diff --git a/test/vector.cpp b/test/vector.cpp
index 1931c2fb..66984cad 100644
--- a/test/vector.cpp
+++ b/test/vector.cpp
@@ -61,12 +61,52 @@ test_polar (util::TAP::logger &tap)
 }
 
 
+void
+test_euler (util::TAP::logger &tap)
+{
+    static const struct {
+        util::vector3f dir;
+        util::vector2f euler;
+        const char *name;
+    } TESTS[] = {
+        // y-axis
+        { {  0,  0, -1 }, { 0.5f,  0.5f }, "forward" },
+        { { -1,  0,  0 }, { 0.5f, -1.0f }, "left" },
+        { {  0,  0,  1 }, { 0.5f, -0.5f }, "back" },
+        { {  1,  0,  0 }, { 0.5f,  0.0f }, "right" },
+
+        // x-axis
+        { {  0,  1,  0 }, { 0, 0 }, "up"   },
+        { {  0, -1,  0 }, { 1, 0 }, "down" },
+    };
+
+    // check that simple axis rotations look correct
+    for (auto i: TESTS) {
+        tap.expect_eq (util::to_euler (i.dir),
+                       i.euler * PI<float>,
+                       "to euler, %s", i.name);
+    }
+
+    // check error in round trip through euler angles
+    for (auto i: TESTS) {
+        auto trip = util::from_euler (util::to_euler (i.dir));
+        auto diff = i.dir - trip;
+        auto norm = diff.magnitude ();
+
+        // trig functions reduce precision above almost_equal levels, so we
+        // hard code a fairly low bound here instead.
+        tap.expect_lt (norm, 1e-7, "euler round-trip error, %s", i.name);
+    }
+}
+
+
 int
 main ()
 {
     util::TAP::logger tap;
 
     test_polar (tap);
+    test_euler (tap);
 
     tap.expect (!util::vector3f::ZERO.is_normalised (), "zero isn't normalised");
     tap.expect (!util::vector3f::UNIT.is_normalised (), "unit is normalised");
diff --git a/vector.cpp b/vector.cpp
index 2efe07fc..a9d004ba 100644
--- a/vector.cpp
+++ b/vector.cpp
@@ -136,6 +136,39 @@ util::cartesian_to_polar (vector<2,T> v)
 }
 
 
+///////////////////////////////////////////////////////////////////////////////
+template <typename T>
+vector<3,T>
+util::from_euler (vector<2,T> euler)
+{
+    return {
+         std::sin (euler.x) * std::cos (euler.y),
+         std::cos (euler.x),
+        -std::sin (euler.x) * std::sin (euler.y),
+    };
+}
+
+template util::vector3f util::from_euler (util::vector2f);
+template util::vector3d util::from_euler (util::vector2d);
+
+
+//-----------------------------------------------------------------------------
+template <typename T>
+vector<2,T>
+util::to_euler (vector<3,T> vec)
+{
+    vec.normalise ();
+
+    return {
+        std::acos  (vec.y / vec.magnitude ()),
+        -std::atan2 (vec.z, vec.x),
+    };
+}
+
+template util::vector2f util::to_euler (util::vector3f);
+template util::vector2d util::to_euler (util::vector3d);
+
+
 ///////////////////////////////////////////////////////////////////////////////
 template <typename T>
 vector<3,T>
diff --git a/vector.hpp b/vector.hpp
index ce194f14..266905aa 100644
--- a/vector.hpp
+++ b/vector.hpp
@@ -63,6 +63,9 @@ namespace util {
     template <typename T> vector<3,T> spherical_to_cartesian (vector<3,T>);
     template <typename T> vector<3,T> cartesian_to_spherical (vector<3,T>);
 
+    template <typename T> vector<2,T> to_euler   (vector<3,T>);
+    template <typename T> vector<3,T> from_euler (vector<2,T>);
+
     // output and serialisation operators
     template <size_t S, typename T> std::ostream& operator<< (std::ostream&, vector<S,T>);