Normalisation

Normalisation

Normalisation is a method used in binary to make sure that the binary number is as accurate as possible using the number of given bits.  Similar to standard form, where a large number can be represented easily with less characters.

This idea links closley  to that of Two's complement and floating point binary.

EXAMPLE:

234000 can be shown as 
23400 x 10^1
2.34 x 10^5  <-- this would be the more logical choice.

 

When this method is used on a binary number, the binary digit has then been 'normalised' or is in 'normal form'. 

Here is an example of normalising a binary number, this method uses floating point binary.

EXAMPLE:

108 in decimal = 01101100 ( 8 bit )

once normalised = 0.1101100  <--The binary point is ALWAYS placed just before the MSB(Most Significatnt Bit)
                                                                     <-- This is called the Mantissa 

How ever to get our normalised binary number back to our decimal equivalent, we need to move the binary point 7 places to the right. So our exponent must be 7.

Exponent = 7 
7 = 0111

Now our fully normalised Binary number is

Mantissa         Exponent
0.1101100 | 0111