clearer comments about the Lanczos2D scaling, which at this point I believe is the best possible one for small ellipses
diff --git a/magick/resize.c b/magick/resize.c
index 7ff2999..8577395 100644
--- a/magick/resize.c
+++ b/magick/resize.c
@@ -888,30 +888,29 @@
resize_filter->support = (MagickRealType) MagickSQ2; /* which times blur => 2.0 */
break;
case Lanczos2DFilter:
- /* Special 2 lobed cylindrical Jinc-Jinc filter,
- * with a special blur adjustment to remove the blurring
- * effect of the windowing of the Jinc function (in the 2
- * lobed case only). To be used as the default filter for EWA
+ /* Special 2-lobed cylindrical Jinc-Jinc filter with a special
+ * "blur" adjustment (actually, a shortening of the
+ * support). To be used as the default filter for EWA
* Resampling and Distorts.
*
* Derivation: Set the scaling s=1/blur of the Lanczos2D
* filter function so that
- * Lanczos2D(s)=-2*Lanczos2D(s*sqrt(2))-Lanczos2D(s*2)
- * which implies that with a no-op a single vertical line is
- * not amplified (although it has negative ripples), and also
- * so that the high frequency checkerboard mode, as well as
- * the high frequency vertical, and horizontal, stripe modes
- * are slightly damped.
+ * Lanczos2D(s)=-2*Lanczos2D(s*sqrt(2))-Lanczos2D(s*2).
*
- * (The value which preserves the high frequency vertical, and
- * horizontal, stripe modes (and slightly dampens the
- * checkerboard mode) satisfies
- * Lanczos2D(s)=-2*Lanczos2D(s*sqrt(2))
- * which gives 0.9549921738 instead of 0.958033808.)
+ * This choice ensures that the height of single vertical line
+ * is exactly preserved when "no-op" (the "identity
+ * geometrical transformation) is applied to the image. It
+ * also ensures that, with no-op, the nearest two columns (one
+ * on the left and one on the right) are not changed (they are
+ * kept to zero). The only side-effect of no-op on a single
+ * vertical line is a "ripple" in the two columns located two
+ * pixel away. The size of this ripple, for a unit column
+ * (constant 1), is -5.86142e-4 (essentially non-existant).
+ * In addition, with this choice of "r-scaling," the high
+ * frequency checkerboard mode, as well as the high frequency
+ * vertical and horizontal stripe modes are slightly damped.
*
- * Note from Nicolas: It is still not totally clear what
- * scaling of the Lanczos2D kernel is optimal for Clamped-EWA
- * upsampling.
+ * Derived by Nicolas Robidoux of Laurentian University.
*/
resize_filter->blur *= (MagickRealType) 0.958033808;
default: