This function is not yet fully documented. This is a transcript of the text-formatted help.

bbnoisy  Create a noise version of a black-box operator.
   BB=bbnoisy(BB1,L) creates a version of the BB1 where a
   matrix-vector product is perturbed by a random vector
   of norm L.

   If L=[L1,L2] the perturbation is a sum of uncorrelated
   noise norm with norm L1 and L2*norm(R), where R is the
   result of the computation. This can be used to model
   an operator with a high relative accuracy.

   BB=bbnoisy(BB1,L,LH) uses L as parameters for A*X and
   LH as parameters for A'*X. This models an asymmetry 
   in the accuracy of the products.

   This function can be used to test the reponse of iterative
   routines facing a limited accuracy of the product routines.

   See also bbtauest, bbcompare.