release: 1.3.29

[lilypond.git] / scm / paper.scm

diff --git a/scm/paper.scm b/scm/paper.scm
index d4eb2d352093b73dae8e802b21fc6820b907ce3f..9a6b05a8c02f2a700f22a096158ee5ebd3a13e9f 100644 (file)
--- a/scm/paper.scm
+++ b/scm/paper.scm
@@ -2,16 +2,10 @@
 ;;;
 ;;;  source file of the GNU LilyPond music typesetter
 ;;; 
-;;; (c) 1999 Jan Nieuwenhuizen <janneke@gnu.org>
+;;; (c) 1999--2000 Jan Nieuwenhuizen <janneke@gnu.org>
 
 ;;; All dimensions are measured in staff-spaces
 
-;; TODO
-;;  - make easily customisable from mudela
-;;  - take #forced stems into account (now done in C++)
-;;  - take y-position of chord or beam into account
-(define (stem-shorten flags) 0.5)
-(define (beamed-stem-shorten multiplicity) 0.5)
 
 
 ;; Beams should be prevented to conflict with the stafflines, 
@@ -27,39 +21,116 @@
 
 ;; inter seems to be a modern quirk, we don't use that
 
+(define staff-line 0.10)
+(define beam-thickness (* 0.52 (- 1 staff-line)))
+(define beam-straddle 0)
+(define beam-sit (/ (+ beam-thickness staff-line) 2))
+(define beam-hang (- 1 (/ (- beam-thickness staff-line) 2)))
+
+;; Note: quanting period is take as quants.top () - quants[0], 
+;; which should be 1 (== 1 interline)
+
 (define beam-normal-dy-quants
-  '(0 (/2 (+ beam-thickness staff-line) 2) (+ beam-thickness staff-line) 1))
+  (list 0 (/ (+ beam-thickness staff-line) 2) (+ beam-thickness staff-line) 1))
 
 ;; two popular veritcal beam quantings
 ;; see params.ly: #'beam-vertical-quants
 (define (beam-normal-y-quants multiplicity dy)
-  (let ((quants `(,beam-hang 1)))
+  (let ((quants (list beam-hang)))
     (if (or (<= multiplicity 1) (>= (abs dy) (/ staff-line 2)))
        (set! quants (cons beam-sit quants)))
     (if (or (<= multiplicity 2) (>= (abs dy) (/ staff-line 2)))
        (set! quants (cons beam-straddle quants)))
-    quants))
+    ;; period: 1 (interline)
+    (append quants (list (+ 1 (car quants))))))
 
 (define (beam-traditional-y-quants multiplicity dy)
-  (let ((quants '(1)))
+  (let ((quants '()))
     (if (>= dy (/ staff-line -2))
        (set! quants (cons beam-hang quants)))
     (if (and (<= multiplicity 1) (<= dy (/ staff-line 2)))
        (set! quants (cons beam-sit quants)))
     (if (or (<= multiplicity 2) (>= (abs dy) (/ staff-line 2)))
        (set! quants (cons beam-straddle quants)))
-    quants))
+    ;; period: 1 (interline)
+    (append quants (list (+ 1 (car quants))))))
 
 
-;;; Default variables and settings
+;; There are several ways to calculate the direction of a beam
+;;
+;; * majority: number count of up or down notes
+;; * mean    : mean centre distance of all notes
+;; * median  : mean centre distance weighted per note
 
-(define staff-line 0.10)
-(define beam-thickness (* 0.52 (- 1 staff-line)))
-(define beam-straddle 0)
-(define beam-sit (/ (+ beam-thickness staff-line) 2))
-(define beam-hang (- 1 (/ (- beam-thickness staff-line) 2)))
+(define (dir-compare up down)
+  (if (= up down)
+      0
+      (if (> up down)
+         1
+         -1)))
+
+;; (up . down)
+(define (beam-dir-majority count total)
+  (dir-compare (car count) (cdr count)))
+
+(define (beam-dir-mean count total)
+  (dir-compare (car total) (cdr total)))
+
+(define (beam-dir-median count total)
+  (if (and (> (car count) 0)
+          (> (cdr count) 0))
+      (dir-compare (/ (car total) (car count)) (/ (cdr total) (cdr count)))
+      (dir-compare (car count) (cdr count))))
+           
+
+;;; Default variables and settings
 
 (define beam-height-quants beam-normal-dy-quants)
 (define beam-vertical-position-quants beam-normal-y-quants)
 
 
+;; [Ross] states that the majority of the notes dictates the
+;; direction (and not the mean of "center distance")
+;;
+;; But is that because it really looks better, or because he wants
+;; to provide some real simple hands-on rules?
+;;     
+;; We have our doubts, so we simply provide all sensible alternatives.
+(define beam-dir-algorithm beam-dir-majority)
+
+
+;; array index flag-2 (what a name!!), last if index>size
+;; unbeamed stems
+(define stem-length '(3.5 3.5 3.5 4.5 5.0))
+(define grace-length-factor 0.8)
+(define grace-stem-length
+  (map (lambda (x) (* grace-length-factor x)) stem-length))
+
+;; array index multiplicity, last if index>size
+;; beamed stems
+(define beamed-stem-shorten '(0.5))
+(define beamed-stem-length '(0.0 2.5 2.0 1.5))
+(define grace-beamed-stem-length '(0.0 2.5 2.0 1.5))
+(define beamed-stem-minimum-length '(0.0 1.5 1.25 1.0))
+(define grace-beamed-stem-minimum-length
+  (map (lambda (x) (* grace-length-factor x)) beamed-stem-minimum-length))
+
+;;  Stems in unnatural (forced) direction should be shortened,
+;;  according to [Roush & Gourlay].  Their suggestion to knock off
+;;  a whole staffspace seems a bit drastical: we'll do half.
+
+;; TODO
+;;  - take #forced stems into account (now done in C++)?
+;;  - take y-position of chord or beam into account
+
+(define stem-shorten '(0.5))
+(define grace-stem-shorten '(0.0))
+
+;; urg
+(define pi (* 2 (acos 0)))
+
+(define (slur-default-height h-inf r-0 b)
+  (let ((alpha (/ (* 2.0 h-inf) pi))
+       (beta (/ (* pi r-0) (* 2.0 h-inf))))
+    (* alpha (atan (* beta b)))))
+