Subexposure Times with the Hyperstar

There are an upper and lower bound to consider when considering the optimum sub-exposure time to use. The lower bound is governed by the shortest sub you can take before CCD noise becomes intrusive. For the SXV-H9C camera, with the Hyperstar, and my typical imaging conditions, this would typically be around 5 seconds or so. The upper bound is where sky glow limits your integration time so that your dim deep-sky objects get lost in the sky background. You must clearly operate well below the time when your image is completely "washed out" by sky glow, but at the same time it is good to maximise your sub-exposure time in order to obtain "deep" images - that is images that are able to pick up very faint objects.

Now I will readily admit that I got into a lot of arguments over the "optimum" sub-exposure time with people on imaging Forums, and the main reason for the disagreements I had was that they were not familiar with very fast systems like the Hyperstar, and I was not familiar with "bog-standard" i.e. the depressingly slow imaging that most conventional imagers suffer. Eventually we did iterate to some sort of agreeable conclusion and it all proved to be a very interesting exercise. I will try to get the main points across in a logical fashion, but it is a very confusing business.

Because you are using a very "fast" optical system, your sub-exposure times can be very short - this after all is the main selling point of the Hyperstar. If your sub-exposure times are very short, then your tracking doesn't need to be very precise and yet you will still get good round stars. Also, if your sub-exposure times are very short, then you will take a very large number of sub-exposures during your imaging session, and when you stack these together to form your final image it will have a very high signal to noise ratio and the image will be as smooth as glass! Excellent, that's just what we want, but it's not a win-win situation. I was surprised when a supernova hunting colleague of mine examined one of my Hyperstar images (composed of more than a hundred sub-exposures) and he said that although it was a very nice low-noise image, it didn't go particularly deep, i.e. it didn't pick up very faint stars or galaxies, and that he could go deeper with a similar aperture scope using a higher f#. He also used a longer sub-exposure time of course, that goes naturally with the bigger f#, but why could he image deeper than me even if I used a much larger number of sub-exposures? The answer really is just the sub-exposure time. I typically used short sub-exposure times of less than a minute and this basically limits how deep you can go. On Moonless nights with good seeing I could approach nearly two minutes for sub-exposures, but I was always fighting against the intrusion of sky glow and the associated problem of removing gradients from my images, as well as having the problem of saturating (and therefore losing information) bright parts of objects.

So the power of the Hyperstar is the ability to collect a large number of subs in a reasonably short time so that the resulting low-noise (high quality) image can be obtained in a matter of only an hour or two. However, if the discussion above is true, I am going to find I run into trouble when trying to image very faint objects and the speed of the Hyperstar will not help me much in these instances. Does this agree with experiment? The answer is yes it does seem to. I was very surprised to find that even with very long total exposure times I couldn't get the quality of image I was used to when imaging rather faint objects like the Crescent Nebula and the Jellyfish nebula - I was running into the problem of not being able to take long enough sub-exposures to image these faint objects without simultaneously increasing my noise level to the point where I didn't gain - because I was coming up against sky glow. I was starting to find the imaging limitations with the seemingly limitless Hyperstar system. A way around this problem, with the emission nebulae at least, is to take narrowband H-alpha images of the region and combine them with the RGB data. The narrowband filter will cut back the sky glow enormously allowing you to use much longer sub-exposure times and giving you a much greater contrast image at the H-alpha wavelengths. You could also image in this way with other narrowband filters and combine the different narrowband images to give a colour image of great depth. Many imagers follow this approach, and narrowband imaging is very popular, especially in regions with bad light pollution. Since you are only imaging one wavelength at a time with narrowband filters you can use higher resolution black and white imagers and there is no need to consider the one-shot colour cameras for narrowband filter imaging. You do however have a problem with trying to use these filters with the Hyperstar! The only place to put a filter in the Hyperstar system is the gap between the end of the Hyperstar and the CCD camera. Physically that is fine, optically it is not so good. The very fast Hyperstar optics means that the cone angle of the light rays passing through the filter is pretty sharp. Now you will find that interference filters have very good specifications for parallel light rays striking the filter surface perpendicularly, but that as the angle moves away from the perpendicular, then not surprisingly, the filter characteristic changes. This is why filter manufacturers often quote the lowest f# system you can use their filters with. Unfortunately, the f#1.85 Hyperstar is not particularly useful for working with narrowband filters, a filter specified with a bandwidth of 6nm would have a much greater bandwidth if used with the Hyperstar with the associated loss in contrast.

So how long were my typical sub-exposure times using the Hyperstar and the IDAS LP filter? If seeing conditions were average there was little to be gained in using sub-exposure times in excess of 60 seconds. Under good clear dark skies, a 90 second sub-exposure time seemed to work well for me. I saw no improvement in final image quality, or depth, if I used 120-second sub-exposure times for the same total length of time under good seeing conditions. However, these are parameters you must work out for yourself (experimentally) under your own local viewing conditions.

+2 0

Post a comment