NOTATIO

John Ruggero wrote:Looking at it very casually, it seemed that beams have approx. a 45 degree limit in both directions, which presents an approx. 90 degree scope. Wouldn't every angle need to be taken in to account = 90 glyphs?

Regardless of how many you need, this would be very easy to generate using a simple script in FontForge, given max/min angles (e.g., +45/-45) and increment (e.g. 1). I'll give it a whirl and get back to you. The real challenge I think you'll deal with is actually USING it effectively.

Fantastico, tisimst!

tisimst, will the sides of the slashes remain perpendicular to the staff lines for all degrees of tilt? These seems highly desirable to me, and I thought I saw this in your LilyPond example as well.

John Ruggero wrote:tisimst, will the sides of the slashes remain perpendicular to the staff lines for all degrees of tilt? These seems highly desirable to me, and I thought I saw this in your LilyPond example as well.

The script I've got can do the following:
1. Just shear it like the beam
2. Sheared, but with thickness corrected (you probably wouldn't use this, since the beam itself doesn't do this)
3. Rotated with thickness like the beam
4. Rotated, but with thickness corrected (again, you probably wouldn't use this)

You can take your pick. I'll post all four versions here.

tisimst wrote:

John Ruggero wrote:Looking at it very casually, it seemed that beams have approx. a 45 degree limit in both directions, which presents an approx. 90 degree scope. Wouldn't every angle need to be taken in to account = 90 glyphs?

Regardless of how many you need, this would be very easy to generate using a simple script in FontForge, given max/min angles (e.g., +45/-45) and increment (e.g. 1). I'll give it a whirl and get back to you. The real challenge I think you'll deal with is actually USING it effectively.

This is the reason why I asked for someone to do the testing. The task of creating these glyphs is no match for FontLab either, but the actual usability of the font would depend on limiting it's scope as much as possible. 90 glyphs for tremolos alone seems rather impractical.

John Ruggero wrote:tisimst, will the sides of the slashes remain perpendicular to the staff lines for all degrees of tilt? These seems highly desirable to me, and I thought I saw this in your LilyPond example as well.

Keeping the sides vertical is no problem. The only (slight) challenge is that the script needs to compensate the thickness of the stroke to prevent the glyphs from becoming thinner as the slant increases. This is one of the main problems with the way Finale handles it's beams by default, and one of the great advantages of using the Patterson Beams plug-in.

Edit: You beat me by a few minutes, tisimst, and addressed my last point, so it doesn't seem to be an issue.

First, I don't believe that you need to create 45+45. Maybe 20+20 - it is to short to be visible if uneven; also it is hard to know what angle degree the beams are in Finale; also any re-spacing would affect the beam's angle.
Secondly, creating a font with these tremolos is a very odd way to achieve the goal. The software should do it. Not speaking of making parts from the score or any other edits which will affect angle of the beams (note spacing, as said above).

Ok. Anyone interested can get the files here. Below are two sets of files. One that goes -45 to +45 in 1 deg increments (so, 90 total shapes) and another that goes from -20 to +20 in 1 deg increments (so, 40 total shapes).

Each file contains the four previously mentioned "styles":
1. su (sheared, uncorrected thickness)
2. sc (sheared, corrected thickness)
3. ru (rectangular, uncorrected thickness)
4. rc (rectangular, corrected thickness)

Any other configuration can be generated on demand. Also, if the placement isn't quite right, let me know. Each glyph is placed such that the center of the left "face" is at the glyph's origin (0, 0). I didn't have a chance to cross-check the equivalent glyph in Maestro.

Molto fantastico, tisimst!!! The uncorrected 90 seems best to me, but the 40 is probably sufficient. More later.

Dear friends,
I haven't participated for a long time at this holly place, however I follow almost every day and most of the subjects you discuss.
With this post, I would like to share my humble experience related to this problem without any, even modest pretension that I am going to post now something distinctive. Not at all.
Slanted tremolos are definitely needed as an element in order one to prepare a good looking score.
Many of us use to generate custom fonts for such needs. So did I. The only issue is to find the balance with the real world and to decide how many elements are necessary. I put into practice just 3 attempts and the last one worked for me. The crucial thing in all of them was the width of the elements (no matter single, double or tripple tremolo symbol). Even the angle of the slant was less important. Finally I found that the width of the element for the tremolo should be equal to the width of the quarter note head. This was the meeting point where my aesthetically needs obtained complete satisfaction.
I prepared two sets of tremolos each with 10 different slants.
They are approximately + (plus) and – (minus) 1.7º, 3.5º, 5.2º, 6.9º, 8.7º, 10.3º, 12.0º, 13.7º, 15.3º and 16.9º
The second set is less usable due to its very steep slope: 18.6º, 21.6º, 24.5º, 27.3º, 30.0º, 32.5º, 35.0º, 37.2º, 39.4º, 41.3º.
Each instance has 6 derivates — single (left and right slanted), double (also left and right) and tripple (the same)
In this case I had to fill all an all 120 slots in a new font.
The biggest problem would be to find which one have to be selected when is required? Shell I try this one or that one?
Of course, not. The hesitation is my biggest enemy.
Therefore I prepared kind of ruler within the font with all different angles.
My philosophy reads: The precision does not contradict with the approximation.
Even more – we could not predict the software to generate beams with only 10 or some more fixed slants. But we could slightly manipulate the beam angles afterwards. The differences compare to the original are negligible.
To illustrate my idea, I will upload screen shots.
On one of them I overlapped the original cello part of the sample. The angles of the beams (as appeared) are as follow: 17.1º 15.2º 11.8º 13.0º.
There is no doubt, that if one of us must copy this exactly as it is, this will be time consuming work. Finale (Patterson beams plug-in) slanted all groups on ca. 8.7º and I find them more suitable for me.

Here are the samples.

The ruler is a glyph – also part of the font. In this situation its size (within Articulation tool) is 255p
I use it only for reference. How it works: if the slanted beam is touched with the appropriate "line" it refers to which slot in the font are the necessary tremolo elects. In this case there is no need to experiment. Even slightly different slanted tremolo (however close to the angle of the beam) doesn't appear improperly.

Best regards,
Wess

Wess, thank you for your very enlightening exposition of your method, which is ingenious and seems ideal to me given the circumstances, i.e. that Finale (and Sibelius) do not handle this automatically as it should, as pointed out by OCTO.

A question: why are the fractional divisions necessary given the approximate nature of the final operation with the ruler?

Also, I am not experienced with the Patterson plugin. Is it correct that the slashes should have a constant width, as should the beams at every angle?