diff --git a/src/_imaging.c b/src/_imaging.c
index d1262f5b8..f023a6328 100644
--- a/src/_imaging.c
+++ b/src/_imaging.c
@@ -740,6 +740,7 @@ _prepare_lut_table(PyObject* table, Py_ssize_t table_size)
         } else {
             prepared[i] = table_data[i] * (255 << PRECISION_BITS) + 0.5;
         }
+        // printf("%f, %d ", table_data[i], prepared[i]);
     }
     
     #undef PRECISION_BITS
diff --git a/src/libImaging/ColorLUT.c b/src/libImaging/ColorLUT.c
index 6288625a1..dab9eb627 100644
--- a/src/libImaging/ColorLUT.c
+++ b/src/libImaging/ColorLUT.c
@@ -2,30 +2,104 @@
 #include <math.h>
 
 
-#define ROUND_UP(f) ((int) ((f) >= 0.0 ? (f) + 0.5F : (f) - 0.5F))
-
-
 /* 8 bits for result. Table can overflow [0, 1.0] range,
    so we need extra bits for overflow and negative values.
    NOTE: This value should be the same as in _imaging/_prepare_lut_table() */
 #define PRECISION_BITS (16 - 8 - 2)
+#define PRECISION_ROUNDING (1<<(PRECISION_BITS-1))
+
+
+UINT8 _lookups2[1024] = {
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+    0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15,
+    16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31,
+    32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47,
+    48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63,
+    64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79,
+    80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95,
+    96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111,
+    112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127,
+    128, 129, 130, 131, 132, 133, 134, 135, 136, 137, 138, 139, 140, 141, 142, 143,
+    144, 145, 146, 147, 148, 149, 150, 151, 152, 153, 154, 155, 156, 157, 158, 159,
+    160, 161, 162, 163, 164, 165, 166, 167, 168, 169, 170, 171, 172, 173, 174, 175,
+    176, 177, 178, 179, 180, 181, 182, 183, 184, 185, 186, 187, 188, 189, 190, 191,
+    192, 193, 194, 195, 196, 197, 198, 199, 200, 201, 202, 203, 204, 205, 206, 207,
+    208, 209, 210, 211, 212, 213, 214, 215, 216, 217, 218, 219, 220, 221, 222, 223,
+    224, 225, 226, 227, 228, 229, 230, 231, 232, 233, 234, 235, 236, 237, 238, 239,
+    240, 241, 242, 243, 244, 245, 246, 247, 248, 249, 250, 251, 252, 253, 254, 255,
+    255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,
+    255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,
+    255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,
+    255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,
+    255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,
+    255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,
+    255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,
+    255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,
+    255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,
+    255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,
+    255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,
+    255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,
+    255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,
+    255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,
+    255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,
+    255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255,
+};
+
+UINT8 *lookups2 = &_lookups2[512];
+
+
+static inline UINT8 clip8(int in)
+{
+    return lookups2[(in + PRECISION_ROUNDING) >> PRECISION_BITS];
+}
 
 
 static inline void
-interpolate3(INT16 out[3], const INT16 a[3], const INT16 b[3], float shift)
+interpolate3(INT16 out[3], const INT16 a[3], const INT16 b[3], INT16 shift)
 {
-    out[0] = a[0] * (1-shift) + b[0] * shift + 0.5;
-    out[1] = a[1] * (1-shift) + b[1] * shift + 0.5;
-    out[2] = a[2] * (1-shift) + b[2] * shift + 0.5;
+    out[0] = (a[0] * (0x1000-shift) + b[0] * shift + (1<<11)) >> 12;
+    out[1] = (a[1] * (0x1000-shift) + b[1] * shift + (1<<11)) >> 12;
+    out[2] = (a[2] * (0x1000-shift) + b[2] * shift + (1<<11)) >> 12;
 }
 
 static inline void
-interpolate4(INT16 out[3], const INT16 a[3], const INT16 b[3], float shift)
+interpolate4(INT16 out[4], const INT16 a[4], const INT16 b[4], INT16 shift)
 {
-    out[0] = a[0] * (1-shift) + b[0] * shift + 0.5;
-    out[1] = a[1] * (1-shift) + b[1] * shift + 0.5;
-    out[2] = a[2] * (1-shift) + b[2] * shift + 0.5;
-    out[3] = a[3] * (1-shift) + b[3] * shift + 0.5;
+    out[0] = (a[0] * (0x1000-shift) + b[0] * shift + (1<<11)) >> 12;
+    out[1] = (a[1] * (0x1000-shift) + b[1] * shift + (1<<11)) >> 12;
+    out[2] = (a[2] * (0x1000-shift) + b[2] * shift + (1<<11)) >> 12;
+    out[3] = (a[3] * (0x1000-shift) + b[3] * shift + (1<<11)) >> 12;
 }
 
 static inline int
@@ -94,19 +168,45 @@ ImagingColorLUT3D_linear(Imaging imOut, Imaging imIn, int table_channels,
     }
 
     for (y = 0; y < imOut->ysize; y++) {
-        UINT8 *rowIn = (UINT8 *)imIn->image[y];
-        UINT8 *rowOut = (UINT8 *)imIn->image[y];
-        for (x = 0; x < imOut->xsize; x++) {
-            float scaled1D = rowIn[x*4 + 0] * scale1D;
-            float scaled2D = rowIn[x*4 + 1] * scale2D;
-            float scaled3D = rowIn[x*4 + 2] * scale3D;
-            int index1D = (int) scaled1D;
-            int index2D = (int) scaled2D;
-            int index3D = (int) scaled3D;
-            float shift1D = scaled1D - index1D;
-            float shift2D = scaled2D - index2D;
-            float shift3D = scaled3D - index3D;
-            
+        UINT8* rowIn = (UINT8 *)imIn->image[y];
+        UINT8* rowOut = (UINT8 *)imOut->image[y];
+        if (table_channels == 3) {
+            for (x = 0; x < imOut->xsize; x++) {
+                float scaled1D = rowIn[x*4 + 0] * scale1D;
+                float scaled2D = rowIn[x*4 + 1] * scale2D;
+                float scaled3D = rowIn[x*4 + 2] * scale3D;
+                int index1D = (int) scaled1D;
+                int index2D = (int) scaled2D;
+                int index3D = (int) scaled3D;
+                INT16 shift1D = (scaled1D - index1D) * 0x1000 + 0.5;
+                INT16 shift2D = (scaled2D - index2D) * 0x1000 + 0.5;
+                INT16 shift3D = (scaled3D - index3D) * 0x1000 + 0.5;
+                int idx = table3D_index3(index1D, index2D, index3D, size1D, size1D_2D);
+                INT16 result[3], left[3], right[3];
+                INT16 leftleft[3], leftright[3], rightleft[3], rightright[3];
+
+                interpolate3(leftleft, &table[idx + 0], &table[idx + 3], shift1D);
+                interpolate3(leftright, &table[idx + size1D*3],
+                             &table[idx + size1D*3 + 3], shift1D);
+                interpolate3(left, leftleft, leftright, shift2D);
+
+                interpolate3(rightleft, &table[idx + size1D_2D*3],
+                             &table[idx + size1D_2D*3 + 3], shift1D);
+                interpolate3(rightright, &table[idx + size1D_2D*3 + size1D*3],
+                             &table[idx + size1D_2D*3 + size1D*3 + 3], shift1D);
+                interpolate3(right, rightleft, rightright, shift2D);
+
+                interpolate3(result, left, right, shift3D);
+
+                rowOut[x*4 + 0] = clip8(result[0]);
+                rowOut[x*4 + 1] = clip8(result[1]);
+                rowOut[x*4 + 2] = clip8(result[2]);
+                rowOut[x*4 + 3] = rowIn[x*4 + 3];
+
+                // printf("%d: %f -> %d, %f; idx: %d; (%d, %d, %d)\n",
+                //     rowIn[x*4 + 0], scaled1D, index1D, shift1D,
+                //     idx, nearest[0], nearest[1], nearest[2]);
+            }
         }
     }