I’m going to send a wide field rig to a remote observatory in the fall. It is very unlikely I will have a flat panel so I need to do sky flats. I also want to take dark flats but I can’t wrap my mind around a simple question.
if I use Nina and the sky flats wizard, they won’t all have the same exposure so how do I match that ?
If I send the Stellavita instead, I can make them all the same exposure but they won’t all have the same ADU.
So how do people do that ? Is there just no way to take dark flats and sky flats ? Or is it so simple I’m just not seeing it because I’m looking for a complicated solution ?
Flat darks with sky flats in remote observatory
Well written Respectful Engaging
Either way it won’t matter. For time sensitive dark-flats the latter is the better choice while for time-insensitive dark-flats the former is the better choice. I do both, irrespective.
Eric Gagné · Jul 3, 2026, 04:18 PM
I’m going to send a wide field rig to a remote observatory in the fall. It is very unlikely I will have a flat panel so I need to do sky flats. I also want to take dark flats but I can’t wrap my mind around a simple question.
if I use Nina and the sky flats wizard, they won’t all have the same exposure so how do I match that ?
If I send the Stellavita instead, I can make them all the same exposure but they won’t all have the same ADU.
So how do people do that ? Is there just no way to take dark flats and sky flats ? Or is it so simple I’m just not seeing it because I’m looking for a complicated solution ?
Dark current in most modern CMOS camera is 100% insignificant when compared to a 50% level of light signal, particularly at exposures well under a minute. You don’t need flat-darks and can instead use a quality bias master and achieve what you need very simply.
Well written Concise
Rick Krejci · Jul 3, 2026 at 05:38 PM
Eric Gagné · Jul 3, 2026, 04:18 PM
I’m going to send a wide field rig to a remote observatory in the fall. It is very unlikely I will have a flat panel so I need to do sky flats. I also want to take dark flats but I can’t wrap my mind around a simple question.
if I use Nina and the sky flats wizard, they won’t all have the same exposure so how do I match that ?
If I send the Stellavita instead, I can make them all the same exposure but they won’t all have the same ADU.
So how do people do that ? Is there just no way to take dark flats and sky flats ? Or is it so simple I’m just not seeing it because I’m looking for a complicated solution ?Dark current in most modern CMOS camera is 100% insignificant when compared to a 50% level of light signal, particularly at exposures well under a minute. You don’t need flat-darks and can instead use a quality bias master and achieve what you need very simply.
I have always been under the impression that this was true for broadband filters that have very short exposures on flats but on narrowband the flats are often around 10 seconds. These would still be properly calibrated with bias frames ?
My darks are from a library as is my bias. The only calibration frames I shoot with the session are sky flats. I’ve always worked that way and it’s fine. I do cool the camera to session temp for the flats but I’m not sure that really does much. If you dither properly you can likely do away with the darks but I have them so I use them. It simple and it works.
Helpful Concise
Eric Gagné · Jul 3, 2026, 06:03 PM
Rick Krejci · Jul 3, 2026 at 05:38 PM
Eric Gagné · Jul 3, 2026, 04:18 PM
I’m going to send a wide field rig to a remote observatory in the fall. It is very unlikely I will have a flat panel so I need to do sky flats. I also want to take dark flats but I can’t wrap my mind around a simple question.
if I use Nina and the sky flats wizard, they won’t all have the same exposure so how do I match that ?
If I send the Stellavita instead, I can make them all the same exposure but they won’t all have the same ADU.
So how do people do that ? Is there just no way to take dark flats and sky flats ? Or is it so simple I’m just not seeing it because I’m looking for a complicated solution ?Dark current in most modern CMOS camera is 100% insignificant when compared to a 50% level of light signal, particularly at exposures well under a minute. You don’t need flat-darks and can instead use a quality bias master and achieve what you need very simply.
I have always been under the impression that this was true for broadband filters that have very short exposures on flats but on narrowband the flats are often around 10 seconds. These would still be properly calibrated with bias frames ?
I did a test of using darks vs. biases for various exposure times for lights (up to 5 minute exposures) just to prove to myself whether darks were beneficial (IMX455 sensor). For broadband frames (still waaaay under 50% histogram), there was no difference in SNR measurably or visibly between darks vs biases. For 5 minute 3nm narrowband sub, where the signal was barely just above the noise floor, there was a measurable benefit to using darks vs. biases, which is why I always use darks for all my light frames where I’m controlling the exposure times.
But even if you flats were a few minutes, as long as the signal is near the 50% of histogram range completely swamping what little dark and read noise there is, there would be no measurable difference using darks vs biases.
If you sensor has amp glow, then the answer may be different.
Helpful Engaging
Rick Krejci · Jul 3, 2026 at 06:32 PM
Eric Gagné · Jul 3, 2026, 06:03 PM
Rick Krejci · Jul 3, 2026 at 05:38 PM
Eric Gagné · Jul 3, 2026, 04:18 PM
I’m going to send a wide field rig to a remote observatory in the fall. It is very unlikely I will have a flat panel so I need to do sky flats. I also want to take dark flats but I can’t wrap my mind around a simple question.
if I use Nina and the sky flats wizard, they won’t all have the same exposure so how do I match that ?
If I send the Stellavita instead, I can make them all the same exposure but they won’t all have the same ADU.
So how do people do that ? Is there just no way to take dark flats and sky flats ? Or is it so simple I’m just not seeing it because I’m looking for a complicated solution ?Dark current in most modern CMOS camera is 100% insignificant when compared to a 50% level of light signal, particularly at exposures well under a minute. You don’t need flat-darks and can instead use a quality bias master and achieve what you need very simply.
I have always been under the impression that this was true for broadband filters that have very short exposures on flats but on narrowband the flats are often around 10 seconds. These would still be properly calibrated with bias frames ?
I did a test of using darks vs. biases for various exposure times for lights (up to 5 minute exposures) just to prove to myself whether darks were beneficial (IMX455 sensor). For broadband frames (still waaaay under 50% histogram), there was no difference in SNR measurably or visibly between darks vs biases. For 5 minute 3nm narrowband sub, where the signal was barely just above the noise floor, there was a measurable benefit to using darks vs. biases, which is why I always use darks for all my light frames where I’m controlling the exposure times.
But even if you flats were a few minutes, as long as the signal is near the 50% of histogram range completely swamping what little dark and read noise there is, there would be no measurable difference using darks vs biases.
If you sensor has amp glow, then the answer may be different.
I never use flat darks but I want to make sure that everyone understands the importance of dark calibration for image subs. I’ve used both QHY600M and Moravian C1x-61000 pro cameras that use the IMX455 sensor and I see noticeable dark current—mostly in the form of warm pixels in all of those cameras. The reason that you aren’t seeing the effect of dark current is that you are probably using the statistical stacking filters to remove that stuff. There’s an argument that you get a cleaner image that way, but I prefer to subtract the dark current before stacking the data. With a master dark composed of around 16 frames used in a reasonably large stack, the noise contribution due to dark correction is very small. Unless you use statistical filtering, this stuff will be pretty obvious in a stacked result so I recommend being careful about completely ignoring dark current in any CMOS camera. Dark calibration is easy and if it’s properly done, it contributes very little to the noise in an integrated master so there is no reason not to do it.
John
Well written Helpful Engaging Supportive
John Hayes · Jul 3, 2026, 07:18 PM
Rick Krejci · Jul 3, 2026 at 06:32 PM
Eric Gagné · Jul 3, 2026, 06:03 PM
Rick Krejci · Jul 3, 2026 at 05:38 PM
Eric Gagné · Jul 3, 2026, 04:18 PM
I’m going to send a wide field rig to a remote observatory in the fall. It is very unlikely I will have a flat panel so I need to do sky flats. I also want to take dark flats but I can’t wrap my mind around a simple question.
if I use Nina and the sky flats wizard, they won’t all have the same exposure so how do I match that ?
If I send the Stellavita instead, I can make them all the same exposure but they won’t all have the same ADU.
So how do people do that ? Is there just no way to take dark flats and sky flats ? Or is it so simple I’m just not seeing it because I’m looking for a complicated solution ?Dark current in most modern CMOS camera is 100% insignificant when compared to a 50% level of light signal, particularly at exposures well under a minute. You don’t need flat-darks and can instead use a quality bias master and achieve what you need very simply.
I have always been under the impression that this was true for broadband filters that have very short exposures on flats but on narrowband the flats are often around 10 seconds. These would still be properly calibrated with bias frames ?
I did a test of using darks vs. biases for various exposure times for lights (up to 5 minute exposures) just to prove to myself whether darks were beneficial (IMX455 sensor). For broadband frames (still waaaay under 50% histogram), there was no difference in SNR measurably or visibly between darks vs biases. For 5 minute 3nm narrowband sub, where the signal was barely just above the noise floor, there was a measurable benefit to using darks vs. biases, which is why I always use darks for all my light frames where I’m controlling the exposure times.
But even if you flats were a few minutes, as long as the signal is near the 50% of histogram range completely swamping what little dark and read noise there is, there would be no measurable difference using darks vs biases.
If you sensor has amp glow, then the answer may be different.
I never use flat darks but I want to make sure that everyone understands the importance of dark calibration for image subs. I’ve used both QHY600M and Moravian C1x-61000 pro cameras that use the IMX455 sensor and I see noticeable dark current—mostly in the form of warm pixels in all of those cameras. The reason that you aren’t seeing the effect of dark current is that you are probably using the statistical stacking filters to remove that stuff. There’s an argument that you get a cleaner image that way, but I prefer to subtract the dark current before stacking the data. With a master dark composed of around 16 frames used in a reasonably large stack, the noise contribution due to dark correction is very small. Unless you use statistical filtering, this stuff will be pretty obvious in a stacked result so I recommend being careful about completely ignoring dark current in any CMOS camera. Dark calibration is easy and if it’s properly done, it contributes very little to the noise in an integrated master so there is no reason not to do it.
John
I agree with you, John, which is why I use them for all my lights. Another benefit using darks is that I believe it allows PI to be better at doing image cosmetic correction as well as straight dark current subtraction.
But for flats at 50% histogram, and usually under 30 seconds, bias works perfectly fine.
Concise
You would probably be fine without flat-darks or bias but I use bias because it’s really simple and helps at least a little. I’m more concerned about removing fixed pattern (bias) noise than dark noise.
Two reasons:
1) These very short exposures have little time for dark noise to accumulate beyond what is contained in the bias
2) With an average ADU count around 30,000 the dark noise is a much lower portion of the flat image than a light image.
Just be sure you either dither or turn off tracking while capturing flats so that rejection can occur.
Kevin
Well written Helpful Respectful Engaging Supportive
Kevin Morefield · Jul 5, 2026 at 12:09 AM
You would probably be fine without flat-darks or bias but I use bias because it’s really simple and helps at least a little. I’m more concerned about removing fixed pattern (bias) noise than dark noise.
Two reasons:
1) These very short exposures have little time for dark noise to accumulate beyond what is contained in the bias
2) With an average ADU count around 30,000 the dark noise is a much lower portion of the flat image than a light image.
Just be sure you either dither or turn off tracking while capturing flats so that rejection can occur.
Kevin
Do you consider 9-10 seconds very short exposure ? Those flats on narrowband filters are the ones I can’t wrap my head around. Does a nearly 0 seconds bias do anything on a 10 seconds flat ?
Eric Gagné · Jul 5, 2026 at 12:16 AM
Kevin Morefield · Jul 5, 2026 at 12:09 AM
You would probably be fine without flat-darks or bias but I use bias because it’s really simple and helps at least a little. I’m more concerned about removing fixed pattern (bias) noise than dark noise.
Two reasons:
1) These very short exposures have little time for dark noise to accumulate beyond what is contained in the bias
2) With an average ADU count around 30,000 the dark noise is a much lower portion of the flat image than a light image.
Just be sure you either dither or turn off tracking while capturing flats so that rejection can occur.
Kevin
Do you consider 9-10 seconds very short exposure ? Those flats on narrowband filters are the ones I can’t wrap my head around. Does a nearly 0 seconds bias do anything on a 10 seconds flat ?
Yes, I find that the Bias and 10 seconds aren’t that different. Test for yourself. Take a bias sub and a 10 seconds dark sub, zoom in on a hot pixel in Pixinsight in the bias. Then probe with your mouse to see what ADU level that hot pixel reads. Then look at the exact same pixel on the 10 second dark and see what its ADU level is.
In the big picture, it’s not practical to shoot matching flat darks and bias is better than nothing. But really, nothing works OK too due to hot pixel rejection.
Helpful Engaging
I have my gear at Starfront, and last night I experimented with this for the first time.
I do have a flat panel, but I needed to take some darks. My flat panel is not light-tight, but the room at night, when the roof is closed, is very very dark. I simply set my filter wheel to HA and that took care of the last little bit of light, it worked great.
Starfront does offer to cap your scope around the full moon. So if you need a library of dark-flats, darks, or new bias frames, most people just schedule them for that night.
Well written Helpful Respectful Concise Engaging Supportive
Eric Gagné · Jul 5, 2026 at 12:16 AM
Kevin Morefield · Jul 5, 2026 at 12:09 AM
You would probably be fine without flat-darks or bias but I use bias because it’s really simple and helps at least a little. I’m more concerned about removing fixed pattern (bias) noise than dark noise.
Two reasons:
1) These very short exposures have little time for dark noise to accumulate beyond what is contained in the bias
2) With an average ADU count around 30,000 the dark noise is a much lower portion of the flat image than a light image.
Just be sure you either dither or turn off tracking while capturing flats so that rejection can occur.
Kevin
Do you consider 9-10 seconds very short exposure ? Those flats on narrowband filters are the ones I can’t wrap my head around. Does a nearly 0 seconds bias do anything on a 10 seconds flat ?
If you don’t use either a bias or dark to calibrate your flats, they won’t correctly flat calibrate light frames that have been dark subtracted. At 10 seconds or even 30 seconds, there is so little dark current above the bias level that it is truly in the noise when your exposure it at 50% histogram.
Rather than believing and asking since you’ve already gotten your answer, I encourage you to go ahead and try and compare even a 60 second close to 50% histogram light image calibrated with a bias and with a dark. Zoom in 400% and blink between the 2.
Helpful Concise
Eric Gagné · Jul 3, 2026, 04:18 PM
I’m going to send a wide field rig to a remote observatory in the fall. It is very unlikely I will have a flat panel so I need to do sky flats. I also want to take dark flats but I can’t wrap my mind around a simple question.
if I use Nina and the sky flats wizard, they won’t all have the same exposure so how do I match that?
If I send the Stellavita instead, I can make them all the same exposure but they won’t all have the same ADU.
So how do people do that ? Is there just no way to take dark flats and sky flats ? Or is it so simple I’m just not seeing it because I’m looking for a complicated solution?
When I do my calibration frames, I tried the doubled over T-shirt and such. I quickly saw the need for a dedicated flat panel. Sky flats vary by the brightness of the Sun, and are by their nature variable for reference frames. And I use NINA’s Flats Wizard. I got a real flat panel I can use with the Flats Wizard, but got mine before the automated panels were prevalent. Now, only a few make one for larger aperture telescopes. But many are available for <4” scopes.
An automated flat panel is the obvious best choice for remote operation. It can automate your calibration frame collecting and NINA does automatic adjusting of the brightness then runs your set number of Flats, Darks. And can be used as an automated cap for your telescope. A definite bonus for us remote operators (Even though my remote operation is just 75’ distant, it is fully remote capable.) So my manually operated Flats Panel is fine for me.
I will do a library of Bias, Flats, and Darks and use them over and over for up to 6 months. Because they are reference frames I found it works fine for me. I used to do my Darks and Bias at the end of a nights collecting and even made a Dark Filter in my EAF so I could program the function at the end of a session as Filter Number 8 in my wheel. Worked great for my smaller sensored OSC and for MONO cameras. Up to my ASI1600MM Pro. But now that I have reverted to my OSC, an ASI2600MC Pro, I don’t use filters anymore, just one of 2 light pollution filters in a filter drawer. Probably the best camera I have ever had.
But I use a library of Flats, Darks, and Bias frames for calibration frames. I’ve found it works fine for me and the way I prefer to do my Astrophotography. In reality, things do not change enough from one night to the next to warrant doing the reference frames over every night. Just library files work fine for me.
But for your away rig, I would heartedly recommend you consider one of the automated Flats Panels. And work with your equipment for a few months to make sure everything is working to your hopes and expectations before shipping it away.
I still prefer my walking distance remote and idiot filters. If it takes time to straighten anything out it is usually because I got side-tracked on my way, like a bathroom stop. 🫢
Helpful
I absolutely agree that an automated flat panel is best, especially on a remote rig. I have one on my 5” refractor and I haven’t ever had a problem with it. It is nice that I can take fresh flats every night, and with a little bit of code, I can stack them and file away everything before I get to my desk in the morning.
But that panel, for a 5” scope, was fairly expensive. I’ve been needling a buddy of mine to get one for his 9.25”, but it is crazy expensive.
So I agree that for any smaller scope, it’s a no brainer… but there just aren’t a lot of good options for larger scopes.
My friend with the 9.25”, I’d buy it from him if he had a flat panel or it was affordable, but alas, he is super cheap
(I know he’s reading this at some point, hehe)
Well written Engaging
Flat panels are certainly nice, but I have a remote rig at SFRO (10”) and DVO(8”) and for both just use sky flats without issue. I have a script (using Nina’s twilight sky flat command) that is more automated than Nina’s flat wizard…it slews to the side away from the sun at the time relative to sunrise or sunset I choose, takes all the flats in the order and histogram level I want and automatically adjusts exposure of each flat (even within the same filter) to maintain the desired levels. Can be automated easily for morning or evening flats. So not really any more trouble than a flat panel.
Again, if I had smaller scopes I’d probably have a flat panel, but they add weight and area to catch the wind. So sky flats it is.
Well written Helpful Respectful Concise Engaging
I’ve always done automated sky flats and it works well and the best part, it’s totally free. The only problem you can run into are clouds. To guard against that I’ll run an automated set in the evening and one in the morning. If I have clouds on both ends (which is extreamly rare) I’ll do them next session.
Well written Helpful Respectful Concise Engaging Supportive