[smufl-discuss] Re: Glyph Registration and Graphical Metadata

[David Webber]

From: Daniel Spreadbury

> At this point, I'm agnostic about whether or not it should be possible to
draw every possible musical symbol at any conceivable size without having
to use the font at more than one point size. My guess is that this is, in
its full generality, not a practical goal.

Certainly not "at any conceivable size" just at sizes needed for symbols in
music which is drawn at a given size.  Thus only two clef sizes are needed.

In my experience  (with two exceptions for which the reasons don't
necessarily apply here - see below) it is a practical goal, and certainly a
very worthwhile one.

>> I am happy with the recommendation not to use precomposed glyphs.  But I
>> *would* like a grace note heads, grace-note-sized accidentals, and
>> grace-note tail flags so I don't have to select the font at another size
>>>to draw grace notes.

> You would also, then, presumably want all noteheads, all accidentals and
all note tails available at a third size for cue-sized notes, which are a
different size than grace notes? Cue notes are (per Gould, p.659) 75% the
size of normal notes, and grace notes are "slightly smaller" (Gould,
p.125).<

Yes.   For grace notes I do it in the Mozart font.  There's really no
problem.

Originally I did it also for cue notes, but changed to using a full font at
a smaller size (one of the exceptions I mentioned above).  I'd rather have
gone on including cue notes as separate characters, but this was before
Unicode became ubiquitous, and I was running out of space in the Windows
symbol font, which admits rather fewer than 256 characters.  (Just to prove
a point: this is one area where I'd rather do it *differently* from the way
Mozart does it - but for cue notes, I'm not sure that it *is* practical -
see below.)

>  How much smaller than cue notes should grace notes be? You and I
may have different opinions about that!

But we're *allowed* different opinions!   Just as different text fonts are
allowed different length ascenders and descenders, different music fonts can
have different relative sizes of note and grace notes.  The different
relative dimensions give a different feel to the appearance of different
fonts.

> any case, we're now talking about adding 82 x 2 glyphs to the standard
to duplicate all noteheads, all accidentals, and all tails at three sizes. <

Yes (assuming it can be done for cue notes, and depending on whether you
feel shape note heads are necessary for grace notes).   But

(a) The PUA runs from U+E000 to U+F8FF  (if we consider the BMP only - there
are lots more elsewhere) which, if my sums are right, is 2034 points  (and
so 164 is only about 7% of them).     This is luxuriously spacious compared
with what I'm used to with symbol fonts.

(b) They're very quick for font designers to produce - just copy the main
one and change size to 75% or whatever.  they don't require much in the way
of extra design.

I'd also like a different set of accidentals designed to align with text.
(% more glyphs).

> And what about tremolo slashes? Do we need to provide them at three sizes?

Not for me:  like the stems, I use line-drawing to produce them.   I was
wondering why you included them given the recommendations to draw other
straight line objects (eg stems and bar lines).  But if you want, why not?
(I'm not sure that tremolo slashes on grace notes make any sense though, and
see below for cue notes.)

> What about articulations? What about other symbols that might be drawn on
stems, such as buzz rolls or sprechstimme crosses? What about jazz
articulations that are positioned to the left or right of notes? And on
and on.
> This is why I don't believe it is possible in general to provide the means
of drawing all conceivable music on a given staff size using separate
glyphs in a single font at a single size: the amount of duplication
required is really not insignificant.

It is also why, in the end, I didn't follow that route for cue notes.

>This seems to me a classic trade-off: duplicate hundreds of glyphs at two
smaller sizes in a font, and have the complexity of requiring calls to
separate code points for every possible symbol that could be drawn at
three sizes, versus having a consuming application using the font at up to
three sizes but use the same code points in each case.<

Yes.

> I remain deeply unconvinced about this, and my proposal is still to encode
glyphs only at their normal, i.e. full, size, and leave it up to consuming
applications to handle scaling.<

I can certainly appreciate it in the case of cue notes, but not at all in
the case of clefs and grace notes.

In order to make decisions in the case of a trade off such as this, one
really needs a guiding philosophy.   Mine, which evolved over a long period,
is broadly (with regard to the particular features we're discussing) as
follows.

I draw a score.  I want one of the staves and the music it contains to be at
a smaller size.  Typically this would be for a bassoon (say) soloist's part
reproduced on the piano accompaniment.  That is a 'cue stave' which can
contain just about any musical element drawn at the smaller size.   It makes
sense therefore to demand that the whole font is used at the smaller size.

The bassoon switches between bass and tenor clef.   For its own part it
needs CLEFs (defined for current purposes to be at the start of lines) and
CLEF-CHANGES which are 2/3 the size.   These are both also drawn on the
pianist's cue at reduced size.   It therefore makes sense to implement the
CLEFs and the CLEF-CHANGEs as different glyphs in the same font.

The bassoon part may have grace notes in it.   Again these would be drawn
reduced on the pianist's cue stave.  So again it makes sense that grace
notes belong in the font separately from notes, and if they are composed of
heads, tails and accidentals, so be it.    You didn't think of cued grace
notes?  Glad I got them in first  :-)  :-)

At this point it seems to me to be a coherent philosophy that the font
should contain both clefs and clef changes, and grace notes (with the
necessary accoutrements), but that the cued in part is written with a
smaller implementation of the font.

For cue notes added to a normal-sizes part, then it is consistent and
probably more practical (I still wish it weren't) to add them with a smaller
implementation of the font too.

So your 'extra' symbols are now reduced to 3 clefs, and whatever's needed
for grace notes.  The latter may be little more than two note-heads (black
and minim, which can be used for appoggiature) 5 up flags, 5 accidentals.
About 15.   You may want more for cross and diamond headed grace notes, and
grace notes with stems down, etc,  but even so it is not an inordinate
number.

In summary: I think your objections are very sound (sadly) for cue notes,
but not for clefs or grace notes.

Dave

David Webber
Mozart Music Software
http://www.mozart.co.uk/


















#############################################################
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]>