diff --git a/community/kernel/src/main/java/org/neo4j/kernel/impl/api/store/StorePropertyPayloadCursor.java b/community/kernel/src/main/java/org/neo4j/kernel/impl/api/store/StorePropertyPayloadCursor.java
index 72bfc5e359996..5f5e666e44bfc 100644
--- a/community/kernel/src/main/java/org/neo4j/kernel/impl/api/store/StorePropertyPayloadCursor.java
+++ b/community/kernel/src/main/java/org/neo4j/kernel/impl/api/store/StorePropertyPayloadCursor.java
@@ -194,8 +194,7 @@ long longValue()
     String shortStringValue()
     {
         assertOfType( SHORT_STRING );
-        Bits bits = valueAsBits();
-        return LongerShortString.decode( bits );
+        return LongerShortString.decode( data, position, currentBlocksUsed() );
     }
 
     String stringValue()
@@ -356,9 +355,6 @@ private static Object readArrayFromBuffer( ByteBuffer buffer )
 
     private void assertOfType( PropertyType expected )
     {
-        if ( type() != expected )
-        {
-            throw new IllegalStateException( "Expected type " + expected + " but was " + type() );
-        }
+        assert type() == expected : "Expected type " + expected + " but was " + type();
     }
 }
diff --git a/community/kernel/src/main/java/org/neo4j/kernel/impl/store/LongerShortString.java b/community/kernel/src/main/java/org/neo4j/kernel/impl/store/LongerShortString.java
index 49b8a687356c5..44aaf247fc06f 100644
--- a/community/kernel/src/main/java/org/neo4j/kernel/impl/store/LongerShortString.java
+++ b/community/kernel/src/main/java/org/neo4j/kernel/impl/store/LongerShortString.java
@@ -497,6 +497,7 @@ char decTranslate( byte codePoint )
     public static final int ALL_BIT_MASK = bitMask( LongerShortString.values() );
     public static final int ENCODING_UTF8 = 0;
     public static final int ENCODING_LATIN1 = 10;
+    private static final int HEADER_SIZE = 39; // bits
 
     final int encodingHeader;
     final long mask;
@@ -763,20 +764,18 @@ private static void writeHeader( Bits bits, int keyId, int encoding, int stringL
      */
     public static String decode( PropertyBlock block )
     {
-        Bits bits = Bits.bitsFromLongs( block.getValueBlocks() );
-        return decode( bits );
+        return decode( block.getValueBlocks(), 0, block.getValueBlocks().length );
     }
 
-    public static String decode(Bits bits)
+    public static String decode( long[] blocks, int offset, int length )
     {
-        long firstLong = bits.getLongs()[0];
+        long firstLong = blocks[offset];
         if ( ( firstLong & 0xFFFFFF0FFFFFFFFFL ) == 0 ) return "";
-        bits.getInt( 24 ); // Get rid of the key
-        bits.getByte( 4 ); // Get rid of the type
-        int encoding = bits.getByte( 5 ); //(int) ( ( firstLong & 0xF00000000L ) >>> 32 );
-        int stringLength = bits.getByte( 6 ); //(int) ( ( firstLong & 0xFC000000L ) >>> 26 );
-        if ( encoding == LongerShortString.ENCODING_UTF8 ) return decodeUTF8( bits, stringLength );
-        if ( encoding == ENCODING_LATIN1 ) return decodeLatin1( bits, stringLength );
+        // key(24b) + type(4) = 28
+        int encoding = (int) ((firstLong & 0x1F0000000L) >>> 28); // 5 bits of encoding
+        int stringLength = (int) ((firstLong & 0x7E00000000L) >>> 33); // 6 bits of stringLength
+        if ( encoding == LongerShortString.ENCODING_UTF8 ) return decodeUTF8( blocks, offset, stringLength );
+        if ( encoding == ENCODING_LATIN1 ) return decodeLatin1( blocks, offset, stringLength );
 
         LongerShortString table = getEncodingTable( encoding );
         assert table != null: "We only decode LongerShortStrings after we have consistently read the PropertyBlock " +
@@ -784,16 +783,29 @@ public static String decode(Bits bits)
         char[] result = new char[stringLength];
         // encode shifts in the bytes with the first char at the MSB, therefore
         // we must "unshift" in the reverse order
-        for ( int i = 0; i < stringLength; i++ )
-        {
-            byte codePoint = bits.getByte( table.step );
-            result[i] = table.decTranslate( codePoint );
-        }
+        decode( result, blocks, offset, table );
 
         // We know the char array is unshared, so use sharing constructor explicitly
         return UnsafeUtil.newSharedArrayString( result );
     }
 
+    private static void decode( char[] result, long[] blocks, int offset, LongerShortString table )
+    {
+        int block = offset;
+        int maskShift = HEADER_SIZE;
+        long baseMask = table.mask;
+        for ( int i = 0; i < result.length; i++ )
+        {
+            byte codePoint = (byte) ((blocks[block] >>> maskShift) & baseMask);
+            maskShift += table.step;
+            if ( maskShift >= 64 && block + 1 < blocks.length )
+            {
+                maskShift %= 64;
+                codePoint |= (blocks[++block] & (baseMask >>> (table.step-maskShift))) << (table.step-maskShift);
+            }
+            result[i] = table.decTranslate( codePoint );
+        }
+    }
 
     // lookup table by encoding header
     // +2 because of ENCODING_LATIN1 gap and one based index
@@ -896,22 +908,40 @@ private void translateData(Bits bits, byte[] data, int length, final int step)
         }
     }
 
-    private static String decodeLatin1( Bits bits, int stringLength )
-    { // see decode
+    private static String decodeLatin1( long[] blocks, int offset, int stringLength )
+    {
         char[] result = new char[stringLength];
-        for ( int i = 0; i < stringLength; i++ )
+        int block = offset;
+        int maskShift = HEADER_SIZE;
+        for ( int i = 0; i < result.length; i++ )
         {
-            result[i] = (char) bits.getShort( 8 );
+            char codePoint = (char) ((blocks[block] >>> maskShift) & 0xFF);
+            maskShift += 8;
+            if ( maskShift >= 64 )
+            {
+                maskShift %= 64;
+                codePoint |= (blocks[++block] & (0xFF >>> (8-maskShift))) << (8-maskShift);
+            }
+            result[i] = codePoint;
         }
-        return new String( result );
+        return UnsafeUtil.newSharedArrayString( result );
     }
 
-    private static String decodeUTF8( Bits bits, int stringLength )
+    private static String decodeUTF8( long[] blocks, int offset, int stringLength )
     {
         byte[] result = new byte[stringLength];
-        for ( int i = 0; i < stringLength; i++ )
+        int block = offset;
+        int maskShift = HEADER_SIZE;
+        for ( int i = 0; i < result.length; i++ )
         {
-            result[i] = bits.getByte();
+            byte codePoint = (byte) (blocks[block] >>> maskShift);
+            maskShift += 8;
+            if ( maskShift >= 64 )
+            {
+                maskShift %= 64;
+                codePoint |= (blocks[++block] & (0xFF >>> (8-maskShift))) << (8-maskShift);
+            }
+            result[i] = codePoint;
         }
         try
         {
@@ -930,13 +960,6 @@ public static int calculateNumberOfBlocksUsed( long firstBlock )
          */
         int encoding = (int) ( ( firstBlock & 0x1F0000000L ) >> 28 );
         int length = (int) ( ( firstBlock & 0x7E00000000L ) >> 33 );
-        /*
-        Bits bits = Bits.bitsFromLongs( new long[] {firstBlock} );
-        bits.getInt( 24 ); // key
-        bits.getByte( 4 ); // type
-        int encoding = bits.getByte( 5 );
-        int length = bits.getByte( 6 );
-        */
         if ( encoding == ENCODING_UTF8 || encoding == ENCODING_LATIN1 )
         {
             return calculateNumberOfBlocksUsedForStep8(length);