How do I select my final scale and pixfrac?

bourquebourque Member, Institute_Staff Posts: 2
My program used a UVIS-DITHER-BOX pattern for subsampling the PSF. How do I find the best parameters for final_scale and final_pixfrac? What tests can I do to determine the optimal values for my data?


  • gunninggunning Member, Institute_Staff Posts: 10
    A good example for finding the best final_scale and final_pixfrac can be found at

    This method works for both WFC3 and ACS.
  • mutchlermutchler Member, Institute_Staff, Moderator_DrizzlePac Posts: 6
    I'm glad to hear you have well-sampled data. Most people who ask this question should NOT modify the default scale and pixfrac, because they do not have optimally sampled images. Obtaining subsampled images is expensive -- e.g. 4 exposures in a dither box pattern -- so most typical (smallish) programs can't afford to do it. But if you do have subsampled data, then there is good advice in the link above. The large and legacy programs that have delivered High Level Science Products (HLSP) to the archive always have a nice webpage and detailed documentation describing how their team of experts processed the data. They can often afford to obtain very well-sampled data, so their decisions about scale and pixfrac can be very good examples to follow. Watching the resulting weight map is the best overall advice: make sure the exposure weight doesn't vary more than 30% (and certainly should lead to empty pixels).
Sign In or Register to comment.