The idea is to use one of several fractal characterisation as a criterion to stop the pre-whitening analysis. This would be done by analysing the fractal behaviour of the residuals released by the PWHT method at each step (i.e. after removal of selected frequencies).

Ideas

• Establish a threshold (arbitrary, statistical?) to stop the PWHT
• Look at the change of the fractal characterisation along the process. This might provides additional information (for instance, the criterion would rely not only on a number but also on a first derivative at the same time)

<todo>Provide access to DFE server to S (JRR, JPG) </todo>
<todo>Perform a first test on a selected star (S)</todo>
<todo>Discussion of the first results and determine the further steps (ALL)</todo>