some distribution:  f(y)dy  with free parameters
Gaussian distribuion:  g(z)dz  fixed mean and sigma
x and y independent.
Distribution of x: h(x)dx  where x=y+z.
Fit h(x) to a number of data points (x, H(x))
and get the parameres for f(y).


Getting  h(x):

h(x)dz= dx Integral delta(x- y-z) f(y)dy g(z)dz
      = dx Itegral dy  f(y) g(x-y)
      
	   



