samp=samp_vis(ad,u,v,vis) Sample visbility surface at a set of u,v points If vis is 3D "stack" of different frequencies then the 3rd dimension of u,v will be assumed to be freqs also When working on szasim power spectrum runs found that using interp other than 'linear' can result in serious bias problems in the final bandpowers.
0001 function samp=samp_vis(ad,u,v,vis) 0002 % samp=samp_vis(ad,u,v,vis) 0003 % 0004 % Sample visbility surface at a set of u,v points 0005 % If vis is 3D "stack" of different frequencies then the 3rd dimension 0006 % of u,v will be assumed to be freqs also 0007 % 0008 % When working on szasim power spectrum runs found that using interp 0009 % other than 'linear' can result in serious bias problems in the 0010 % final bandpowers. 0011 0012 if(size(u)~=size(v)) 0013 error('u,v must have same size'); 0014 end 0015 0016 if(ndims(vis)==2) 0017 samp=interp2(ad.u_val,ad.u_val',vis,u,v,'linear'); 0018 else 0019 0020 if(size(u,3)~=size(vis,3)) 0021 error('u,v and vis must have the same 3rd dimension size (freqs)'); 0022 end 0023 0024 for i=1:size(u,3) 0025 samp(:,:,i)=interp2(ad.u_val,ad.u_val',squeeze(vis(:,:,i)),... 0026 u(:,:,i),v(:,:,i),'linear'); 0027 end 0028 end