[JohnM]

Currently the REW filter recommendations are aimed at modal resonances. REW examines the response, identifies features that have the characteristics of modal resonances and calculates filters to counter them. The filter recommendations depend on the target level, as only features above the target are considered for filtering. The identification process is not foolproof, REW may miss some features that should be countered or it may apply a filter where it would be more correct to have 2 separate filters or no filter at all. It also does not apply any filters to shape the response to a desired overall trend. I'm working on improvements to the process of choosing and optimising filters, but there needs to be an element of user judgement applied to the REW recommendations. REW also does not currently consider any audibility criteria for the resonances it detects - some research indicates even prolonged resonances at low frequencies have little audibility, see the papers on the Genelec web site for example.