Hi Hans, many thanks for having released the new beta. I experience opposite results with opbd. The first sample is: \definefontfeature[default][default][protrusion=quality, opbd=yes] \definefontfamily[mainface][rm][Linux Libertine O] \setupbodyfont[mainface, 300pt] \setuppagenumbering[location=] \setupalign[hanging] \showframe \starttext W \stoptext Although the value for that lfbd is only -29, the W is moved too much from its original position. The second sample is (font attached at https://mailman.ntg.nl/pipermail/ntg-context/2018/092527.html): \showframe \definefontfeature[default][default][protrusion=quality, opbd=yes] \definefontfamily[mainface][rm][TestOPBD] \setupbodyfont[mainface, 60pt] \setupalign[hanging] \starttext Ἠ\\ Ἠ\hfill Ἠ\\ Ἦ\hfill ὩὩ\\ ἯἯ\hfill ὩὩ \stoptext Here the changes are tiny. The right values are used in the image at https://github.com/wspr/fontspec/issues/282#issuecomment-287155442. Could you check it? Many thanks for your help, Pablo -- http://www.ousia.tk
On 11/18/2018 10:22 PM, Pablo Rodriguez wrote:
Hi Hans,
many thanks for having released the new beta.
I experience opposite results with opbd. The first sample is:
\definefontfeature[default][default][protrusion=quality, opbd=yes] \definefontfamily[mainface][rm][Linux Libertine O] \setupbodyfont[mainface, 300pt] \setuppagenumbering[location=] \setupalign[hanging] \showframe \starttext W \stoptext
Although the value for that lfbd is only -29, the W is moved too much from its original position.
i have no time now but keep in mind that we also have a multiplier (the system is set up to use ratios / fractions, not hard coded numbers also, it makes no sense to experiment too much with only some demo fonts with incomplete coverage ... adapting the low level mechanism to just one font with one character protruding makes no sense (one can tnen better use the non-opbd way)
The second sample is (font attached at https://mailman.ntg.nl/pipermail/ntg-context/2018/092527.html):
\showframe \definefontfeature[default][default][protrusion=quality, opbd=yes] \definefontfamily[mainface][rm][TestOPBD] \setupbodyfont[mainface, 60pt] \setupalign[hanging]
\starttext Ἠ\\ Ἠ\hfill Ἠ\\ Ἦ\hfill ὩὩ\\ ἯἯ\hfill ὩὩ \stoptext
Here the changes are tiny. The right values are used in the image at https://github.com/wspr/fontspec/issues/282#issuecomment-287155442.
Could you check it?
Many thanks for your help,
Pablo
-- ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | www.pragma-ade.nl | www.pragma-pod.nl -----------------------------------------------------------------
On 11/18/18 10:37 PM, Hans Hagen wrote:
On 11/18/2018 10:22 PM, Pablo Rodriguez wrote:
[...] \starttext W \stoptext
Although the value for that lfbd is only -29, the W is moved too much from its original position.
i have no time now but keep in mind that we also have a multiplier (the system is set up to use ratios / fractions, not hard coded numbers
Many thanks for your reply, Hans. Although optical bounds in OpenType and character protrusion are very similar in their goals, I think they are different. Some glyphs need optical bounds, such as in the following example: Hanging diacritics seems to be a typographic practice in the Greek tradition. In poetry, uppercase letters with diacritical marks at the beginning of the verse are aligned as if they had no diacrical mark (https://cloud.disroot.org/s/23wXgw2FHN9CHtM). The only way to achieve it is the use of the lfbd tag. But if the values are recalculated, it is impossible to get the desired results. It is impossible to make the desired glyphs work as expected.
also, it makes no sense to experiment too much with only some demo fonts with incomplete coverage ... adapting the low level mechanism to just one font with one character protruding makes no sense (one can tnen better use the non-opbd way)
I love character protrusion and I avoid optical bounds as much as I can (I used the Linux Libertine font because you used it, I would never use optical bounds with Linux Libertine). But there are some cases in which optical bounds are needed and their values shouldn’t be altered. Would it be possible that ConTeXt deploys the optical bound values exactly from the font without multiplying them? If the font has wrong values, and character protrusion should be the solution there. Many thanks for your help, Pablo -- http://www.ousia.tk
participants (2)
-
Hans Hagen -
Pablo Rodriguez