[smufl-discuss] Proposing the removal of the wind fingering charts ranges from SMuFL 1.0

[dspreadbury]

Dear community,

Further to my last message concerning the current state of progress
towards 1.0, the remaining issue concerns zero-width characters, i.e.
characters that have an advance width of zero. The tl:dr version is that I
propose removing the wind fingering charts ranges from SMuFL in order to
eliminate the use of zero-width characters (likewise removing them from
Bravura and instead releasing them as a separate font under the SIL Open
Font License), and ask that any strong objections or counter-proposals are
made before this Friday, 30 May. Please do read on for the full story,
however.

Advance width is the amount by which the input caret advances when you
enter a character in e.g. a text editor or a word processor; normally, the
full horizontal width of a glyph as measured from its origin to the origin
of the next glyph on the line, including the side bearings on both sides.
Having an advance width of zero allows glyphs to be overprinted when they
are entered successively in a run of text. This is very useful for some
applications: for example, Bravura Text makes use of zero-width characters
to allow staff lines to be drawn behind notes and other musical symbols.

Using zero-width glyphs can cause problems in different situations,
because their behaviour tends to vary by application, operating system and
drawing/shaping API. The use of zero-width glyphs is practically
unavoidable in Bravura Text (because of the requirement for overprinting
described above), but for Bravura it would definitely be desirable to
eliminate their use. Bravura is likely to be used in scenarios where the
precise horizontal and vertical positioning and measurement of glyphs is
critical, so it is definitely preferable to have glyphs that render
reliably in as many contexts as possible.

Bravura 0.9 currently uses zero-width characters in the following
situations:

1. Control characters with empty paths (e.g. timeSigNumerator,
staffPosRaise1, controlBeginTie).
2. Percussion beater glyphs tilted leftwards (e.g.
pictBeaterSoftXylophoneLeft and its ilk).
3. Certain characters in the 'Beamed groups of notes' range (e.g.
textCont8thBeamShortStem).
4. The ligated time signature digits positioned to be the numerator of a
time signature (e.g. timeSig4Numerator, at code point F54D).
5. All of the glyphs in the wind fingering chart ranges.

Taking each of these use cases in turn, #1 presents no real problem:
control characters aren't designed to print anything anyway, so having an
empty, zero-width glyph is acceptable.

For #2 and #4, it's possible to achieve the same results through judicious
use of pair kerning. The intended use of the left-tilted beaters is to be
overlapped with the right-tilted beater, to show which beaters should be
held by a player. To achieve this result using pair kerning, the
left-tilted beater can have its regular advance width, and when used
adjacent (either before or after) the right-tilted beater, the pair can be
kerned such that they overlap as intended. Likewise, for the time
signature digits, when the appropriate pair of glyphs appears (e.g.
timeSig4Numerator followed immediately by timeSig4Denominator), the pair
kerning can be set such that the second glyph is kerned leftwards to the
point that it sits directly underneath the first glyph.

For #3 it's more complicated, because it's useful to be able to have a
zero-width character, like the above-mentioned textCont8thBeamShortStem,
and either use a space character to advance or put another glyph at the
same position, e.g. a rhythm dot, to make a dotted 8th-16th figure, or
whatever. To achieve good results with kerning will require around 400
kerning pairs to be defined, but it's within the realms of practicality to
achieve this.

However, eliminating zero-width characters from the wind fingering charts
(#5) really seems to be a non-starter. Each fingering chart contains
dozens of glyphs, which could be typed in any order, which means that the
kerning pairs required to ensure that the filled or half-filled glyphs are
correctly registered with the blank fingering chart will number into the
thousands. Using zero-width characters really is the simplest way to
achieve the right result, and unfortunately even making the advance width
of each glyph 1 du won't produce good results when that 1 du "rounding
error" in registration is multiplied out over a dozen or so glyphs in a
single completed fingering chart. It is possible that there is another
advanced glyph substitution or glyph positioning feature tied to a
particular font technology (OpenType or AAT) that would provide a solution
to this problem, but in general we are trying to avoid the use of
font-technology-specific solutions to these kinds of problems in SMuFL.
(Pair kerning, by contrast, seems to be common to enough font formats etc.
to be possible to recommend.)

My proposal therefore is the following:

1. Specify that SMuFL-compliant fonts intended for use in scoring
applications should only use zero-width glyphs for control characters.
2. Recommend the use of pair kerning to achieve overprinting effects for
beaters, ligated time signatures, and beamed groups of precomposed notes
set in a run of text.
3. Eliminate the wind fingering charts from SMuFL, since it is not in
general possible to achieve the correct result without the use of
zero-width characters.

Although the wind fingering chart ranges would be eliminated from SMuFL,
the work done on them would not be wasted: I would release a separate font
containing those glyphs, together with documentation for how to use them,
under the SIL Open Font License so that they can be used by anybody
interested in using them.

Unless there are strong objections expressed before this Friday, 30 May,
to the above proposal, I will go ahead with this plan for SMuFL 1.0.

Daniel

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Steinberg Media Technologies GmbH, Frankenstrasse 18b, D-20097 Hamburg, Germany
Phone: [phone] | Fax: [phone] | www.steinberg.net
President: Andreas Stelling | Managing Director: Hiroshi Sasaki, Hirofumi Osawa
Registration Court: Hamburg HRB 86534
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

#############################################################
This message is sent to you because you are subscribed to
  the mailing list <[hidden email]>.
To unsubscribe, E-mail to: <[hidden email]>
To switch to the DIGEST mode, E-mail to <[hidden email]>
To switch to the INDEX mode, E-mail to <[hidden email]>
Send administrative queries to  <[hidden email]>