Am I underexposed? Imaging help.

[focaldepth]

There are many exposure questions on SGL but here's another.

Where am I so far. I intend to end up CCD imaging but for now I am playing with DSLR. I want to get the techniques right and not worry about the actual images just yet.

Working under quite bad LP I have been amazed that I can actually pull out some galaxy details from a sea of mush and noise.

Here is one sub of M81. Shown in DPP at 100% and Pixinsight.

This tells me I am under exposed.

The shot was at ISO800 and 5min.

LP filter fitted.

It was taken last week (very few clear nights since I got my scope on a pier) so is obviously in summer skies.

From the camera histogram I thought I had exposed for as long as I dare but looking on the PC it looks as though I can go a lot further.

So 3 questions.

1) Am I correct and I need much more exposure, despite what the camera tells me?

2) Is that amount of noisy background gradient normal, as I am finding it difficult to process out?

3) Can you spot anything else I am doing wrong?

Go on, don't pull punches.

[jgs001]

That looks about normal for a deep sky image. What you want is many 5 minute ISO800 images, and combine them using something like Deep Sky Stacker. Then the fun starts... The combined image will have all the data buried in the black point (or at least, it'll look like it on a histogram display). You have to drag it out by processing the data. It does look like you're off to a good start there though. As for the gradient, LP filters can help a lot. I believe there is something (Dynamic Background Extraction) in Pixinsight... but I've never used it or tried it... to deal with it.

[roundycat]

what exactly was it that told you the picture is underexposed? Depending on the f ratio of the optical system I would think 5m is about right. If you expose any longer what happens to the bright stars? If they get bloated and saturated the exposure is too long. If the bright stars are nowhere near saturated try upping the exposure to 10m and study the result. Whatever gives you the best result, shoot at that exposure and don't mix up different exposure lengths.

You do not have a background gradient that I can see. The bright core of the pic is down to optical considerations and can be neutralised by proper flat fielding.

The post processing looks a bit suspect with that bright pink background but you need to attend to the initial problems first before you can expect to produce a good result.

Check exposure by measuring a bright star, don't let it saturate.

Calibrate the sub frames as best you can. That means dark subtraction and proper flats. Calibration makes more difference than you will ever believe.

Dennis

[focaldepth]

Sorry I was a bit lacking in explaining what I had posted.

The image in Pixinsight has simply had the auto adjust in screen transfer function applied. (I.e. to the display but not the image)

The one in DPP is simply viewed at 100% with nothing done.

This is just ONE sub from many over a few nights.

This image has had NO stacking or PP. (I have stacked with DSS using bias, dark, flat and dark-flat(flat-dark) but not posted this.)

Looking at the histograms in DPP and PI, the brightest part is well away from the right-hand end, even though it was near the right on the camera.

Am I reading the tools correctly and they are telling me that I can go for longer exposures? To my thinking, longer means more faint data out of the 12(14?) bit A/D quantisation noise on the DSLR.

[harry page]

Hi

Most images carry colour casts , particularly with DSLR cameras

You need to remove the gradients and balance the background

Have a quick look at this video http://www.harrysastroshed.com/pixinsighthomeinter.html the sword of orion one

Should get you going

Harry

[roundycat]

place your cursor over a bright star and measure the level in ADU. For 12bit anything approaching 4096ADU is saturated. 14bit would be 16,384ADU. You probably have to interact with your software to do this. It can't tell you by itself. Forget the histogram, it is not sufficiently accurate.

Dennis

[focaldepth]

place your cursor over a bright star and measure the level in ADU

Doh! Obvious, why did I not think of that.

Great call on that one, thanks.

[focaldepth]

And thanks Harry, I have watched many of the Harry tutorials. They have taught me all I know about Pixinsight.

It is far from obvious how to PI but I am getting used to it.

Often things seem strange at first but then often when they click I can see why they were done that way.

Soon PS will seem counter-intuitive.

[focaldepth]

Just for interested viewers, turns out the stars are 15300 ish and the centre of M81 is 6000 in green channel.

So for the 14bit 7D that is almost saturated.

The background is 2400.

So with help, the answer to my worry is that it is just about right for exposure.

Although doubling the time would bring the galaxy up to 12000 and white out some stars.

I think I will stick with 5mins for now.