Hi,
The basic difference between a fixed and floating point processor is in the way data
is represented.
Data format determines the ability to handle signals of differing precision, dynamic range and signal to noise ratios.
The floating points processors have greater dynamic range and less quantisation noise.
In intensive signal processing apliccations it is better to use floating points processors which yields more accurate results and the implementer need not worry about rounding off, trunaction and saturation issues dealt with in fixed point processors.
In control applications where much multiply accumulation is not involved fixed processors can be used.
floating point processors are costly.