Proposed update of the 'head' table flags

[Levantovsky, Vladimir]

Dear all,

I am writing to you on behalf of the W3C WebFonts WG that is now developing a new webfont compression technology and webfont data delivery format known as WOFF 2.0.

The new compression technology is partially based on what was previously known as MicroType Express compression, which was implemented on the web as Embedded OpenType (EOT format, see [1]) and has also been used as the foundation of the MPEG font compression technology defined by ISO/IEC 14496-18 "Font compression and streaming". The essence of the MTX compression is to remove the built-in redundancy of the generic font format data and restructure the data to make it suitable for high efficiency entropy coding. As a result, when the font data is compressed and decompressed it is subjected to "lossless modifications", where the input and output font files retain the exact same functionality and features but are no longer bitwise compatible binary files.

The WOFF 2.0 compression technology is being developed with the same underlying principles in mind - it actually reuses portions of the MTX preprocessing and optimization steps but restructures the data to assemble it in seven different sub-streams to make the entropy coding more efficient. There is also a new entropy coding mechanisms that was developed to improve compression efficiency.

The OFF 'head' table contains 'flags field I believe that specify various font properties. Bit 11 of that field, as it is currently defined, indicates that
"Font data is 'lossless,' as a result of having been compressed and decompressed with the Agfa MicroType Express engine." I believe that the original intent was to make an emphasis that a font has been subjected to the lossless transform where the original font functionality is preserved but binary compatibility is lost - MTX (referenced as part of the bit 11 description) was the only known technology at that time to apply this kind of transformation but the use of this bit should not be limited to MTX only.

WebFonts WG believes that WOFF2 compliant encoders should also set this flag to indicate the fact that a font file was subjected to the lossless transform (which is similar but not identical to the original MTX) and suggests to modify the bit 11 description as follows:

"Bit 11: Font data is 'lossless' as a results of having been subjected to optimizing transformation and/or compression (such as e.g. compression mechanisms defined by ISO/IEC 14496-18, MicroType Express, WOFF 2.0 or similar) where the original font functionality and features are retained but the binary compatibility between input and output font files is not guaranteed. As a result of the applied transform, the 'DSIG' Table may also be invalidated."

I believe this new definition preserves the original intent of this flag and extends use cases when this flag setting may be a benefit. It will also provide an indication that certain transforms may cause Digital Signature table data be invalidated as a result.

Comments?

Thank you,
Vladimir

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.aau.at/pipermail/mpeg-otspec/attachments/20140430/38baac4d/attachment.html>