How to take Flats that work? (Canon EOS M50)

Hi all,

Been taking DSO photos for about 6 months now, and I still am not getting great results from my flats.

My images still have donuts which are hard to remove in post-processing.

Does anyone have a technique that would work better?

How I do it:

Bring gear inside but leave all in original settings
Use an elastic to hold a single layer of a tshirt in-place over the scope
Set M50 Camera to "AV"
Hold my laptop over the t-shirt, flush with the lens
Take some photos and make sure that they are in the right part of the histogram (in my case 1/3 of the way)
Profit?

Thanks for any help.

Bad practice. Either use the sky at dusk as illumination source (with no clouds) or get an EL panel of the right size to use. No computer screens. And do it either before or just after your session. If eveything is left untouched then you'll be allright.

andrea tasselli:

Hi all,

Been taking DSO photos for about 6 months now, and I still am not getting great results from my flats.

My images still have donuts which are hard to remove in post-processing.

Does anyone have a technique that would work better?

How I do it:

Bring gear inside but leave all in original settings
Use an elastic to hold a single layer of a tshirt in-place over the scope
Set M50 Camera to "AV"
Hold my laptop over the t-shirt, flush with the lens
Take some photos and make sure that they are in the right part of the histogram (in my case 1/3 of the way)
Profit?

Thanks for any help.

Bad practice. Either use the sky at dusk as illumination source (with no clouds) or get an EL panel of the right size to use. No computer screens. And do it either before or just after your session. If eveything is left untouched then you'll be allright.

andrea tasselli:

Hi all,

Been taking DSO photos for about 6 months now, and I still am not getting great results from my flats.

My images still have donuts which are hard to remove in post-processing.

Does anyone have a technique that would work better?

How I do it:

Bring gear inside but leave all in original settings
Use an elastic to hold a single layer of a tshirt in-place over the scope
Set M50 Camera to "AV"
Hold my laptop over the t-shirt, flush with the lens
Take some photos and make sure that they are in the right part of the histogram (in my case 1/3 of the way)
Profit?

Thanks for any help.

Bad practice. Either use the sky at dusk as illumination source (with no clouds) or get an EL panel of the right size to use. No computer screens. And do it either before or just after your session. If eveything is left untouched then you'll be allright.

Thanks, good tips, I want a flatmaster for sure, but $300...

I use automation to take sky flats every morning for all filters. In my case I use Voyager and it executes a flats specific sequence that has all of the parameters set for taking and verifying my sky flats for me. It slews, snaps a photo to determine the ADU and adjusts the exposure until it fits the parameters I gave it. It dithers every flat to ensure proper rejection of any stars. 

After that, it parks the mount and warms up the camera and waits until the next night to wake up and run the whole show again.

I used to do 16 sky flats per filter, but after talking to Mike Selby he said he has used 9 sky flats per filter for years without any issues. I switched to that to save space and it's worked perfectly for all of my data.

Bill

Jon Rista:

Hi all,

Been taking DSO photos for about 6 months now, and I still am not getting great results from my flats.

My images still have donuts which are hard to remove in post-processing.

Does anyone have a technique that would work better?

How I do it:

Bring gear inside but leave all in original settings
Use an elastic to hold a single layer of a tshirt in-place over the scope
Set M50 Camera to "AV"
Hold my laptop over the t-shirt, flush with the lens
Take some photos and make sure that they are in the right part of the histogram (in my case 1/3 of the way)
Profit?

Thanks for any help.

Don't bring your gear in. Wait till the sun is up, with a clear blue sky and get flats then. If you have a relatively bright and even overcast sky, you can try to put a diffuser of some kind in front of the aperture. But, don't bother with dusk flats. At dusk the sky is changing all the time, and you can get gradients. A clear, flat blue sky during the day (or morning, as long as the sun is up and the sky is even) is one of the flattest sources of light you can find.

I never thought oh doing flats before imaging! that can work, IF I get clear skies, which seems to be rare where I live.

As long as the skies aren't changing in intensity a lot, wrap a t-shirt around the aperture and go for it even if its overcast. Before, after your session, doesn't really matter. The key is, the DAYTIME sky can be a nice, bright, flat light source that won't introduce gradients, etc. Dusk flats were always problematic for me with changing intensity and changing gradients, and then of course the eventual appearance of stars. Daytime blue sky flats, or diffused overcast flats, have always worked much better for me. The one thing about diffusing...is that changes the distance of the light source, which can affect how any internal reflections behave. The best light source, is a daytime empty blue sky, far enough away from the sun not to exhibit any gradients. It can be exceptionally flat.

The automation capabilities these days are really powerful for getting perfect sky flats, every time. No more T-shirts, no more panels, no more waking up super early. You can take Dusk or Dawn sky flats (or both!) very easily these days. The sequences are specific to flats, and do it all for you. 

Example from Voyager:

So I gave it a shot and tried to do some flats pointint up at the evening sky, few clouds. 

What a strange phenomenon, all the images had a black bar at the bottom, like a gradient. No matter where the scope was pointing.

That got my head scratching

Jon Rista:

andrea tasselli:
Daytime sky isn't really flat (thanks the nearest star to us, the Sun) and we don't know what the OP is using for imaging. Wide FOV will reveal that sure enough. Small FOV, you'll get away with it, IF pointing well away from the Sun. I have always taken flats at dusk (specifically at sunset or shortly before) and NEVER had any issue. YMMV.

A daytime flat is significantly flatter than a dusk flat. You can easily point away from the sun enough that the blue sky is very flat. At dusk, you have constantly changing variables that are integrated into your individual frames. Some software allows you to normalize the average ADU count, but you can still have some pretty rouge gradients. You can also pick up stars (a near-impossibility during the day). A blue sky flat, even if it has a gradient, will be orders of magnitude smaller than most dusk/dawn sky flats, and you won't have the intensity shift of dusk/dawn flats.

Jon,

The software in use nowadays, does all of this for people. Picking up stars is not a problem. The automation slews and dithers as well. I would bet you $10 I can take better flats with Voyagers automation, than you can with a T-shirt futzing around in the daytime.

This image was taken using the same software, same method of flats, and is up on the "Sky" Facebook page of NASA and will probably show up soon as an APOD. That is "The Eyes" taken with a 1 meter CDK in Chile by Mike Selby.

What you are saying here, is old man get off my lawn levels of old school thought and point in time perspective based on what software BACK THEN did. We are way past this, my friend.

Jon Rista:

andrea tasselli:
Daytime sky isn't really flat (thanks the nearest star to us, the Sun) and we don't know what the OP is using for imaging. Wide FOV will reveal that sure enough. Small FOV, you'll get away with it, IF pointing well away from the Sun. I have always taken flats at dusk (specifically at sunset or shortly before) and NEVER had any issue. YMMV.

A daytime flat is significantly flatter than a dusk flat. You can easily point away from the sun enough that the blue sky is very flat. At dusk, you have constantly changing variables that are integrated into your individual frames. Some software allows you to normalize the average ADU count, but you can still have some pretty rouge gradients. You can also pick up stars (a near-impossibility during the day). A blue sky flat, even if it has a gradient, will be orders of magnitude smaller than most dusk/dawn sky flats, and you won't have the intensity shift of dusk/dawn flats.

Jon,

The software in use nowadays, does all of this for people. Picking up stars is not a problem. The automation slews and dithers as well. I would bet you $10 I can take better flats with Voyagers automation, than you can with a T-shirt futzing around in the daytime.

This image was taken using the same software, same method of flats, and is up on the "Sky" Facebook page of NASA and will probably show up soon as an APOD. That is "The Eyes" taken with a 1 meter CDK in Chile by Mike Selby.

What you are saying here, is old man get off my lawn levels of old school thought and point in time perspective based on what software BACK THEN did. We are way past this, my friend.

With a blue daytime flat, you don't use a diffuser. Its just the blue sky with an open aperture. If you try an overcast sky, then a diffuser would be necessary, but not with a blue sky. Its exceptionally flat.

If software can take care of stars for you and all of that, great. I have always had gradients in my flats at dusk (or dawn). Does the software also correct the gradient? Is a corrected gradient better than a truly flat flat? If you are correcting a gradient, then that would be multiplying and dividing the signal, so I wonder about PRNU correction and how good it would be. I am aware that modern software can automate flat acquisition, and do things like automatically adjust to maintain an average ADU level. The image you shared exhibits a dark ring artifact, around the center field, and looks slightly brighter at the top than the bottom. Is that due to the flats?

I still disagree that a dusk sky has less gradient than a blue sky. ;) 

With that said, I won't continue to debate the point. If someone wants to get their flats at dusk, I am certainly not going to tell them they are wrong, or are an old geezer for saying so. I'm sure voyager could make short work of open aperture daytime blue sky flats too, and I think they can be just as viable an option for those who have troubles with dusk flats. Beyond that, I don't think this all warrants any deeper debate. I've offered my 2 pence, and people can decide on their own what kind of flat to get.

@Jon Rista 

https://apod.nasa.gov/apod/ap240320.html?fbclid=IwAR05zFBbTojMRczUhOz-2yzEGU4Tth1pFwGNvnkgORQr7aYs9fImzCZ-NL8

I knew that would win APOD.

The flats timing and direction are all set by Voyager. It points away from the sun and determines when to start based on the degrees under the horizon the sun is. Then it samples the flats from the sky until they are within the parameters set and then begins to take flats and checks each as they come in. It tweaks the exposure time as it takes the flats little by little to ensure they remain in the percentage threshold I've given. It dithers for each flat to make sure any stars caught in the flats can be rejected.

This flat sequence starts up when my main imaging sequence (RoboTarget Scheduler) has completed for the night (end of nautical night).