[Open2x-svn-logs] SF.net SVN: open2x:[324] trunk/libs/SDL-1.2.13/src/video/gp2x/gp2x_tslib.c

[<sen...@us...>]

Revision: 324
          http://open2x.svn.sourceforge.net/open2x/?rev=324&view=rev
Author:   senquack
Date:     2008-08-18 19:38:19 +0000 (Mon, 18 Aug 2008)

Log Message:
-----------


Modified Paths:
--------------
    trunk/libs/SDL-1.2.13/src/video/gp2x/gp2x_tslib.c

Modified: trunk/libs/SDL-1.2.13/src/video/gp2x/gp2x_tslib.c
===================================================================
--- trunk/libs/SDL-1.2.13/src/video/gp2x/gp2x_tslib.c	2008-08-13 00:01:07 UTC (rev 323)
+++ trunk/libs/SDL-1.2.13/src/video/gp2x/gp2x_tslib.c	2008-08-18 19:38:19 UTC (rev 324)
@@ -722,6 +722,206 @@
 	int y;
 } XYHIST;
 
+#define XYHISTLENGTH 15
+XYHIST average_xyhist[XYHISTLENGTH];
+
+static inline int average_read(TSDEV *ts, TS_SAMPLE *samp, int nr)
+{
+
+	int ret = 0;
+	int newx, newy;
+
+	int flush_history = 0; // variance module will tell us if we are getting fast movements
+	static int flush_count = 0; // when we get enough fast movements, flush the history
+	static int xyhist_full = 0; //DKS - when we have a full four samples, this is 1 
+	static int xyhist_counter = 0; // DKS - how many entires in history?
+	static int xyhist_index = 0;	//xyhist_index - 1 is index to most-recent sample in array
+
+	ret = variance_read(ts, samp, nr, &flush_history);
+	flush_count += flush_history;
+	struct ts_sample *s;
+	s = samp + (ret - 1); // this module can handle more than one sample, but the gp2x
+								// touchscreen driver only ever provides 1 sample each read so
+								// it's not really necessary but not really inefficient either way
+	int i;
+
+	for( i = ret; i > 0; i--, s--) {
+		if ((s->pressure == 0) || (flush_count > 7)) {
+			// when pen is lifted, or a fast movement is detected, flush stored history
+			xyhist_full = 0;
+			xyhist_counter = 0;
+			xyhist_index = 0;
+			flush_count = 0;
+		} else {
+			average_xyhist[xyhist_index].x = s->x;
+			average_xyhist[xyhist_index].y = s->y;
+			
+			xyhist_counter++;
+			if (xyhist_counter == XYHISTLENGTH) {
+				xyhist_full = 1;
+			}
+
+			xyhist_index++;
+			if (xyhist_index == XYHISTLENGTH) {
+				xyhist_index = 0; // point back at beginning since array is full
+			}
+
+			if (xyhist_full) {
+				//we have a full sample history, we can average this sample with the others
+
+				int j = xyhist_index - 1; // point j to most recent entry in history
+				if (j < 0) {
+					j = XYHISTLENGTH - 1;
+				}
+
+				// 15-sample weighted average, pyramid, provides great smoothing for precise
+				// operations like menus and drawing
+
+				// sample 1 has weight of * 1
+				newx = average_xyhist[j].x;
+				newy = average_xyhist[j].y;
+
+				j--;
+				if (j < 0) {
+					j = XYHISTLENGTH - 1;
+				}		
+				
+				// sample 2 has weight of * 2
+				newx += average_xyhist[j].x << 1;
+				newy += average_xyhist[j].y << 1;
+
+				j--;
+				if (j < 0) {
+					j = XYHISTLENGTH - 1;
+				}		
+
+				// sample 3 has weight of * 3
+				newx += (average_xyhist[j].x << 1) + average_xyhist[j].x;
+				newy += (average_xyhist[j].y << 1) + average_xyhist[j].y;
+
+				j--;
+				if (j < 0) {
+					j = XYHISTLENGTH - 1;
+				}		
+
+				// sample 4 has weight of * 4
+				newx += average_xyhist[j].x << 2;
+				newy += average_xyhist[j].y << 2;
+
+				j--;
+				if (j < 0) {
+					j = XYHISTLENGTH - 1;
+				}		
+
+				// sample 5 has weight of * 5
+				newx += (average_xyhist[j].x << 2) + average_xyhist[j].x;
+				newy += (average_xyhist[j].y << 2) + average_xyhist[j].y;
+
+				j--;
+				if (j < 0) {
+					j = XYHISTLENGTH - 1;
+				}		
+
+				// sample 6 has weight of * 6
+				newx += (average_xyhist[j].x << 2) + (average_xyhist[j].x << 1);
+				newy += (average_xyhist[j].y << 2) + (average_xyhist[j].y << 1);
+
+				j--;
+				if (j < 0) {
+					j = XYHISTLENGTH - 1;
+				}		
+
+				// sample 7 has weight of * 7
+				newx += (average_xyhist[j].x << 3) - average_xyhist[j].x;
+				newy += (average_xyhist[j].y << 3) - average_xyhist[j].y;
+
+				j--;
+				if (j < 0) {
+					j = XYHISTLENGTH - 1;
+				}		
+
+				// sample 8, middle sample,  has weight of * 8
+				newx += (average_xyhist[j].x << 3);
+				newy += (average_xyhist[j].y << 3);
+
+				j--;
+				if (j < 0) {
+					j = XYHISTLENGTH - 1;
+				}		
+
+				// sample 9 has weight of * 7
+				newx += (average_xyhist[j].x << 3) - average_xyhist[j].x;
+				newy += (average_xyhist[j].y << 3) - average_xyhist[j].y;
+
+				j--;
+				if (j < 0) {
+					j = XYHISTLENGTH - 1;
+				}		
+
+				// sample 10 has weight of * 6
+				newx += (average_xyhist[j].x << 2) + (average_xyhist[j].x << 1);
+				newy += (average_xyhist[j].y << 2) + (average_xyhist[j].y << 1);
+
+				j--;
+				if (j < 0) {
+					j = XYHISTLENGTH - 1;
+				}		
+
+				// sample 11 has weight of * 5
+				newx += (average_xyhist[j].x << 2) + average_xyhist[j].x;
+				newy += (average_xyhist[j].y << 2) + average_xyhist[j].y;
+
+				j--;
+				if (j < 0) {
+					j = XYHISTLENGTH - 1;
+				}		
+
+				// sample 12 has weight of * 4
+				newx += average_xyhist[j].x << 2;
+				newy += average_xyhist[j].y << 2;
+
+				j--;
+				if (j < 0) {
+					j = XYHISTLENGTH - 1;
+				}		
+			
+				// sample 13 has weight of * 3
+				newx += (average_xyhist[j].x << 1) + average_xyhist[j].x;
+				newy += (average_xyhist[j].y << 1) + average_xyhist[j].y;
+
+				j--;
+				if (j < 0) {
+					j = XYHISTLENGTH - 1;
+				}		
+
+				// sample 14 has weight of * 2
+				newx += average_xyhist[j].x << 1;
+				newy += average_xyhist[j].y << 1;
+
+				j--;
+				if (j < 0) {
+					j = XYHISTLENGTH - 1;
+				}		
+
+				// sample 15 has weight of * 1
+				newx += average_xyhist[j].x;
+				newy += average_xyhist[j].y;
+
+				j--;
+				if (j < 0) {
+					j = XYHISTLENGTH - 1;
+				}		
+				
+				// divide results by 64 to provide average
+				samp->x = newx >> 6;
+				samp->y = newy >> 6;
+			}
+		}
+	}
+
+  return ret;
+}
+//DKS - non-pyramid form of the above, I cannot tell a real difference, will use pyramid for now.
 //#define XYHISTLENGTH 15
 //XYHIST average_xyhist[XYHISTLENGTH];
 //
@@ -777,70 +977,69 @@
 //				// 15-sample weighted average, pyramid, provides great smoothing for precise
 //				// operations like menus and drawing
 //
-//				// sample 1 has weight of * 1
-//				newx = average_xyhist[j].x;
-//				newy = average_xyhist[j].y;
+//				// sample 1 has weight of * 15
+//				newx = average_xyhist[j].x * 15;
+//				newy = average_xyhist[j].y * 15;
 //
 //				j--;
 //				if (j < 0) {
 //					j = XYHISTLENGTH - 1;
 //				}		
 //				
-//				// sample 2 has weight of * 2
-//				newx += average_xyhist[j].x << 1;
-//				newy += average_xyhist[j].y << 1;
+//				// sample 2 has weight of * 14
+//				newx += average_xyhist[j].x * 14;
+//				newy += average_xyhist[j].y * 14;
 //
 //				j--;
 //				if (j < 0) {
 //					j = XYHISTLENGTH - 1;
 //				}		
 //
-//				// sample 3 has weight of * 3
-//				newx += (average_xyhist[j].x << 1) + average_xyhist[j].x;
-//				newy += (average_xyhist[j].y << 1) + average_xyhist[j].y;
-//
+//				// sample 3 has weight of * 13
+//				newx += average_xyhist[j].x * 13;
+//				newy += average_xyhist[j].y * 13;
 //				j--;
 //				if (j < 0) {
 //					j = XYHISTLENGTH - 1;
 //				}		
 //
-//				// sample 4 has weight of * 4
-//				newx += average_xyhist[j].x << 2;
-//				newy += average_xyhist[j].y << 2;
+//				// sample 4 has weight of * 12
+//				newx += average_xyhist[j].x * 12;
+//				newy += average_xyhist[j].y * 12;
 //
 //				j--;
 //				if (j < 0) {
 //					j = XYHISTLENGTH - 1;
 //				}		
 //
-//				// sample 5 has weight of * 5
-//				newx += (average_xyhist[j].x << 2) + average_xyhist[j].x;
-//				newy += (average_xyhist[j].y << 2) + average_xyhist[j].y;
+//				// sample 5 has weight of * 11 
+//				newx += average_xyhist[j].x * 11;
+//				newy += average_xyhist[j].y * 11;
 //
 //				j--;
 //				if (j < 0) {
 //					j = XYHISTLENGTH - 1;
 //				}		
 //
-//				// sample 6 has weight of * 6
-//				newx += (average_xyhist[j].x << 2) + (average_xyhist[j].x << 1);
-//				newy += (average_xyhist[j].y << 2) + (average_xyhist[j].y << 1);
+//				// sample 6 has weight of * 10
+//				newx += average_xyhist[j].x * 10;
+//				newy += average_xyhist[j].y * 10;
 //
 //				j--;
 //				if (j < 0) {
 //					j = XYHISTLENGTH - 1;
 //				}		
 //
-//				// sample 7 has weight of * 7
-//				newx += (average_xyhist[j].x << 3) - average_xyhist[j].x;
-//				newy += (average_xyhist[j].y << 3) - average_xyhist[j].y;
+//				// sample 7 has weight of * 9
+//				newx += average_xyhist[j].x * 9;
+//				newy += average_xyhist[j].y * 9;
 //
 //				j--;
 //				if (j < 0) {
 //					j = XYHISTLENGTH - 1;
 //				}		
 //
-//				// sample 8, middle sample,  has weight of * 8
+//				// sample 8, middle sample,  has weight of * 8 
 //				newx += (average_xyhist[j].x << 3);
 //				newy += (average_xyhist[j].y << 3);
 //
@@ -912,213 +1111,15 @@
 //					j = XYHISTLENGTH - 1;
 //				}		
 //				
-//				// divide results by 64 to provide average
-//				samp->x = newx >> 6;
-//				samp->y = newy >> 6;
+//				samp->x = newx / 120;
+//				samp->y = newy / 120;
 //			}
 //		}
 //	}
 //
 //  return ret;
 //}
-#define XYHISTLENGTH 5
-XYHIST average_xyhist[XYHISTLENGTH];
 
-static inline int average_read(TSDEV *ts, TS_SAMPLE *samp, int nr)
-{
-
-	int ret = 0;
-	int newx, newy;
-
-	int flush_history = 0; // variance module will tell us if we are getting fast movements
-	static int flush_count = 0; // when we get enough fast movements, flush the history
-	static int xyhist_full = 0; //DKS - when we have a full four samples, this is 1 
-	static int xyhist_counter = 0; // DKS - how many entires in history?
-	static int xyhist_index = 0;	//xyhist_index - 1 is index to most-recent sample in array
-
-	ret = variance_read(ts, samp, nr, &flush_history);
-	flush_count += flush_history;
-	struct ts_sample *s;
-	s = samp + (ret - 1); // this module can handle more than one sample, but the gp2x
-								// touchscreen driver only ever provides 1 sample each read so
-								// it's not really necessary but not really inefficient either way
-	int i;
-
-	for( i = ret; i > 0; i--, s--) {
-		if ((s->pressure == 0) || (flush_count > 7)) {
-			// when pen is lifted, or a fast movement is detected, flush stored history
-			xyhist_full = 0;
-			xyhist_counter = 0;
-			xyhist_index = 0;
-			flush_count = 0;
-		} else {
-			average_xyhist[xyhist_index].x = s->x;
-			average_xyhist[xyhist_index].y = s->y;
-			
-			xyhist_counter++;
-			if (xyhist_counter == XYHISTLENGTH) {
-				xyhist_full = 1;
-			}
-
-			xyhist_index++;
-			if (xyhist_index == XYHISTLENGTH) {
-				xyhist_index = 0; // point back at beginning since array is full
-			}
-
-			if (xyhist_full) {
-				//we have a full sample history, we can average this sample with the others
-
-				int j = xyhist_index - 1; // point j to most recent entry in history
-				if (j < 0) {
-					j = XYHISTLENGTH - 1;
-				}
-
-				// 15-sample weighted average, pyramid, provides great smoothing for precise
-				// operations like menus and drawing
-
-				// sample 1 has weight of * 15
-				newx = average_xyhist[j].x * 15;
-				newy = average_xyhist[j].y * 15;
-
-				j--;
-				if (j < 0) {
-					j = XYHISTLENGTH - 1;
-				}		
-				
-				// sample 2 has weight of * 14
-				newx += average_xyhist[j].x * 14;
-				newy += average_xyhist[j].y * 14;
-
-				j--;
-				if (j < 0) {
-					j = XYHISTLENGTH - 1;
-				}		
-
-				// sample 3 has weight of * 13
-				newx += average_xyhist[j].x * 13;
-				newy += average_xyhist[j].y * 13;
-				j--;
-				if (j < 0) {
-					j = XYHISTLENGTH - 1;
-				}		
-
-				// sample 4 has weight of * 12
-				newx += average_xyhist[j].x * 12;
-				newy += average_xyhist[j].y * 12;
-
-				j--;
-				if (j < 0) {
-					j = XYHISTLENGTH - 1;
-				}		
-
-				// sample 5 has weight of * 11 
-				newx += average_xyhist[j].x * 11;
-				newy += average_xyhist[j].y * 11;
-
-				j--;
-				if (j < 0) {
-					j = XYHISTLENGTH - 1;
-				}		
-
-				// sample 6 has weight of * 10
-				newx += average_xyhist[j].x * 10;
-				newy += average_xyhist[j].y * 10;
-
-				j--;
-				if (j < 0) {
-					j = XYHISTLENGTH - 1;
-				}		
-
-				// sample 7 has weight of * 9
-				newx += average_xyhist[j].x * 9;
-				newy += average_xyhist[j].y * 9;
-
-				j--;
-				if (j < 0) {
-					j = XYHISTLENGTH - 1;
-				}		
-
-				// sample 8, middle sample,  has weight of * 8 
-				newx += (average_xyhist[j].x << 3);
-				newy += (average_xyhist[j].y << 3);
-
-				j--;
-				if (j < 0) {
-					j = XYHISTLENGTH - 1;
-				}		
-
-				// sample 9 has weight of * 7
-				newx += (average_xyhist[j].x << 3) - average_xyhist[j].x;
-				newy += (average_xyhist[j].y << 3) - average_xyhist[j].y;
-
-				j--;
-				if (j < 0) {
-					j = XYHISTLENGTH - 1;
-				}		
-
-				// sample 10 has weight of * 6
-				newx += (average_xyhist[j].x << 2) + (average_xyhist[j].x << 1);
-				newy += (average_xyhist[j].y << 2) + (average_xyhist[j].y << 1);
-
-				j--;
-				if (j < 0) {
-					j = XYHISTLENGTH - 1;
-				}		
-
-				// sample 11 has weight of * 5
-				newx += (average_xyhist[j].x << 2) + average_xyhist[j].x;
-				newy += (average_xyhist[j].y << 2) + average_xyhist[j].y;
-
-				j--;
-				if (j < 0) {
-					j = XYHISTLENGTH - 1;
-				}		
-
-				// sample 12 has weight of * 4
-				newx += average_xyhist[j].x << 2;
-				newy += average_xyhist[j].y << 2;
-
-				j--;
-				if (j < 0) {
-					j = XYHISTLENGTH - 1;
-				}		
-			
-				// sample 13 has weight of * 3
-				newx += (average_xyhist[j].x << 1) + average_xyhist[j].x;
-				newy += (average_xyhist[j].y << 1) + average_xyhist[j].y;
-
-				j--;
-				if (j < 0) {
-					j = XYHISTLENGTH - 1;
-				}		
-
-				// sample 14 has weight of * 2
-				newx += average_xyhist[j].x << 1;
-				newy += average_xyhist[j].y << 1;
-
-				j--;
-				if (j < 0) {
-					j = XYHISTLENGTH - 1;
-				}		
-
-				// sample 15 has weight of * 1
-				newx += average_xyhist[j].x;
-				newy += average_xyhist[j].y;
-
-				j--;
-				if (j < 0) {
-					j = XYHISTLENGTH - 1;
-				}		
-				
-				samp->x = newx / 120;
-				samp->y = newy / 120;
-			}
-		}
-	}
-
-  return ret;
-}
-
 /******************************
  * linear
  ******************************/


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.