using wavelet denoising as preprocessing function with real time data.

When I train a neural network I need to process the training data X with multivariate wavelet denoising obtaining a new data set denoised X_den.

   level = 4;           
   wname = 'sym2';         
   tptr  = 'heursure';  
   sorh  = 's';       
   mode  = 'asym';    
   SCAL  ='mln';     
   npc_app = 'none';
   npc_fin = 'none';

   [X_den, npc, nestco] = wmulden(X, level,wname,'mode',mode, npc_app, ... 
                                   npc_fin, tptr, sorh);

   [mynet,tr]=train(mynet,X,Y);

After training I need to use 'mynet' to calculate the output of unknown data X(i).

   output(i)=mynet(X(i));

Unknown data is obtained in realtime one by one and ,to be consistent with the trained network, I must denoise X(i) using the same Wavelet's parameters calculated previously.

But manual doesn't help me...

Thanks.