diff --git a/gfx/filter.c b/gfx/filter.c
index 631e0b5307..878eb322d0 100644
--- a/gfx/filter.c
+++ b/gfx/filter.c
@@ -97,6 +97,8 @@ static softfilter_get_implementation_t softfilter_get_implementation_from_idx(un
          return darken_get_implementation;
       case SOFTFILTER_TWOXSAI:
          return twoxsai_get_implementation;
+      case SOFTFILTER_SUPERTWOXSAI:
+         return supertwoxsai_get_implementation;
       case SOFTFILTER_HQ2X:
          return hq2x_get_implementation;
       case SOFTFILTER_SCALE2X:
diff --git a/gfx/filter.h b/gfx/filter.h
index 2826d13fcb..58555d5568 100644
--- a/gfx/filter.h
+++ b/gfx/filter.h
@@ -53,6 +53,7 @@ enum
    SOFTFILTER_DARKEN,
    SOFTFILTER_HQ2X,
    SOFTFILTER_TWOXSAI,
+   SOFTFILTER_SUPERTWOXSAI,
    SOFTFILTER_SCALE2X,
    SOFTFILTER_LAST,
 };
@@ -60,6 +61,7 @@ enum
 #ifdef HAVE_FILTERS_BUILTIN
 const struct softfilter_implementation *hq2x_get_implementation(softfilter_simd_mask_t simd);
 const struct softfilter_implementation *twoxsai_get_implementation(softfilter_simd_mask_t simd);
+const struct softfilter_implementation *supertwoxsai_get_implementation(softfilter_simd_mask_t simd);
 const struct softfilter_implementation *twoxbr_get_implementation(softfilter_simd_mask_t simd);
 const struct softfilter_implementation *darken_get_implementation(softfilter_simd_mask_t simd);
 const struct softfilter_implementation *scale2x_get_implementation(softfilter_simd_mask_t simd);
diff --git a/gfx/filters/2xsai.c b/gfx/filters/2xsai.c
index 41db3eb280..1ac03d18c5 100644
--- a/gfx/filters/2xsai.c
+++ b/gfx/filters/2xsai.c
@@ -152,9 +152,8 @@ static inline int result2_rgb565(uint16_t A, uint16_t B, uint16_t C, uint16_t D)
    return r;
 }
 
-static void write2_rgb565(uint16_t *out, uint16_t val0, uint16_t val1)
+static void twoxsai_write2_rgb565(uint16_t *out, uint16_t val0, uint16_t val1)
 {
-   /* TODO: swap this for big endian */
    *((uint32_t*)out) = ((uint32_t)(val0) | ((uint32_t)(val1) << 16));
 }
 
@@ -286,8 +285,8 @@ static void twoxsai_generic_rgb565(unsigned width, unsigned height,
                product1 = interpolate_rgb565(colorA, colorC);
          }
 
-         write2_rgb565(out, colorA, product);
-         write2_rgb565(out + dst_stride, product1, product2);
+         twoxsai_write2_rgb565(out, colorA, product);
+         twoxsai_write2_rgb565(out + dst_stride, product1, product2);
 
          ++in;
          out += 2;
diff --git a/gfx/filters/Makefile b/gfx/filters/Makefile
index 9a4aebe97c..f2f58a5d6b 100644
--- a/gfx/filters/Makefile
+++ b/gfx/filters/Makefile
@@ -38,7 +38,7 @@ ASMFLAGS := -INEON/asm
 asflags += -mfpu=neon
 endif
            
-objects += 2xsai.so 2xbr.so darken.so hq2x.so scale2x.so
+objects += 2xsai.so super2xsai.so 2xbr.so darken.so hq2x.so scale2x.so
 
 all: build;
 
diff --git a/gfx/filters/hq2x.c b/gfx/filters/hq2x.c
index 619987b73b..64e8881b61 100644
--- a/gfx/filters/hq2x.c
+++ b/gfx/filters/hq2x.c
@@ -219,6 +219,8 @@ static uint16_t blend(unsigned colfmt, unsigned rule, uint16_t E, uint16_t A,
 {
    unsigned shift;
    uint32_t mask;
+
+   mask = 0;
    switch (colfmt)
    {
       case SOFTFILTER_FMT_RGB565:
diff --git a/griffin/griffin.c b/griffin/griffin.c
index 7fe66370d9..ea60dfd392 100644
--- a/griffin/griffin.c
+++ b/griffin/griffin.c
@@ -478,6 +478,7 @@ FILTERS
 
 #ifdef HAVE_FILTERS_BUILTIN
 #include "../gfx/filters/2xsai.c"
+#include "../gfx/filters/super2xsai.c"
 #include "../gfx/filters/2xbr.c"
 #include "../gfx/filters/darken.c"
 #include "../gfx/filters/hq2x.c"