0.00
#...
·
 |
|---|
Using my 1600MM and 533MC I take 50 darks with same temp, gain settings, then on the night (morning) 40 flats and 20 dark flats. This seems to work pretty good for me. Good luck with yours.  |
0.90
#...
·
|
|---|
| I use an ASI 533 MM and I take 200 bias, 25 darks and 25 flats that I renew twice a year. |
1.81
#...
·
·
1
like
|
|---|
Ivaylo Stoynov: what about flats frames…surely you can’t keep the optical elements perfectly free of dust all the time..? |
9.58
#...
·
·
1
like
|
|---|
AstroShed:Ivaylo Stoynov: In fact I keep the system clean. It is a permanent setup... Every time when there is a need to open something in the train reducer-wheel-camera am taking care to leave the things clean. So far this approach is working well for me in last few years as I don't see any artifacts in my images...I keep one virgin sample of these (the first one of just opened box, without the handle of course) in a plastic bag. Taking it out only when it is used. Before putting two things together, very gently without any pressure, I am moving only the white part over the surfaces to be sure that there are no dust particles left.  |
|
9.58
#...
·
|
|---|
| BTW if there is need to use flats, then at least dark-flats and biases will be needed in order to make the things right. My logic is - use only the raw data in big amounts + dithering and leave the math to do its job... |
2.41
#...
·
|
|---|
| 5-10 darks,5-10 flats -sufficient for short session 1-2 hours, of course, if you have 20-30 hours total imaging time -need x10 times darks and flats. But our weather prevents thus madness, for the last three months barely few hours of clear night 👹👹👹 |
22.40
#...
·
·
4
likes
|
|---|
Adel Kildeev: Who told you that you "need 10x times darks and flats"? That is simply not correct. 16 darks and even fewer flats will be fine for 20-30 hours of total imaging time. I'm not just making this up. Please look at the calculations and the charts that I've posted. It is fairly straightforward to show that using such large numbers of dark, flat, and bias frames provides very little advantage over ~16-20 frames--for even up to 100s of integrated frames. John |
0.00
#...
·
|
|---|
|
30 Darks, no bias, the bias information is in the dark frames, there's no point subtracting it twice... IMX294 colour - so despite the noise being quite reasonable, I need to shoot darks to control the amp glow. |
6.65
#...
·
|
|---|
I want to know how many people take, and also why they take that amount, so I can learn from them. It looks like you are working with CMOS cameras. For these cameras if you are going to do flats you must do either biases or darks, but not both, that match your lights (biases need the same gain and temperature as the lights, darks also need the same exposure). And you need to do biases or dark flats that match you flats, with the same criteria as with the lights. The reason is your flats will not correct your lights properly at all unless you subtract the bias at the very least. Same with flats, without at least subtracting your bias your flats will not correct properly. If you don't use flats, you may possibly get aw If you need to do darks, darks include the bias, so you should not do a bias. Biases are only required with darks if your darks are not at the same temperature and you are trying to correct for that problem. Whether you need darks or only biases depends on your read noise and exposure time and temperature and the dark noise. You need to do a very simple calculation to decide if you need darks. Look at the camera spec for dark current in e/px/s for your camera at your chosen temperature. Multiply by your sub exposure time. This gives you the dark current in your sub. Take the square root, this is the dark noise in your sub. As long as the dark noise is less than about 1/10 of your read noise at your chosen gain, the dark noise is neglible compared to read noise. You are okay then to just use a bias. Otherwise use a dark. For my 2600 I am now using 100 bias frames, which reduces the read noise in the bias to be about 1/10 of that of a single frame. This is overkill, but biases are so fast to capture it seems silly not to do more. Do at least 20 though just to make sure that no odd issues in one frame mess up the results, and use some sort of rejection in your master bias again for any weird pixels in your stack. Finally, in most instances you must dither. If you take 20 biases/darks and 20 lights, the read noise from the biases/darks is equal to the read noise from the lights, and you reduce S/N by the square root of 2, by 1.41X. You need the number of biases/darks x number of dithers to be about 3.5X the number of lights, this ensures the biases/darks add less than 5% to the read noise from the lights. So for 20 biases say, you can do about 6 frames before dithering, to keep the 3.5 x ratio to the lights. Flats do not contribute as much noise since they are taken near 50% adu, where read noise is insignificant. But still do at least 20 so you can use rejection in your stack. Hope this helps Rick |
1.20
#...
·
|
|---|
|
Asi2600mc pro When shooting narrowband 300s subs at gain 100 through the L-Ultimate, I use no calibration frames. When shooting broadband, typically between 60s - 240s at gain 0, I take 15 flats and 15 dark flats. Exposure is between 0,5 - 2s. These approaches give me excellent results, every single time. |
4.53
#...
·
|
|---|
| ASI533MM Pro and MC Pro, 20 or so flats and bias, no darks or dark flats. Normally the darks would make no difference, and occasionally would give me a worse result than if I omit them. Dark current in the 533 is very low and it doesn’t suffer from amp glow. |
0.00
#...
·
|
|---|
| 21 darks, flats, bias (imx571m) |
0.00
#...
·
|
|---|
| Something like 30 to 50 of each! More than that is just wasting time. For a good dark in a cooled sensor, the minimum should be 20. |
0.00
#...
·
|
|---|
| I currently use no darks. Only 25 biases and 25 flats. There is no need for enormous amount of calibration frames. |
1.51
#...
·
·
1
like
|
|---|
|
Maybe my suggestion is a little out of the norm, but I would recommend that you spend the time doing some testing and determining what works best for you. Lot's of great information has so far been provided, and this information can be used to determine a place to start, but how many calibration frames and which ones you think you should be taking can only be decided by you. I think that the only aspects of imaging that you need to consider are image quality, time involved and the effort expended. Remember that the main goal should be that you are satisfied with your images. Miguel  . |
0.90
#...
·
·
2
likes
|
|---|
CVZ_Astrophotography: I basically do the same thing…but with 25 vs 20. my camera is a ZWO 533 which supposedly doesn’t require darks due to no amp glow…but I find it’s good practice and can’t hurt to keep the library and use it. |
11.91
#...
·
·
7
likes
|
|---|
| I’m reading some of these posts here and wonder - do people bother to actually learn what calibration frames are and why they are needed? Or just come up with some random protocol and claim, with no basis in fact or evidence, that is how it should be done. For example, claims that flats are not needed. Or darks are only needed if the camera has amp glow. Or one needs 10x the number of darks. If people bother to actually read the thread, there is excellent information with basis in science on what numbers of frames are practically needed. On the other hand, there is some joy to be had, I suppose, in coming up with one’s own concoction. |
|
1.51
#...
·
·
1
like
|
|---|
Arun H: When I was a little boy, I looked at the heavens through the cardboard tube from an empty roll of paper towels. I was sure that I could see things more clearly, and everything seemed to be much closer. Taping two tubes together improved things dramatically. Since then I have learned a lot. The magic of innocence may have drifted away, but it has been relentlessly replaced with reality and the acknowledgement of facts. But to this day, I am still not sure which is more pleasing. Whether fact or fancy, grown up or child, viewing through two side by side tubes is an improvement over viewing through just one. I guess that some things never change no matter how much older and wiser one gets. Miguel . |
0.00
#...
·
·
1
like
|
|---|
Just one more word : I won't argue about the proper (theory-wise) amount of dark frames . Personally, I would :
I have never understood the use of biases frame for lights, since bias error is included in both lights and darks, and thus any subtraction will eliminate biases... I use biases with flats for standards exposures, including Nb filters as soon as exposure time is < 10s. Clear skies |
7.46
#...
·
·
1
like
|
|---|
not forget to use dithering : it costs nothing Of course they cost, they cost oberving time. May be a lot, may be a little but it will still cost something. |
|
1.51
#...
·
·
1
like
|
|---|
andrea tasselli:not forget to use dithering : it costs nothing Yes, but one small dither for every 10 minute frame cost very little in imaging time, but the gains are unquestionably required. Have you ever seen the damage that can be done by walking noise? FYI... One shouldn't dither when observing. Miguel . |
|
7.46
#...
·
|
|---|
Miguel Morales: It costs time and that's a fact and how much depends on the specific circumstances and I won't go into those. When I rent time dithering adds to the bill in non marginal way but it has to be done for other reasons (time cost). Personally, I never dither and I have never dithered in the whole of my observing life and that is quite a stretch. I don't have walking noise. But I do calibrate my data very very accurately. Others' experience may differ... |
0.00
#...
·
·
1
like
|
|---|
|
I find it fascinating to read this thread and how misunderstood this topic is. Simply put, this topic is all about math. It's just that simple. It should not be ignored, or your end result will have the consequential flaws. They may not be discernible to some, but trust me and the others who have taken the time to study this topic, the flaws are there. I would point to Arun's earlier post where he directs us to John Hayes' previous work where he and others have taken the time to study the math and outline it to us in layman terms. If you take the time to understand it, it will help you to appreciate why the calibration process is so important to your processing routine. https://www.astrobin.com/forum/c/astrophotography/deep-sky/lets-discuss-about-dark-bias-dark-flats/?page=1 The way I look at this and I think why the reason the OP has asked this question, is that we have spent so much time, energy and money collecting our data, why wouldn't we want to get the most out of what we collected? Afterall, a pretty picture isn't as pretty if it has mathematical defects within. My 2 cents. |
|
1.51
#...
·
·
1
like
|
|---|
Barry Schellenberg: Yes, some people will disagree, and some other will argue until their heritage clearly shows. But math seems to work for nearly everything, even though nobody can explain exactly why. I for one like to read what John Hayes has to say. He is the most knowledgeable person that I know of for the subject that we enjoy. If I were to believe anyone, it would be John Hayes, everyone else can go kick rocks. Miguel . |
|
11.91
#...
·
·
1
like
|
|---|
Miguel Morales:Barry Schellenberg: I mean, there is a pretty good reason math works. All this is based on reasonably simple statistics that is covered in many college and even some advanced high school courses. More importantly, things like Poisson statistics and the central limit theorem show excellent agreement with observation. In a separate thread here, John showed how someone took hundreds of darks and the noise profile basically fell exactly in line with the theoretical prediction. What would be surprising is if it didn’t. Our world works on the basis of well understood laws of physics. Learn them, believe them, use them. |
