This chapter outlines recommendations for creating OpenType fonts.
All OpenType fonts use Motorola-style byte ordering (Big Endian).
OpenType fonts that contain TrueType outlines should use the value of 1.0 for the sfnt version. OpenType fonts containing CFF data should use the tag 'OTTO' as the sfnt version number.
Both Microsoft and Adobe recommend against mixing outline formats within a single font. Choose the format that meets your feature requirements.
OpenType fonts may have the extension .OTF, .TTF, or .TTC, depending on the type of outlines in the font and the presence of OpenType layout tables.
All tables should be aligned to begin at offsets which are multiples of four bytes. While this is not required by the TrueType rasterizer, it does prevent ambiguous checksum calculations and greatly speeds table access on some processors.
All tables should be recorded in the table directory with their actual length. To ensure that checksums are calculated correctly, it is suggested that tables begin on LONG word boundries. Any extra space after a table (and before the next LONG word boundry) should be padded with zeros.
TrueType outline fonts should have the following four glyphs at the glyph ID indicated. These were listed in Apple's original TrueType specification. These glyphs are recommended to allow for the same version of the font to work on both Windows and Macintosh.
|Glyph ID||Glyph name||Unicode value|
The .notdef glyph is very important for providing the user feedback that a glyph is not found in the font. This glyph should not be left without an outline as the user will only see what looks like a space if a glyph is missing and not be aware of the active font's limitation.
It is recommended that the shape of the .notdef glyph be either an empty rectangle, a rectangle with a question mark inside of it, or a rectangle with an “X”. Creative shapes, like swirls or other symbols, may not be recognized by users as indicating that a glyph is missing from the font and is not being displayed at that location.
The 'BASE' table allows for different scripts in the font to specify different values for the same baseline tag. This situation could arise when a developer makes a Unicode font, for example, by combining glyphs from fonts that use different baseline systems.
However, glyphs from different scripts in this font may not appear correctly aligned relative to each other when used with applications that either don't support the 'BASE' table or that support it but assume that a particular baseline will not vary across scripts. Furthermore, it is not always possible to determine the script of every glyph in the font, some “weakly-scripted” characters such as punctuation may be used in several scripts, and some glyphs such as ornaments may not have a script at all.
Thus, it is strongly recommended that developers construct their fonts so that all scripts in the 'BASE' table record the same value for a particular baseline if they want their fonts to work as expected in the above situations.
If baselines vary by script, it is strongly recommended that the vendor add a DFLT script entry to the BASE table, which can be used if the script requested by the client is not matched, or if the client does not or cannot determine the script.
When building a Unicode font for Windows, the platform ID should be 3 and the encoding ID should be 1 (this subtable must use cmap format 4). When building a symbol font for Windows, the platform ID should be 3 and the encoding ID should be 0.
When building a font to support surrogate characters i.e. the UCS-4 (4 byte) form of ISO 10646 (ISO 10646 UCS-4 contains 2^31 code positions and the Unicode transformation formats UTF-8 and UTF-16 access a subset of these code positions using surrogate characters), use platform ID 3, encoding ID 10 and format 12. Depending on support installed and the content of text being displayed, Windows 2000 may use either the format 4 or format 12 cmap. Therefore the first 64k codepoint to glyph mappings must be identical for any font containing both cmap format 4 and format 12. Please note, that the content of format 12 subtable, needs to be a super set of the content in the format 4 subtable. The format 4 subtable needs to be included, for backward compatibility needs.
The number of glyphs that may be included in one font is limited to 64k.
Remember that, despite references to 'first' and 'second' subtables, the subtables must be stored in sorted order by platform and encoding ID.
When building a font containing Roman characters that will be used on the Macintosh, an additional subtable is required, specifying platform ID of 1 and encoding ID of 0 (this subtable may use cmap formats 0, 2, 4, or 6).
In order for the Macintosh 'cmap' table to be useful, the glyphs required for the Macintosh must have glyph indices less than 256 (since the 'cmap' subtable format 0 uses BYTE indices and therefore cannot index any glyph above 255).
The Apple 'cmap' subtable should be constructed according to Apple guidelines.
Should be defined only if required by font instructions.
Should be defined only if required by TrueType font instructions.
The 'glyf' table contains TrueType outline data, and can be optimized by Agfa MicroType Compression. It is recommended that developers perform this optimization prior to finalizing and adding a digital signature to the font. This is necessary for the creator's signature to remain valid in embedded OFF fonts.
This table improves the performance of OpenType fonts with TrueType outlines. This table is not necessary at all unless instructions are used to control the “phantom points,” and should be omitted if bits 2 and 4 of the flags field in the 'head' table are zero. (See the 'head' table documentation in Chapter 2.) Microsoft recommends that this table be included for fonts with one or more non-linearly scaled glyphs (i.e., bit 2 or 4 of the 'head' table flags field are set).
Device records should be defined for all sizes from 8 through 14 point, and even point sizes from 16 through 24 point. However, the table requires pixel-per-em sizes, which depend on the horizontal resolution of the output device. The records in 'hdmx' should cover both 96 dpi devices (CGA, EGA, VGA) and 300 dpi devices (laser and ink jet printers).
Thus, 'hdmx' should contain entries for the following pixel sizes (PPEM): 11, 12, 13, 15, 16, 17, 19, 21, 24, 27, 29, 32, 33, 37, 42, 46, 50, 54, 58, 67, 75, 83, 92, 100. These values have been rounded to the nearest pixel. For instance, 12 points at 300 dpi would measure 37.5 pixels, but this is rounded down to 37 for this list.
This will add approximately 9,600 bytes to the font file. However, there will be a significant improvement in speed when a client requests advance widths covered by these device records.
If the font includes an 'LTSH' table, the hdmx values are not needed above the linearity threshold.
Although historical usage of the fontRevision value is varied, the recommended use of the field is to set it as a Fixed 16.16 value, and to report it rounded and zero-padded to three fractional decimal places. Examples: Decimal 1.5 is set as 0x00018000 and is reported as “1.500”; decimal 1.001 is set as 0x00010041 and is reported as “1.001”. All data required. If the font has been compressed with Agfa MicroType Compression, this must be indicated in the flags field of the 'head' table.
All data required. It is suggested that monospaced fonts set numberOfHMetrics to three (see hmtx).
All data required. It is suggested that monospaced fonts have three entries in the numberOfHMetrics field. OpenType fonts that include CFF data must set numberOfHMetrics equal to the number of glyphs in the font and therefore cannot use the “repeat last width” optimization normally available within the 'hmtx' table.
Should contain a single kerning pair subtable (format 0). Windows will not support format 2 (two-dimensional array of kern values by class); nor multiple tables (only the first format 0 table found will be used) nor coverage bits 0 through 4 (i.e. assumes horizontal data, kerning values, no cross stream, and override).
The OFF specification allows CFF OT fonts to express their kerning in a kern table. Many OFF text layout engines support this. Windows GDI’s CFF OT driver, however, ignores the kern table in a CFF OT font when it prepares kerning pairs to report via its pair kerning API.
When a kern table and GPOS table are both present in a font, and an OFF layout engine is requested to apply kerning to a run of text of a particular script and language system: (a) If the number of kern feature lookups in the resolved language system in the GPOS table is zero, then the kern table should be applied, followed by any remaining GPOS features requested. (b) If the number of kern feature lookups in the resolved language system in the GPOS table is non-zero, then all GPOS lookups, including the kern lookups, should be applied in the usual way and the kern table data ignored.
If a kern table present but no GPOS table is present in the font, then an OFF layout engine should apply the kern table to the text, regardless of the resolved language system of the text.
If compatibility with legacy environments is not a concern, font vendors are encouraged to record kerning in the GPOS table's kern feature and not in the kern table.
All data required for fonts with TrueType outlines. We recommend that local offsets should be word-aligned, in both the short and long formats of this table.
The actual ordering of the glyphs in the font can be optimized based on expected utilization, with the most frequently used glyphs appearing at the beginning of the font file. Additionally, glyphs that are often used together should be grouped together in the file. The will help to minimize the amount of swapping required when the font is loaded into memory.
This table improves the performance of OpenType fonts with TrueType outlines. The table should be used if bit 2 or 4 of flags in 'head' is set.
All data required for a font with TrueType outlines. Fonts with CFF data must only fill the numGlyphs field.
Platform and encoding ID's in the name table should be consistent with those in the cmap table. If they are not, the font will not load in Windows. When building a Unicode font for Windows, the platform ID should be 3 and the encoding ID should be 1. When building a symbol font for Windows, the platform ID should be 3 and the encoding ID should be 0.
When building a font containing Roman characters that will be used on the Macintosh, an additional name record is required, specifying platform ID of 1 and encoding ID of 0.
Each set of name records should appear for US English (language ID = 0x0409 for Microsoft records, language ID = 0 for Macintosh records); additional language strings for the Microsoft set of records (platform ID 3) may be added at the discretion of the font vendor.
Remember that, despite references to “first” and “second,” the name record must be stored in sorted order (by platform ID, encoding ID, language ID, name ID). The 'name' table platform/encoding IDs must match the 'cmap' table platform/encoding IDs, which is how Windows knows which name set to use.
We recommend using name ID's 8-12, to identify manufacturer, designer, description, URL of the vendor, and URL of the designer. URL's must contain the protocol of the site: for example, http:// or mailto: or ftp://. The OpenType font properties extension can enumerate this information to the users.
The Subfamily string in the 'name' table should be used for variants of weight (ultra light to extra black) and style (oblique/italic or not). So, for example, the full font name of “Helvetica Narrow Italic” should be defined as Family name “Helvetica Narrow” and Subfamily “Italic”. This is so that Windows can group the standard four weights of a font in a reasonable fashion for non-typographically aware applications which only support combinations of “bold” and “italic.”
The Full font name string usually contains a concatenation of strings 1 and 2. If the font is 'Regular' as indicated in string 2, then sometimes only the family name contained in string 1 is used for the full font name. This is the font name that Windows will expose to users.
All data required. We recommend applying PANOSE values to fonts to improve the user's experience when using the Windows fonts folder or other font management utilities. If the font is a symbol font, the first byte of the PANOSE value must be set to 'decorative.' The PANOSE evaluation document is on-line at www.fonts.com/hp/panose/greybook/frame.htm.
sTypoAscender is used to determine the optimum offset from the top of a text frame to the first baseline. sTypoDescender is used to determine the optimum offset from the last baseline to the bottom of the text frame. The value of (sTypoAscender - sTypoDescender) is recommended to equal one em.
While the OpenType specification allows for CJK (Chinese, Japanese, and Korean) fonts' sTypoDescender and sTypoAscender fields to specify metrics different from the HorizAxis.ideo and HorizAxis.idtp baselines in the 'BASE' table, CJK font developers should be aware that existing applications may not read the 'BASE' table at all but simply use the sTypoDescender and sTypoAscender fields to describe the bottom and top edges of the ideographic em-box. If developers want their fonts to work correctly with such applications, they should ensure that any ideographic em-box values in the 'BASE' table describe the same bottom and top edges as the sTypoDescender and sTypoAscender fields. See the sections “OpenType CJK Font Guidelines“ and ”Ideographic Em-Box“ for more details.
For Western fonts, the Ascender and Descender fields in Type 1 fonts' AFM files are a good source of sTypoAscender and sTypoDescender, respectively. The Minion Pro font family (designed on a 1000-unit em), for example, sets sTypoAscender = 727 and sTypoDescender = -273.
sTypoAscender, sTypoDescender and sTypoLineGap specify the recommended line spacing for single-spaced horizontal text. The baseline-to-baseline value is expressed by:
OS/2.sTypoAscender - OS/2.sTypoDescender + OS/2.sTypoLineGap
sTypoLineGap will usually be set by the font developer such that the value of the above expression is approximately 120% of the em. The application can use this value as the default horizontal line spacing. The Minion Pro font family (designed on a 1000-unit em), for example, sets sTypoLineGap = 200.
All information required, although the VM Usage fields may be set to zero. OpenType fonts containing CFF outlines use only format 3.0 of the 'post' table. Glyph names must be assigned as described in the Adobe document “Unicode and Glyph Names,“ which specifies glyph naming conventions for all Unicode characters as well as those that don't have standard Unicode values such as certain ligatures or glyphic variants. Note that names for all glyphs must be supplied as it cannot be assumed that all Microsoft platforms will support the default names supplied on the Macintosh.
Should be defined only if required by the TrueType font instructions.
This table improves the performance of OpenType fonts with TrueType outlines. It Should be present if hints cause the font to scale non-linearly. If not present, the font is assumed to scale linearly. Clipping may occur if values in this table are absent and font exceeds linear height.
OpenType fonts with TrueType outlines are more efficient in the Windows operating system when the tables are ordered as follows (from first to last):
head, hhea, maxp, OS/2, hmtx, LTSH, VDMX, hdmx, cmap, fpgm, prep, cvt, loca, glyf, kern, name, post, gasp, PCLT, DSIG
The initial loading of an OpenType font containing CFF data will be more efficiently handled if the following sfnt table ordering is used within the body of the sfnt (listed from first to last):
head, hhea, maxp, OS/2, name, cmap, post, CFF, (other tables, as convenient)
Non-standard fonts such as Symbol or Wingdings(tm) have special requirements for Microsoft platforms. These requirements affect the 'cmap,' 'name,' and 'OS/2' tables; the requirements and recommendations for all other tables remain the same.
For the Macintosh, non-standard fonts can continue to use platform ID 1 (Macintosh) and encoding ID 0 (Roman character set). The 'cmap' subtable should use format 0 and follow the standard PostScript character encodings.
For non-standard fonts on Microsoft platforms, however, the 'cmap' and 'name' tables must use platform ID 3 (Microsoft) and encoding ID 0 (Unicode, non-standard character set). Remember that 'name' table encodings should agree with the 'cmap' table. Additionally, the first byte of the PANOSE value in the 'OS/2' table must be set to 'decorative.'
The 'cmap' subtable (platform 3, encoding 0) must use format 4. The character codes should start at 0xF000, which is in the Private Use Area of Unicode. It is suggested to derive the format 4 encodings by simply adding 0xF000 to the format 0 (Macintosh) encodings.
Under Windows, only the first 224 characters of non-standard fonts will be accessible: a space and up to 223 printing characters. It does not matter where in user space these start, but 0xF020 is suggested. The usFirstCharIndex and usLastCharIndex values in the 'OS/2' table would be set based on the actual minimum and maximum character indices used.
Windows makes use of a logical device resolution. The physical resolution of a device is also available, but fonts will be rendered based on the logical resolution. The table below lists some important logical resolutions in dots per inch (Horizontal x Vertical). The most important ratios (in order) are 1:1, 1.67:1 and 1.33:1.
|CGA||96 x 48||2:1|
|EGA||96 x 72||1.33:1|
|VGA||96 x 96||1:1|
|8514||120 x 120||1:1|
|Dot Matrix||120 x 72||1.67:1|
|Laser Printer||300 x 300||1:1|
|Laser Printer||600 x 600||1:1|
The 'OS/2' table fields sTypoAscender, sTypoDescender, and sTypoLineGap free applications from Macintosh- or Windows-specific metrics which are constrained by backward compatibility requirements. The following discussion only pertains to the platform-specific metrics.
The suggested Baseline to Baseline Distance (BTBD) is computed differently for Windows and the Macintosh, and it is based on different OpenType metrics. However, if the recommendations below are followed, the BTBD will be the same for both Windows and the Mac.
The Windows metrics in the table below are returned as part of the logical font data structure by the GDI CreateLogFont( ) API.
|Windows Metric||OpenType Metric|
|internal leading||usWinAscent + usWinDescent - unitsPerEm|
|external leading||MAX(0, LineGap - ((usWinAscent + usWinDescent) - (Ascender - Descender)))|
The suggested BTBD = ascent + descent + external leading
It should be clear that the “external leading” can never be less than zero. Pixels above the ascent or below the descent will be clipped from the character; this is true for all output devices.
The usWinAscent and usWinDescent are values from the 'OS/2' table. The unitsPerEm value is from the 'head' table. The LineGap, Ascender and Descender values are from the 'hhea' table.
Ascender and Descender are metrics defined by Apple and are not to be confused with the Windows ascent or descent, nor should they be confused with the true typographic ascender and descender that are found in AFM files. The Macintosh metrics below are returned by the Apple Advanced Typography (AAT) GetFontInfo( ) API.
|Macintosh Metric||OpenType Metric|
The suggested BTBD = ascent + descent + leading
If pixels extend above the ascent or below the descent, the character will be squashed in the vertical direction so that all pixels fit within these limitations; this is true for screen display only.
If you perform some simple algebra, you will see that the suggested BTBD across both Macintosh and Windows will be identical if and only if:
LineGap >= (yMax - yMin) - (Ascender - Descender)
For backwards compatibility with previous versions of Windows, the macStyle bits in the 'head' table will be used to determine whether or not a font is regular, bold or italic (in the absence of an 'OS/2' table). This is completely independent of the usWeightClass and PANOSE information in the 'OS/2' table, the ItalicAngle in the 'post' table, and all other related metrics. If the 'OS/2' table is present, then the fsSelection bits are used to determine this information.
Drop-out control is needed if there is a difference in bitmaps with dropout control on and off. Two cases where drop-out control is needed are when the font is rotated or when the size of the font is at or below 8 ppem. Do not use SCANCTRL unless needed. SCANCTRL or the drop-out control rasterizer should be avoided for Roman fonts above 8 points per em (ppem) when the font is not under rotation. SCANCTRL should not be used for “stretched” fonts (e.g. fonts displayed at non-square aspect ratios, like that found on an EGA).
Three tables are used to embed bitmaps in OpenType fonts. They are the 'EBLC' table for embedded bitmap locators, the 'EBDT' table for embedded bitmap data, and the 'EBSC' table for embedded bitmap scaling information. OpenType embedded bitmaps are also called 'sbits'.
The behavior of sbits within an OpenType font is essentially transparent to the client. A client need not be aware whether the bitmap returned by the rasterizer comes from an sbit or from a scan-converted outline.
The metrics in 'sbit' tables overrule the outline metrics at all sizes where sbits are defined. Fonts with 'hdmx' tables should correct those tables with 'sbit' values.
'Sbit only' fonts, that is fonts with embedded bitmaps but without outline data, are permitted. Care must be taken to ensure that all required OpenType tables except 'glyf' and 'loca' are present in such a font. Obviously, such fonts will only be able to return glyphs and sizes for which sbits are defined.
This section provides a checklist of links to various CJK-related sections of the OpenType specification. Some items are requirements; others, recommendations:
This page was last updated 28 May 2010.
Comments to the MST group: how to contact us.