diff --git a/lib/base.c b/lib/base.c
index 2142039..8b5a617 100644
--- a/lib/base.c
+++ b/lib/base.c
@@ -29,39 +29,29 @@ word_t word_byteswap(const word_t w)
 hword_t convert_bytes_to_hword(const byte_t *bytes)
 {
   hword_t h = 0;
-  if (LITTLE_ENDIAN)
-    memcpy(&h, bytes, HWORD_SIZE);
-  else
-    for (size_t i = 0; i < HWORD_SIZE; ++i)
-      h |= ((hword_t)(bytes[i]) << (8 * i));
+  memcpy(&h, bytes, HWORD_SIZE);
+  if (!LITTLE_ENDIAN)
+    h = hword_byteswap(h);
   return h;
 }
 
 word_t convert_bytes_to_word(const byte_t *bytes)
 {
   word_t h = 0;
-  if (LITTLE_ENDIAN)
-    memcpy(&h, bytes, WORD_SIZE);
-  else
-    for (size_t i = 0; i < WORD_SIZE; ++i)
-      h |= ((word_t)(bytes[i]) << (8 * i));
+  memcpy(&h, bytes, WORD_SIZE);
+  if (!LITTLE_ENDIAN)
+    h = word_byteswap(h);
   return h;
 }
 
 void convert_hword_to_bytes(hword_t w, byte_t *bytes)
 {
-  if (LITTLE_ENDIAN)
-    memcpy(bytes, &w, HWORD_SIZE);
-  else
-    for (size_t i = 0; i < HWORD_SIZE; ++i)
-      bytes[i] = WORD_NTH_BYTE(w, i);
+  hword_t h = LITTLE_ENDIAN ? w : hword_byteswap(w);
+  memcpy(bytes, &h, HWORD_SIZE);
 }
 
 void convert_word_to_bytes(word_t w, byte_t *bytes)
 {
-  if (LITTLE_ENDIAN)
-    memcpy(bytes, &w, WORD_SIZE);
-  else
-    for (size_t i = 0; i < WORD_SIZE; ++i)
-      bytes[i] = WORD_NTH_BYTE(w, i);
+  word_t h = LITTLE_ENDIAN ? w : word_byteswap(w);
+  memcpy(bytes, &h, WORD_SIZE);
 }