So this flat frame malarky.

[ryandsimmons]

Given the rubbish skies I am trying to redo the images that I took before the skies clouded over.  I have Darks, I have plenty of actual images, but what I have an issue with is vignetting according to the helpful people here.

So I am going to take some flats tonight and add them into the stack to see what happens.  I am aware that they should be taken at the same time as the rest etc, but I figure some practice here won't hurt.

Can I just clarify that I am (planning) doing this right?

1) I will put camera in fridge to replicate the cooler temperatures of when I did the imaging.

2) Point camera at white surface (white screen of ipad probably.)

3) Set ISO etc to match images.  (ISO 800, manual settings.)

4) Take image at...xxxx... exposure time.  *

5) Check histogram. ** Histogram should be a mountain 1/3rd across the screen?

6) Take a few of these.

Does that sound right?  The two main questions I have are:

*-should exposure be anything in particular?  I really have no idea if it should be 1/1000 of a second or 3 minutes?

**-how do I check the histogram.

I have a Canon EOS1000D.

Many thanks

[Leveye]

If you don't take them at the same ambient outside temperature and sensor temp as the lights you're wasting your time.

[DaveS]

You can do the flats any time, all that matters is that the optical train in the same. A flat white screen will do, some people do them in daylight with a couple of layers of white t-shirt over the object lens.

Set you camera to AV and let it set the exposure, though you could dial in a bit of exposure compensation.

For the histogram press the > button on the camera and use info to select what is displayed.

HTH

[bruno]

Have a look at this,it may help: http://deepskystacker.free.fr/english/faq.htm#lightdarkflatoffset

[DaveS]

Only the Darks must be taken at the same time and temp, the flats are just mapping the optical imperfection in the light path.

[Adamar98]

I do my flats straight after a session of imaging. I have forgotten to do this in the past and taken my flats on another day. It did help with vignetting but not with the dust bunnies. I'm not to sure if temperature has anything to do with them. So I don't think you'll be wasting your time. It's worth a try.

[lensman57]

Given the rubbish skies I am trying to redo the images that I took before the skies clouded over.  I have Darks, I have plenty of actual images, but what I have an issue with is vignetting according to the helpful people here.

So I am going to take some flats tonight and add them into the stack to see what happens.  I am aware that they should be taken at the same time as the rest etc, but I figure some practice here won't hurt.

Can I just clarify that I am (planning) doing this right?

1) I will put camera in fridge to replicate the cooler temperatures of when I did the imaging.

2) Point camera at white surface (white screen of ipad probably.)

3) Set ISO etc to match images.  (ISO 800, manual settings.)

4) Take image at...xxxx... exposure time.  *

5) Check histogram. ** Histogram should be a mountain 1/3rd across the screen?

6) Take a few of these.

Does that sound right?  The two main questions I have are:

*-should exposure be anything in particular?  I really have no idea if it should be 1/1000 of a second or 3 minutes?

**-how do I check the histogram.

I have a Canon EOS1000D.

Many thanks

You need to take your at flats the end of a session at the same focus point as your lights. During an imaging session the point of focus shifts as the temprature changes untill a steady state is reached. If you do your imaging at lets say 5C during the evening and then do the  flats inside where the temp is 20C+ then they will not do you any good. There is also the problem of the focus tube being disturbed while the scope is being moved. If you change a filter you need to redo the flats again. Getting good flats is almost as difficult as getting good lights so do not do short cuts.

A.G

[knobby]

I'm sure you're all correct... But he did say the practice will be good, and you can't argue with that :-)

[Thalestris24]

Hi

I suppose one could argue that flats aren't absolutely necessary but they can be quite useful - or they can cause more trouble than they're worth! If you have no dust bunnies but have vignetting then it's possible to remove vignetting and other gradients during post-processing. If you have dust bunnies then maybe it's best to remove them. Flats only make them appear to disappear so what might have been recorded in their positions wasn't. Also, if you don't take flats right then stacking with incorrect flats can introduce unwanted gradients and artefacts, and generally make a mess.

Having said all that I usually do take flats But sometimes it all goes horribly wrong! So I guess the moral of the tale is - yes, take them but be sure you do them properly

Just my thoughts

Louise

[dph1nm]

I think the effect of temperature and focus on flats is somewhat overstated. Professionally speaking, I can tell you that flats are generally taken in twilight, or during the day using the inside of the dome - in neither case is the telescope precisely in focus or at the same temperature as during the night- but they still work.

As an amateur I can say that nothing has made a greater impact on my photos that the addition of flats - I would always use them. Trying to flatten in software does not work anywhere near as well, and usually means subtracting the background (which is incorrect, as flats need to be divided into the data).

NigelM

[laser_jock99]

Something I've been experimenting with lately is a 'flats library'. Whilst this won't help with the aforementioned 'dust bunnies' (you should keep your optics as clean as possible anyway) it could help with the Big One- i.e. vignetting. Vignetting won't change much if the set up is the same from session to session. If you can remember to keep the camera in the same orientation (I align my cameras exactly parallel to the focuser knobs every time) then your 'flats library' could be used to control vignetting at least. You'll need to average out any dust specks in the library images as these aren't constant. Clearly there is no short-cut to 'proper' flats if you have the time & patience.

On the subject of flats it's interesting how wildly they vary from instrument to instrument- some stretched range flats from my 'flats library'

Fast Newtonians seem to give nice, evenly illuminated, dust free flats

10" Newtonian @ F2.9

12" Newtonian @ F2.9

On the other hand my refractor flats are oddly shaped, unevenly illuminated and suffer from dust (which appears to originate from the field flattener/reducers).

ALTAIR STARWAVE ED80

SKYWATCHER ED80 FLAT

Camera lenses are also uneven.

Nikon 50mm @ F1.4

Even a pro lens - Nikon ED180 F2.8 (@ F4 in this image) looks naff.

[ryandsimmons]

Cheers for all the info guys.

I tried last night with the camera in the fridge first (to match up with the cold night I took the pics) but I found the flats didn't really corrospond well.  Whereas the DSO images I have were pretty circular in terms of vignetting, I found the flats tended to be more dark at the bottom and lighter at the top.  I think that may have been the way I took them though. 

Still, practice is practice even if it doesn't succeed I suppose. I will give it a better try tonight.

[frugal]

Cheers for all the info guys.

I tried last night with the camera in the fridge first (to match up with the cold night I took the pics) but I found the flats didn't really corrospond well.  Whereas the DSO images I have were pretty circular in terms of vignetting, I found the flats tended to be more dark at the bottom and lighter at the top.  I think that may have been the way I took them though. 

One thing I have found is that you need to make sure that if you are using a light panel or a laptop/tablet screen; you need to make sure that the panel is square onto the lens / scope or you will introduce more light on one side than the other. I found that the best way to get a good result is to press the screen gently up to the end of the dew shield so that it is perfectly flat.

[Alien 13]

I dont use flats anymore all the info required to generate an artificial flat is in the final image, the problem i allways had is that dust bunnies would move during a session so i had to allways keep modifying any flats i took.

Alan

[ollypenrice]

If you don't take them at the same ambient outside temperature and sensor temp as the lights you're wasting your time.

This is quite incorrect. The temperature doesn't matter in the least, other than that the noise will be higher at higher temperatures. That, though, can be addressed by taking more. It matters for darks, but flats are entirely different.

Let's step back and think what we're doing with flats.

When we take our deep sky images (known as Lights) we take them through a telescope and filters and chip window. In a perfect world the telescope and filters and chip window would have no effect whatever on the image's illumination. In reality the telescope will give a brighter middle than corners. (This is called vignetting, from the French term for small oval portraits.) FIlters and chip window will have dust on them which will create shadows on the image (known as dust bunnies or doughnuts.) These are artefacts of our system and we want to remove them. We can do so by photographing just the artefacts (these photographs are our 'flats') and asking the software to dim or brighten the final image according to the dark and light patches seen on the flats.

In effect our flats photograph the imperfect light path through which our lights were captured. You must, therefore, recreate exactly that imperfect light path when taking flats. Same camera, same camera orientation, same filters, same point of focus (or your bunnies will contract or expand). It is the light path which must be duplicated, not the characteristics of the chip at the time. If you cannot recreate the same light path you cannot make useful flats.

To work properly, flats must be calibrated with dark frames like any other digital image, but in reality a master bias will be perfect as a dark for flats. A master bias is made by averaging about fifty dark frames of the shortest exposure your camera can take.

To take flats, photograph an even light source and expose till the histogram peak is about a third of the way to saturation (a third of the way to the right.)

A flat can look truly awful with dark corners, a light middle and dark blobs abounding. A flat corrected by an application of itself will simply look like a flat, featureless even surface. This is what happens to your images. The uneven lighting caused by your system errors is removed and yu get the result you'd get from perfect optics (which cannot ever be made.)

Olly

[ollypenrice]

I dont use flats anymore all the info required to generate an artificial flat is in the final image, the problem i allways had is that dust bunnies would move during a session so i had to allways keep modifying any flats i took.

Alan

This is true in the sense that all the defects are in the final image. But how do you gain access to this information? Information which you cannot discover is not information. The reason flats are useful is that the information they contain can be discovered because they contain no other information. Only the defects are present. But in a real image a dark patch could be real or artificial. How can you distinguish them? I don't see how you can. DBE in Pixinsight is very powerful for general gradients provided you have genuine background sky available throughout the image, but often you don't. And it can't deal with local effects like bunnies.

Olly

[Alien 13]

This is true in the sense that all the defects are in the final image. But how do you gain access to this information? Information which you cannot discover is not information. The reason flats are useful is that the information they contain can be discovered because they contain no other information. Only the defects are present. But in a real image a dark patch could be real or artificial. How can you distinguish them? I don't see how you can. DBE in Pixinsight is very powerful for general gradients provided you have genuine background sky available throughout the image, but often you don't. And it can't deal with local effects like bunnies.

Olly

I agree that some dark patches etc can be real useful data but i only extract the information that is obvious vignetting or clearly defined dust bunnies, it is by no means as good as a well taken flat and with some targets it can be difficult.

The main reason i do this is because dust in the camera migrates during a session because i keep forgetting to switch the ultrasonic sensor cleaning routine off  so every time the camera is powered down i get a new set of dust artifacts.

Alan

[dph1nm]

This is true in the sense that all the defects are in the final image. But how do you gain access to this information?

So you can dither and then median together the individual subs - this will produce a flat  (providing you have no scattered light getting into the camera).  You do have an issue that the s/n in the flat is then no better than in the final stack.  And if you are observing large objects you would need a very big dither, or to take subs with your object out of the field (which wastes good observing time of course).  You do however, get very flat images this way!

NigelM

[ollypenrice]

So you can dither and then median together the individual subs - this will produce a flat  (providing you have no scattered light getting into the camera).  You do have an issue that the s/n in the flat is then no better than in the final stack.  And if you are observing large objects you would need a very big dither, or to take subs with your object out of the field (which wastes good observing time of course).  You do however, get very flat images this way!

NigelM

I don't understand this. You cannot dither with respect to the uneveven illumination. When you dither you move the imaging train (and with it the bunnies and the vignetting). Yes, if you like you move the image 'underneath' these defects but dithering takes place on a pixel or sub pixel scale and this scale cannot possibly identify the uneven illumination caused by flats and bunnies which might be on a thousands of pixels scale. Or what am I missing?

Hi

Flats only make them appear to disappear so what might have been recorded in their positions wasn't. Also, if you don't take flats right then stacking with incorrect flats can 

Louise

If they appear to disappear they disappear. Isn't it that simple? If a defect is present and you calibrate it out so that it cannot be detected then it cannot be detected and does not exist. When you calibrate a blotchy image of a flat panel with itself you get a (near) perfect image of that flat panel.

I'm five foot five. If I stand in a one inch deep ditch I appear to be five fout four. If you measure the depth of the ditch and add it to five foot four you get five foot five. Now I appear to be five foot five. That is good because I AM five foot five!!! And so with flatfield calibration, surely? (I know that flats are operated by division but that arises from the need to find a baseline from the flat.)

Olly

[Thalestris24]

If they appear to disappear they disappear. Isn't it that simple? If a defect is present and you calibrate it out so that it cannot be detected then it cannot be detected and does not exist. When you calibrate a blotchy image of a flat panel with itself you get a (near) perfect image of that flat panel.

Olly

Hi  Olly

I know you are way more knowledgeable and expert than little me I'm a beginner and have only been imaging for a few months - much to learn. All I was saying was that, by interrupting the light path, dust bunnies must surely prevent something being recorded that otherwise would have been? It might only have been a star or two, perhaps in a big field of stars, but still. Using flats to remove the dust bunnies certainly does the job but only cosmetically as far as I understand it. As for gradients, well, that's a different story. Making proper flats will do a good job of removing unwanted gradients that exist in the optical train e.g. vignetting.  I'm not sure by how much vignetting can be minimised when using an APS-C size sensor. Standard T2 fittings don't seem to help much...

Cheers

Louise

[Alien 13]

Hi  Olly

I know you are way more knowledgeable and expert than little me I'm a beginner and have only been imaging for a few months - much to learn. All I was saying was that, by interrupting the light path, dust bunnies must surely prevent something being recorded that otherwise would have been? It might only have been a star or two, perhaps in a big field of stars, but still. Using flats to remove the dust bunnies certainly does the job but only cosmetically as far as I understand it. As for gradients, well, that's a different story. Making proper flats will do a good job of removing unwanted gradients that exist in the optical train e.g. vignetting.  I'm not sure by how much vignetting can be minimised when using an APS-C size sensor. Standard T2 fittings don't seem to help much...

Cheers

Louise

Hi Louise

That is an interesting point you make i am not sure at what point a dust bunny completely obliterates a star on the image but i imagine some must or at least distort it but in my images i cant say ive ever seen any obvious holes where the dust bunnies have been processed away.

I do however often switch my RA drive off between sets of subs to shift the image a little to compensate for any dust on the sensor.

Alan

[ollypenrice]

Sorry to be slow on this. Louise, the flats are more than cosmetic. They are an accurate measure of the dimming and brightening created by an imperfectly even illumination. They will correct this un-even illumination. They are, therefore, a scientific calibration and professional observatories use them. Flats are not used to correct total obscuration of some pixels. That has to be done by cleaning the chip! Sure, if you have black pixels because of contamination or because they are dead then anything you do to remove them is guesswork and cosmetic, but flats deal with dimmming rather than total obscuration.

Olly

[Thalestris24]

Sorry to be slow on this. Louise, the flats are more than cosmetic. They are an accurate measure of the dimming and brightening created by an imperfectly even illumination. They will correct this un-even illumination. They are, therefore, a scientific calibration and professional observatories use them. Flats are not used to correct total obscuration of some pixels. That has to be done by cleaning the chip! Sure, if you have black pixels because of contamination or because they are dead then anything you do to remove them is guesswork and cosmetic, but flats deal with dimmming rather than total obscuration.

Olly

Hi Olly

Thanks for the message! Yeah, I think I appreciate the purpose of flats better now I'm still such a rank beginner/amateur... I managed to do a little imaging last night (yay!) and have to do a set of flats later when it's dark.  Actually, forecast for tonight is also good - it's otherwise been terrible here for months now Plus, of course, it starts getting light not much after 2am at the moment! Fingers crossed the clouds stay away again.

Cheers

Louise

[lensman57]

Hi Louise,

Following from Olly's advice, dead pixels or coloumns can to a great dergree be corrected by " High Dithering " and then stacking a large number the subs using Sigma Clipping. This will also help a great deal to reduce noise in the capture but you really do need a lot of subs.

A.G

[Thalestris24]

Hi Louise,

Following from Olly's advice, dead pixels or coloumns can to a great dergree be corrected by " High Dithering " and then stacking a large number the subs using Sigma Clipping. This will also help a great deal to reduce noise in the capture but you really do need a lot of subs.

A.G

Hiya

Alas, I'm lucky if I can get 10 x 5min subs what with the window and Glasgow clouds . If it stays clear tonight I hope I'll be able to add to what I did last night - some of which were galaxies!  I've yet to get enough integration time to successfully image a galaxy - sigh. I got a 550mm pedestal mount recently which enables me to point higher in the sky . It was a lot of hassle getting Orion Optics UK to supply me with an azimuth post that would fit but it seems ok now.

Cheers

Louise