diff --git a/rawimage_sse.cpp b/rawimage_sse.cpp
index 87ef938..cf9c8dd 100644
--- a/rawimage_sse.cpp
+++ b/rawimage_sse.cpp
@@ -84,15 +84,15 @@ void fromPlanarSSE(const void *in, void *out, size_t count)
         switch(sizeof(T))
         {
         case 1:
-            for(uint32_t o=0; o<ch; o++)static_cast<uint8_t*>(out)[i + o] = static_cast<const uint8_t*>(in)[i + o + s2];
+            for(uint32_t o=0; o<ch; o++)static_cast<uint8_t*>(out)[i*4 + o] = static_cast<const uint8_t*>(in)[i + o*s2];
             if(ch==3)static_cast<uint8_t*>(out)[i*4 + 3] = 0xff;
             break;
         case 2:
-            for(uint32_t o=0; o<ch; o++)static_cast<uint16_t*>(out)[i + o] = static_cast<const uint16_t*>(in)[i + o + s2];
+            for(uint32_t o=0; o<ch; o++)static_cast<uint16_t*>(out)[i*4 + o] = static_cast<const uint16_t*>(in)[i + o*s2];
             if(ch==3)static_cast<uint16_t*>(out)[i*4 + 3] = 0xffff;
             break;
         case 4:
-            for(uint32_t o=0; o<ch; o++)static_cast<uint32_t*>(out)[i + o] = static_cast<const uint32_t*>(in)[i + o + s2];
+            for(uint32_t o=0; o<ch; o++)static_cast<uint32_t*>(out)[i*4 + o] = static_cast<const uint32_t*>(in)[i + o*s2];
             if(ch==3)
             {
                 if(!std::numeric_limits<T>::is_integer)static_cast<float*>(out)[i*4 + 3] = 1.0;