mdavezac
commented
10 years ago From looking at the van Cittert-Zernike, I would expect a center of inversion at the origin of the visibility domain. However, that does not seem to be (explicitly) the case in the CASA input. Which makes sense, since its more data than needed. However, since purify uses complex to complex FFT transforms, we should include that data explicitly?
Purify expects U, V data in [-pi, pi[. So wrapper scales CASA UVW data by
pi/max |U|andpi/max |V|. CASA's UVW data is in meters, whereas the van Cittert-Zernike theorem involves an FFT over U and V in wave-numbers. However, the conversion to wave numbers would disappear when scaling to [-pi, pi[. So that has no influence.