included in the font tables, it is a natural place to try to setup classification information.

[Dick Pape]

On a related thread I have used FC to create valid Panose codes.

Once when I had a lot of extra time, I built a Panose encoding feature for my home-grown data base of fonts. What was so terrific and relevant was I used the Font Creator Measure Length and Angle tool for the required measurements.

Whereas the Panose documentation says you have to print out 200pt images of certain characters to correctly measure angles and lengths, that is not necessary with FC as the Measure tool gives you the numbers you need directly. I can calculate all Panose measurements online without printing a thing.

There remain holes where a measure asks for visual interpretation of such things as Hand Written tool kind, topology, form or finials. These have pretty clear meanings and can be determined easily, even if they're not measureable.

I have proven that Panose is not a mysterious science and if you have a need Font Creator can help you fill it!!

Thank you, Erwin!!

DP

Sure wish I had more time again.

[William]

I had not heard of the PANOSE system before. It is very interesting. Thank you for your document.

A system with a somewhat similar-sounding name about which some readers may like to know is the Pantone system, which is used for describing colours.

http://www.pantone.com

William Overington

[Dave Crosby]

I say Dick,

a cup of tea or two,

My Panose E-mail file got so heavy it almost tipped my computer over!
Thanks for hanging in there!

[Dick Pape]

Hi Dave -- where are you? Hope you got your connection repaired.

I saw this conversation regarding a current use of PANOSE and thought I'd dump it here.

Subject: Re: When is a monospace font not a monospace font?

I've installed the Excalibur Monospace font, but it won't show up in the list of available fonts in any of my apps that allow only monospace fonts to be selected (NoteTab Pro or JCreator, for example). It *does* show up in programs that allow variable *or* fixed width fonts (Notepad or WordPad, for example). When I use Excalibur Mono in one of these programs, it very clearly does have a fixed width.

What's up with this? Is the font file missing some sort of 'header' or 'flag' or whatever that would prevent fixed-width-font-only apps from recognizing it as such? Or something??


Response: It's just a flag that wasn't set before generating the font. Whoever designed the Excalibur Monospace font didn't designate it as a monospaced font in its Panose identification section. This is the same flag that monospace-fonts-only apps probe when selecting the fonts to be displayed. The app looks at the "Proportion" record of the font's PANOSE table. If it finds the value to be anything other than "Monospaced", it doesn't display the font.

This problem is easily fixable if you have Fontlab or Fontographer or Font Creator. Just open the font and redefine it as monospaced in the Panose table.


Thought it was good to know. Dick

[Bhikkhu Pesala]

I don't know quite how to classify a font like this.



I got as far as latin hand-written (script), tool maybe brush, rough, or wild brush. Then what? How to measure the weight?

The vertical stroke of the E varies enormously depending on where you measure it, and even the caps height is not too obvious. Visually, it looks like at least bold, but strict measurement gives Book with a WeightRat of about 9.

[Dave Crosby]

Ay, there’s the rub.

Much of Panose is in the eye of the Panoser.

Which item 2 about the font catches YOUR eye first,

3.2 Tool Kind,
or 4.2 Class?

I probably would have picked 43 -Non-standard Topology, but that is just me.

You saw what the "experts" did with Common Bullets.

[Dick Pape]

From PANOSE Classificatrion Metrics Guide - Monotype Imaging. (I think its the same for all Families).

"3. Weight

The Weight digit classifies the appearance of a fonts’ stroke thickness in relation to its height. This is expressed as a comparison of the measurements taken on the uppercase E glyph and the Upper H used before.

Sub-digits
0-Any
1-No Fit
2-Very Light
3-Light
4-Thin
5-Book
6-Medium
7-Demi
8-Bold
9-Heavy
10-Black
11-Extra Black

Measurements:

Two measurements are required for classification of the Weight digit.

CapH
The same measurement used in the start of the serif classification is used to begin the weight classification. CapH (Figure 2) is the cap height and is measured on the uppercase H, from the top-most Y-extent to the bottom-most Y-extent at the theoretical midline of the left vertical stroke. The midline is chosen to avoid serifs that extend the height or depth of the character shape. This is a vertical measurement even if the glyph is italic or oblique.

WStem(E)
The width of the vertical stem, WStem(E) (Figure 5), is measured horizontally on the uppercase E at a point halfway between the upper two arms. This measurement is the width of the vertical stem, or back bone, of the character and is taken perpendicular to the stem. In the case of an oblique letter, the horizontal axis is shifted to be perpendicular to the stem. Note: For the purpose of serif designs, this measurement is applied to the large (400 point) uppercase I glyph. This measurement is used to set the nominal weight of the overall font.

Calculated Variables:

Only one calculated variable (WeightRat) is used to determine the Weight digit for the PANOSE Typeface Matching System. The WeightRat variable is calculated by dividing the cap height by the width of the vertical stem.

WeightRat = CapH / WStem(E)

Classification:

To determine the exact PANOSE Weight digit, round the WeightRat value to two decimal places and match it in the following table:

0-Any
1-No fit
2- Very Light …………………WeightRat >= (greater and equal) 35
3-Light………………… 18 <=WeightRat < 35
4-Thin……………………10 <=WeightRat < 18
5-Book ………………….7.5 <=WeightRat < 10
6-Medium……………….5.5 <=WeightRat < 7.5
7-Demi ………………….4.5 <=WeightRat < 5.5
8-Bold …………………..3.5 <=WeightRat < 4.5
9-Heavy………………… 2.5 <=WeightRat < 3.5
10-Black………………… 2.0 <=WeightRat < 2.5
11-Extra Black………………… WeightRat < 2.0


Works every time...! Dick

[Dick Pape]

OK. I used a font called Paintbrush from SWFTE International, Ltd which was very similar and calculated the PANOSE variables using a calculator designed into my fonts data base.

Came up with the following:

Family: 3-Script
Tool Kind: 6-Brush
Weight: 5-Book
Spacing: 2-Proportional Spaced
Aspect: 4-Normal
Contrast: 6-Medium
Topology: Cursive Disconnected
Form: 11 Exaggerated/Some Wrapping (I didn't use hex)
Finials: 8 Tapered/No Loops
X-Ascent: 4-Medium

Needed 13 measurements which are easily done with FC and the Measuring Tool. Took maybe 10 minutes to find my notes on how to do it and 5 minutes to measure and calculate the codes.

Tool Kind, Topology and Finials were judgment calls.

Works faster when you remember what a Ewid and Eout are.

Dick

[Bhikkhu Pesala]

Mostly beyond me, but Family: Script and Tool Kind: Brush are fair enough.

The weight is not right. No way is this font "Book" or "Thin" weight. Maybe Demi rather than Bold, but not medium or book, let alone thin. The recommended measurement method just doesn't fit this kind of font.

Below is a comparison at the same point size with Carita Bold, a bold font with an easily measured stem and caps height giving a Weight Ratio of about 4.3 in the Bold range.



With this script font the stem is weighted heavily towards the bottom, if I measure the stem in the suggested place I get a value of 112.



The Caps Height is not so controversial. I measured it from the middle of the bottom of the left stem.



This gives a value for CapH of 1427 and a Weight Ratio of 12.74 which is well into the Thin range.

The other dilemma is that the droplist in Font Creator's classification dialogue doesn't match the entries for script fonts. There is no "Tool Kind" just "Serif Style." I have to calculate that the tool kind of "Brush" (6) corresponds to the seventh item in the list of Serif Styles = "Thin", the first one in the list "Any" corresponding to a Panose number of zero.

This makes a difficult task almost impossible for Joe Average, or even for Bhikkhu average.
Without the help of someone who had read up on this at some length, I would still be lost.

[Dick Pape]

The real intent of Panose is simply to provide a consistent measurement scheme for all fonts so that for all such "asian brush stroke" fonts the Panose would be the same -- and substitution straightforward. Panose specs do make some disclaimer that the more severe the design the less this works. The calculations for Text fonts is 4 times more extensive and for Symbols there are 5.

Again, there are enough needed "your choice" decisions -- tool kind, topology and finials that there will be a variety of outcomes from several reviewers. With such a wide variation in stroke width you must ask what is the middle ground between Light and Bold -- perhaps Book (well, maybe Medium...)

Also, this is no worse then what some font designers name their creations. It's all in the eye of the beholder as Dave said -- so if I wanna call my font Black -- IT'S BLACK!!

I believe E.D. will correct the Panose classes on the next update. I also believe also you're one of the first to seriously walk through the process except for Dave and me and Erwin and possibly William. Thank you!

[Bhikkhu Pesala]

This Tutorial really helps to understand how to make use of the Panose Classification system without making heavy weather out of it.

The latest version of FontCreator 5.5 has further facilitated the process by correcting the names of the various fields and linking them to Family Kind.



Now you can just cut and paste the Panose Number from another type style or font to quickly fill in most of the fields, then modify just those that need to be changed. If your font is something like Verajja, for example, just cut and paste: 2-11-6-3-3-8-4-2-2-4 then for bold you just modify the weight field and the number changes to: 2-11-8-3-3-6-4-2-2-4