aboutsummaryrefslogtreecommitdiffstats
path: root/Game/Code/lib/ffmpeg/rational.pas
diff options
context:
space:
mode:
Diffstat (limited to '')
-rw-r--r--Game/Code/lib/ffmpeg/rational.pas88
1 files changed, 88 insertions, 0 deletions
diff --git a/Game/Code/lib/ffmpeg/rational.pas b/Game/Code/lib/ffmpeg/rational.pas
new file mode 100644
index 00000000..a996e1df
--- /dev/null
+++ b/Game/Code/lib/ffmpeg/rational.pas
@@ -0,0 +1,88 @@
+(*
+ * Rational numbers
+ * Copyright (c) 2003 Michael Niedermayer <michaelni@gmx.at>
+ *
+ * This library is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2 of the License, or (at your option) any later version.
+ *
+ * This library is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with this library; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ *)
+unit rational;
+
+interface
+
+uses
+ windows;
+
+const
+ av__util = 'avutil-49.dll';
+
+type
+
+(*
+ * Rational number num/den. *)
+ PAVRational = ^TAVRational;
+ TAVRational = record
+ num: integer; ///< numerator
+ den: integer; ///< denominator
+ end;
+
+(**
+ * returns 0 if a==b, 1 if a>b and -1 if a<b.
+ *)
+function av_cmp_q(a: TAVRational; b: TAVRational): integer;
+
+(**
+ * converts the given AVRational to a double.
+ *)
+function av_q2d(a: TAVRational): double;
+
+(**
+ * reduce a fraction.
+ * this is usefull for framerate calculations
+ * @param max the maximum allowed for dst_nom & dst_den
+ * @return 1 if exact, 0 otherwise
+ *)
+function av_reduce(dst_nom: pinteger; dst_den: PInteger; nom: int64; den: int64; max: int64): integer;
+ cdecl; external av__util;
+
+function av_mul_q(b: TAVRational; c: TAVRational): TAVRational;
+ cdecl; external av__util;
+function av_div_q(b: TAVRational; c: TAVRational): TAVRational;
+ cdecl; external av__util;
+function av_add_q(b: TAVRational; c: TAVRational): TAVRational;
+ cdecl; external av__util;
+function av_sub_q(b: TAVRational; c: TAVRational): TAVRational;
+ cdecl; external av__util;
+function av_d2q(d: double; max: integer): TAVRational;
+ cdecl; external av__util;
+
+implementation
+
+function av_cmp_q (a: TAVRational; b: TAVRational): integer;
+var
+ tmp: int64;
+begin
+ tmp := a.num * b.den - b.num * a.den;
+
+ if tmp = 0 then
+ Result := (tmp shr 63) or 1
+ else
+ Result := 0;
+end;
+
+function av_q2d(a: TAVRational): double;
+begin
+ Result := a.num / a.den;
+end;
+
+end.