2.94
#... |
|---|
|
I am thinking about building a library for some of my calibration subs just in case if i didn't take them correctly outdoor or during the imaging time, so i need to know few things if possible, prefer the answer from those who tried it really. 1. How many dark subs you take or how many is enough really? 2. What is the Gain and Offset do you use for LRGB filters and for narrowbanding filters? Assuming 3 factors for above questions:
|
9.90
#... |
|---|
|
Hi Tareq, The Bortle and the focal ratio does not matter for darks. The rule is to use darks at the same temperature (you are trying to subtract the thermal noise) and gain settings. So the darks only have to correspond to the light frames settings you used. It is highly recommended to keep a darks/bias library when using cooled CMOS (maybe renew it from time to time). There is no general agreement for the number of dark frames, but remember that the SNR will vary with the number of subs. Usually between 30 and 50 darks should suffice. I use 50 generally with my ASI1600. Bias are easy to do, so 50 to 100 is a good choice. Since you can use the Master Dark and Bias afterwards in Pixinsight, you do it once. The gain settings with the CMOS camera depend on what you are doing : how bright your sky is (like mine, pretty bright, Bortle 8), how bright your target is (how long you have to expose) and the read noise (how many subs are you going to take). The higher the gain, the narrower your dynamic range is and with a long exposure the pixel will be saturated faster than with a lower gain. The same number of photons (converted into electrons) would have reached the chip, but since the gain is higher, you have less ADU to express the information (for example, all the pixels above value X are considered white, maximal value). The data is clipped, even if you add 60s to that sub, the core of the galaxy is already completely white for example, so no information will be added. It's still going to be white. With a lower gain, you could still be under the maximal pixel value, so you are going to have some shade (=information) in the core. The higher the gain, the lower the read noise - with this chip the read noise has a fixed and variable part, and since the fixed part is constant, you decrease the variable part and the overall read noise. So if you take lots of subs, you are going to introduce a lot of read noise. A higher gain setting will decrease that. But it will also decrease your dynamic range, as explained above. There are mathematical formulas and graphs in order to determine what's the optimal duration under your skies. The safest thing to do is use unity gain (QHY should provide that). For ASI 1600 this is 139/21 or 139/50 (their new default offset is a little higher). I use that for narrowband for example. Some people use 200/50 even for narrowband. It is unlikely that you are going to have very bright targets in narrowband and that you are going to clip information. For LRGB I use 75/15, but I think I'll try unity gain. The thing that gets rapidly burned in LRGB are stars and their color. If your RGB data for stars is clipped, they are all going to be white. So try to use lower gain/shorter exposure to avoid clipping. Since my sky is polluted though, I will use shorter subs (2 to 3 min) so more subs and more read noise. But what determines my noise is the sky, not the read noise. As always, you have to try. I would recommend using unity gain at first much simpler. Use shorter subs for LRGB (if you have no LP filter) and longer for SHO. You can check LRGB images on Astrobin with the QHY and see the gain settings. If you use a lower gain (like 75 for ASI), make a separate bias and dark library/masters. Hope that helps. CS, Bogdan |
