Return to Home Page
Re: Software problem in time-series photometry > If your field rotates, C-offset will not work for long. The > C-offset routine does *NOT* centroid the V and K stars > except for the first frame. That is the idea - it follows > the C star and offsets a certain number of pixels in the X > and Y direction for both V and K stars. That way if the V > star goes very faint, you do not have the problem of > failing to centroid the V star. If you get a value that is > just barely measurable, fine. If there are other stars near > the V and K stars, when the V or stars get faint, the > centroid will gravitate to the other nearby stars. As I stated in the previous message, people must have not used constant C-offset in rotating images. However, as you inferred, the software may have re-centroided somerhing occassionally, and sometimes fails -- but I don't know what is really happening. > If stars are so faint that you can't centroid them, your > statistics are going to be bad, but why compound the > problem by measuring the star poorly? Among the set of data we received, there is a clear instance that the image quality did not apparently changed, but the software gave false results. Occasional poor-quality images can not be a natural explanation for this behavior. I had a chance to re-measure raw images, and the result was sufficiently good, while the initially reported data seemed unusable. This instance is sufficient to tell that some data must have been wasted even if the originally recorded information is useful. Regards, Taichi Kato
Return to the Powerful Daisaku Nogami
vsnet-adm@kusastro.kyoto-u.ac.jp
