What went wrong

[phodso]

Hi all having spent most of last night and the wee hours of this morning getting this data was slightly dissapointed with the results.

It is a stacked image of countless 2 minute exposures. Darks, bias and flats taken. (Although not overly sure how to take flats).

It was also my first attempt at guiding.

All comments and recommendations are appreciated as I am new to this imaging nightmare.

[Bizibilder]

Could you post one of your flats? - I suspect that they are the problem.

[sergeC14]

I fear flats may be the problem here... The best way for us to help you is to post a flat and a raw sub image. For flats, I'll give you a trick. The best way to acquire flats is to use a "flat screen" which can be expensive or a light box which you can make yourself if you have DIY skills.

Now, I you have none of that, at dusk, or dawn, aim your setup at the pale blue sky overhead, and adjust your exposure so that the resulting flats have the proper intensity. Now if you are acquiring color images with DSLR, remember to acquire them in black and white, because using blue flats, will affect the final colour or your image.

I do this with an OSC CCD caméra on C14 EdgeHD and it works quite well for me. You can read a post here:

Using a large format color CCD camera | Amateur Astronomy

It talks about colo CCD postprocessing and flats, but I think some of it might apply to DSLR as well.

In any case, I think you're getting alot of vignetting, it might be worth it looking closely at your setup as to what is causing this.

Regards

Serge

[phodso]

Managed to open the flat eventually.

I suspect you are right about this. Having looked at it I would have expected a circulare gradient getting darker towards the corners.

I assume this is over exposed.

They were taken at about 3am using the t-shirt method at the same exposure time as the lights.

If what I am saying is right then is there an idiots guide to flats?

[RikM]

I fear flats may be the problem here... The best way for us to help you is to post a flat and a raw sub image. For flats, I'll give you a trick. The best way to acquire flats is to use a "flat screen" which can be expensive or a light box which you can make yourself if you have DIY skills.

Now, I you have none of that, at dusk, or dawn, aim your setup at the pale blue sky overhead, and adjust your exposure so that the resulting flats have the proper intensity. Now if you are acquiring color images with DSLR, remember to acquire them in black and white, because using blue flats, will affect the final colour or your image.

I do this with an OSC CCD caméra on C14 EdgeHD and it works quite well for me. You can read a post here:

Using a large format color CCD camera | Amateur Astronomy

It talks about colo CCD postprocessing and flats, but I think some of it might apply to DSLR as well.

In any case, I think you're getting alot of vignetting, it might be worth it looking closely at your setup as to what is causing this.

Regards

Serge

I thought DSS only uses the luminance information from flats so colour is irrelevant?

[RikM]

Managed to open the flat eventually.

I suspect you are right about this. Having looked at it I would have expected a circulare gradient getting darker towards the corners.

I assume this is over exposed.

They were taken at about 3am using the t-shirt method at the same exposure time as the lights.

If what I am saying is right then is there an idiots guide to flats?

Take them in the daytime or with a lightbox. A laptop screen showing a blank white page in notepad can be used as well. Set the camera to Av mode and it will adjust the shutter speed automatically to give you the correct exposure. It is most important that you don't adjust the set up (camera orientation etc) or the flats won't work properly. My Av exposures normally come out at about 1/100th sec or so at ISO800.

[sergeC14]

Upon looking closely at your picture, I then went to check what scope you were using... A nice big Newtonian opened at F/D 4. In the corners and edges (especially to the right), You can see that coma is playing havoc with your stars. Coma being, a common thing with very open Newtonians, you might consider using a field corrector. I'm no sure which, but look it up there are quite a few you can use.

Also, Is your scope well collimated? Using a laser collimator or a cheshire eyepiece make the procedue a breeze.

As or vignetting, look at your focuser, it might be what's causing it. Anyhow, don't hesitate to post picture of your setup, I am sure there are plenty of very talented people here who can help you.

This is very interesting post a it can potentially apply to lots of people.

Serge

[ollypenrice]

It's darks, not flats, that have to have the same exposure time as the lights! Here's your problem, at least.

Your flats should be exposed till the peak of the histogram lies a third to two thirds of the way to saturation. A glance at the histogram on the screen will confirm this and it isn't critical within the range stated above.

It is certainly easiest to use an EL panel for taking them, as Serge says. All other methods can have side effects though they can work in many cases.

Olly

[sergeC14]

There you have it!!!! You used the same exposure time as your subs? Your flats are saturated , probably 65535 ADUs if your camera has a 16 bit ADC, check it out!!!

Ok, so you are using a white Tshirt over the scope aperture and then probably using a light. You need to use short exposure times (around a few 10th of second) so that your flats are not saturated. Now, the good news are that your subs are probably fine, you just need to redo the flats...

Serge

[sergeC14]

Rik,

I use an OSC CCD camera, and, as I have a large tube, I use the blue evening or ealy morning sky to make my flats. The OSC camera has a bayer matrix, so, blue pixels levels (or luminence) is higner that red or green. because the subs are divided by the flats, during calibration, the images blue pixels end up beeing "divided more" by say the green or the red pixels. So at the end the colours are way off if "blue flats" are taken "as is". I'm using Maxim for post processing, and it is what I have noticed.

Rgds

Serge

[RikM]

Rik,

I use an OSC CCD camera, and, as I have a large tube, I use the blue evening or ealy morning sky to make my flats. The OSC camera has a bayer matrix, so, blue pixels levels (or luminence) is higner that red or green. because the subs are divided by the flats, during calibration, the images blue pixels end up beeing "divided more" by say the green or the red pixels. So at the end the colours are way off if "blue flats" are taken "as is". I'm using Maxim for post processing, and it is what I have noticed.

Rgds

Serge

I think I understand but I never heard this before. I used to take blue sky flats with t-shirt over the scope and my DSLR. Using deep sky stacker I didn't notice a colour balance problem but it could be something to watch out for. Thanks.

Using sky illumination would often introduce a gradient in my images so I found much better results with a white laptop screen and now I bought an EL panel which is much easier to use. But not so cheap for a 14" Still, setting the DSLR to Av and checking the histogram makes sure the overall flat levels are correct.

[Photosbykev]

I also thought that the colour was irrelevant taking flats as the master flat is a greyscale image that represents the luminance??

[phodso]

Upon looking closely at your picture, I then went to check what scope you were using... A nice big Newtonian opened at F/D 4. In the corners and edges (especially to the right), You can see that coma is playing havoc with your stars. Coma being, a common thing with very open Newtonians, you might consider using a field corrector. I'm no sure which, but look it up there are quite a few you can use.

Also, Is your scope well collimated? Using a laser collimator or a cheshire eyepiece make the procedue a breeze.

As or vignetting, look at your focuser, it might be what's causing it. Anyhow, don't hesitate to post picture of your setup, I am sure there are plenty of very talented people here who can help you.

This is very interesting post a it can potentially apply to lots of people.

Serge

Collimation carried out when setup and star pattern checked.

The vignetting looks worse in the processed image then in the lights. I am assuming that it mainly caused by using a 1.25 camera nose piece (have to use that setup due to only having a 1.25 lp filter).

A coma corrector and 2inch lp filter are the next items on my wish list (Another purchase to get signed off).

[Ags]

If it was a choice between severe vignetting and LP, I would choose the LP!

[sergeC14]

Rik,

that's the thing! a laptop screen would be way too small for my setup...

But, using a white cloth + a dusk or dawn sky, that might work! I never thought of that, I'll try that thanks!

That said, trying to do flats with an Atik 11000-CM on the sky, you have to wait for the light to be "just right" or ugly columns will appear on the flats. You end up with flats ADU being twice between the first flat to the last one. Very tricky...

Also, if you do not cover the aperture of the scope while the flat is being downloaded to the PC (takes 30 sec), then you end up with a gradient, because the masked lines of the KAI-11000 interline CCD do let some light through!!! So, the "hidden" pixels at the time of dowload get some charge, with the last ones to be encoded getting 30 sec worth while the first lines get very little...

Amazing is not it? And the worst part is, is not written anywhere;)...

Rgds

Serge

[phodso]

Thanks for all the replies they have been very informative. Obtaining a new set of flats might be possible as the scope went to bed with the focusser locked in place the only problem would be getting the camera orientation correct.

Is it possible to set it to various angular positions and generate an average flat, or do I just write this one off to experience.

[RikM]

Thanks for all the replies they have been very informative. Obtaining a new set of flats might be possible as the scope went to bed with the focusser locked in place the only problem would be getting the camera orientation correct.

Is it possible to set it to various angular positions and generate an average flat, or do I just write this one off to experience.

Don't try a few different positions, just try and get it as close as possible to how you had it for this image and have a go. It might work if you are lucky, if not you haven't lost anything.

[Dave7a]

Hmmm....Intersting read! Could someone explain why the camera orientation would make any difference to the image it produces? I'm struggling to understand that?

[daz]

Dave - the flats are designed to remove imperfections in the light:- dust bunnies, vignetting and so on.

If the camera orientation is changed between the lights and the flats, then the resultant images will not line up and the processed result could actually introduce more artifacts

[Shibby]

But, using a white cloth + a dusk or dawn sky, that might work! I never thought of that, I'll try that thanks!

My best ever flats were taken with the morning twilight sky (no T-shirt). Although the second time I tried it they were hopeless for some reason. Maybe the half-asleep, zombie-like state of consciousness after an all-nighter has something to do with it!

[Dave7a]

Ahhhh... I get it now! So it's the change in orientation between the camera and the scope that we're talking about, not (as I was struggling with) the orientation of the camera in it's own '3D space'. Sorry for being a bit slow I was imagining a camera, pointing at the sky, with a lens cap on, wondering what difference a change in orientation could make. lol I'm learning! Thanks for clearing the fog Daz

[phodso]

Right guys have had a chance to recapture the flats, and there is an improvement in the gradient although far from perfect.

I also used PS to create a dummy flat (If thats what you would call it) and both images are attached.

Thanks for all the assistance and hopefully this thread has been of use to others on this steep learning curve. I certainly have learned that you check the flats before breaking the kit down.