Allow interpolation of matrices with negative scale / axis flips

The matrix interpolation function `interp_m3_m3m3()` decomposes the
matrices into rotation and scale matrices, converts the rotation
matrices to quaternions, SLERPs the quaternions, and converts the result
back to a matrix. Since quaternions cannot represent axis flips, this
results in interpolation problems like described in T77154.

Our interpolation function is based on "Matrix Animation and Polar
Decomposition", by Ken Shoemake & Tom Duff. The paper states that it
produces invalid results when there is an axis flip in the rotation
matrix (or negative determinant, or negative scale, those all indicate
the same thing). Their solution is to multiply the rotation matrix with
`-I`, where `I` is the identity matrix. This is the same as element-wise
multiplication with `-1.0f`. My proposed solution is to not only do that
with the rotation matrix `R`, but also with the scale matrix `S`. This
ensures that the decomposition of `A = R * S` remains valid, while also
making it possible to conver the rotation component to a quaternion.

There is still an issue when interpolating between matrices with
different determinant. As the determinant represents the change in
volume when that matrix is applied to an object, interpolating between a
negative and a positive matrix will have to go through a zero
determinant. In this case the volume collapses to zero. I don't see this
as a big issue, though, as without this patch Blender would also produce
invalid results anyway.

Reviewed By: brecht, sergey

Differential Revision: https://developer.blender.org/D8048

1 files changed, 31 insertions, 9 deletions

diff --git a/tests/gtests/blenlib/BLI_math_matrix_test.cc b/tests/gtests/blenlib/BLI_math_matrix_test.cc
index 0baccf9ee60..9c47c02ceaf 100644
--- a/tests/gtests/blenlib/BLI_math_matrix_test.cc
+++ b/tests/gtests/blenlib/BLI_math_matrix_test.cc
@@ -62,16 +62,38 @@ TEST(math_matrix, interp_m3_m3m3_singularity)
   transpose_m3(matrix_a);
   EXPECT_NEAR(-1.0f, determinant_m3_array(matrix_a), 1e-6);
 
-  float matrix_i[3][3];
-  unit_m3(matrix_i);
+  /* This matrix represents R=(0, 0, 0), S=(-1, 0, 0) */
+  float matrix_b[3][3] = {
+      {-1.0f, 0.0f, 0.0f},
+      {0.0f, 1.0f, 0.0f},
+      {0.0f, 0.0f, 1.0f},
+  };
+  transpose_m3(matrix_b);
 
   float result[3][3];
-  const float epsilon = 1e-6;
-  interp_m3_m3m3(result, matrix_i, matrix_a, 0.0f);
-  EXPECT_M3_NEAR(result, matrix_i, epsilon);
+  interp_m3_m3m3(result, matrix_a, matrix_b, 0.0f);
+  EXPECT_M3_NEAR(result, matrix_a, 1e-5);
+
+  interp_m3_m3m3(result, matrix_a, matrix_b, 1.0f);
+  EXPECT_M3_NEAR(result, matrix_b, 1e-5);
 
-  /* This fails for matrices with a negative determinant, i.e. with an axis mirror in the rotation
-   * component. See T77154. */
-  // interp_m3_m3m3(result, matrix_i, matrix_a, 1.0f);
-  // EXPECT_M3_NEAR(result, matrix_a, epsilon);
+  interp_m3_m3m3(result, matrix_a, matrix_b, 0.5f);
+  float expect[3][3] = {
+      {-0.997681f, -0.049995f, 0.046186f},
+      {-0.051473f, 0.998181f, -0.031385f},
+      {0.044533f, 0.033689f, 0.998440f},
+  };
+  transpose_m3(expect);
+  EXPECT_M3_NEAR(result, expect, 1e-5);
+
+  /* Interpolating between a matrix with and without axis flip can cause it to go through a zero
+   * point. The determinant det(A) of a matrix represents the change in volume; interpolating
+   * between matrices with det(A)=-1 and det(B)=1 will have to go through a point where
+   * det(result)=0, so where the volume becomes zero. */
+  float matrix_i[3][3];
+  unit_m3(matrix_i);
+  zero_m3(expect);
+  interp_m3_m3m3(result, matrix_a, matrix_i, 0.5f);
+  EXPECT_NEAR(0.0f, determinant_m3_array(result), 1e-5);
+  EXPECT_M3_NEAR(result, expect, 1e-5);
 }