Big eared stars or the effects of the "CONTOUR" parameters on raw images acquired with Philips (ToUcam or Vesta Pro) webcams
Every one who has taken starfield images with a PWC (ToU or VP philips webcam) has seen the result on the stars apearance of stacking images: 2 black lobes appear beside each star (upper right image); removing them by post-processing images is possible but tedious and of course, avoiding to produce them would be a far better solution. The cause of the problem is in fact a pre-processing algorithm implemented in these cams ; this process is controlled via a register of the cam ; unfortunately, the Windows driver does not give access to this register, and hence does not provide user-control on that process. In fact, what we want is to just turn it off...
The linux driver does provide that control, and so does qastrocam, acquisition software for linux. Here are the results of a (quick) comparison of raw images obtained with CONTOUR (called SHARPNESS in qastrocam) enabled/disabled.
The upper left image () shows 2 shots acquired at the primary focus of a Tak fs102 using a Vesta Pro camera illustrating the devastating effects of the contour parameter on raw images.
Left shot acquired with Contour set to 66%, which is I think its default value. Right shot acquired with Contour set to 0% See in particular the couple of stars in the upper left corner : the alogrithm digs a blackhole between the 2 stars. In fact it does the same on every feature in the image.
Since these effects are more or less systematics, they cumulate and stacking 50 such images results in the big-eared stars we all recognize on the image on the right. Logically, globular cluster images are the preferred victims of CONTOUR. Compare for example the 2 m13 images : the older one is of really poor quality, and not only because of the ears and also it has not been taken with the BW1/4" chip ; but it has been acquired with the windows driver, with contour inevitably enabled, while the recent one has been obtained under linux with CONTOUR disabled. Those 2 images just illustate the fact that in the case of a globular cluster, you will probably get thousands of ears instead of thousands of stars if you cannot disable CONTOUR... (see also that m22 image, CONTOUR disabled)
In conclusion, having the capability of disabling CONTOUR is a real plus in acquirring clean raw images. This might be the occasion to give a try to linux if you're under Windows ; rewriting the windows driver (that would have a lot of benefits apart giving control on CONTOUR and DNR parameters) is the other branch of the alternative.
Meanwhile, you might (must) test all this stuff with lin4astro, a specially designed linux distribution: it is all on only 1 CD and you can boot on the CD without installing the stuff on your hard drive, at the cost of a somewhat longer bootup process. But it works, and once you tested it, no doubt you will want to install it (or any other linux distribution) on your hard drive. Welcome to linux :)
fm, 2003/07/24