If you are simply retrieving image texture you can as well use the GRD images instead of SLC and simply apply the calibration (to Sigma0, no complex output), texture derivation and terrain correction. With GRD there is no need for debursting, multi-looking, resampling.
Also, filtering ruings much of the texture you are expecting to get so I would skip that as well.
See this post for details: Radiometric & Geometric Correction Workflow