Stacking guide

[zief]

Hi,

I've come back to Astrophotography after years out, now with dSLR instead of hypered film. I've decided to be a bit more portable and to start with just want to get some good results with wide field shots. I didn't set my eq mount up properly on my first run out, and have some drift over my images, but each is only ~20 seconds anyway. I wanted to stack these, but I'm having a lot of difficulty getting a good final result. I'm very computer savvy and fairly image processing savvy, so this has just made it all the more frustrating! Following advice and guides, it seems largely centered on Deep sky stacker and Registax. However Deep Sky stacker always returns me with a blurred image as it can't have dealt with the motion properly, and Registax I can barely get to work! I'm in need of good advice on how to use the programs properly I guess or good alternatives?

Thanks for any help!

Joe

[iksose7]

Look forward to the responses here, i'd like a good step by step for DSS aswell. Just to be sure everything is being done properly.

[jgs001]

Hi Joe, welcome to SGL... DSS is the correct tool for wide field and deep sky images... Registax is much better suited to the moon and planets. If you have drift/trails within each frame, then DSS is going to struggle. It may be able to stack them if the drift is not too great, but the end result is going to be trailed. Can you post examples of a single image and the resulting stack ?

[bamus]

You can try IRIS software where you can manually allign images. It does look complicated but does much better job in my opinion

[Steve Ward]

Try here . . .

Steve.

[iksose7]

Just registered my images in RAW format and the result appears to only be a portion of the image. Any ideas what may be going wrong?

[jgs001]

A thin strip is normally an indication that the raw processing engine of dss has not been updated with the raw format for that camera.

[zief]

The Iris software looked promising, but can you recommend a process to follow for exactly what I want to do here? I spent hours 1 night a few weeks ago playing with it and wasn't getting anywhere fast!

An example of one of the original shots, and the final result after Deep Sky Stacker and a bit of an edit with saturation etc can be seen below hopefully. Hopefully not too low resolution to see. You can see Deep Sky Stacker has detected the movement as the foreground trees have been blurred, but I don't think it has done that great a job matching star to star as each seperate image is well focussed, and the resultant Milky Way is pretty blurred. Anything I can do to improve my process?

Thanks

[bamus]

The Iris software looked promising, but can you recommend a process to follow for exactly what I want to do here? I spent hours 1 night a few weeks ago playing with it and wasn't getting anywhere fast!

try this:http://www.astrosurf.com/buil/iris/tutorial3/doc13_us.htm

or this: http://www.astrosurf.com/buil/iris/roadmap/help2_us.htm

[zief]

try this:http://www.astrosurf...l3/doc13_us.htm

or this: http://www.astrosurf...ap/help2_us.htm

Cheers.

So if I don't have dark or flat field images I can skip those 2 steps? And skip them in the boxes in the pre-processing?

[rfdesigner]

I use Iris

I use the command line extensively as you keep a track of what it is you've done and I find it far quicker (button looks like)

>=

>=

Download the command list (pdf) from the same web page as you downloaded IRIS.

I use 'coregister2' as this aligns on the whole image and will take out image rotation and variation in distortion. You can vary how many stars it picks up by varying 'setfindstar'.. edit: and setregister, which allows you to vary the 'order' meaning, just left-right-up-down (1), +rotation (2), +curvature(3) and so on.

there are many different alignment options and it's worth playing.

Derek

[rfdesigner]

Cheers.

So if I don't have dark or flat field images I can skip those 2 steps? And skip them in the boxes in the pre-processing?

yes... it won't look quite as good but it will function.

[zief]

yes... it won't look quite as good but it will function.

Cheers for this mate, I'll have more of a play with it! Sounds like I might need to set a day aside!

Do you mean I should still create the dark and flat frames even though I don't have any actual ones from the shoot, and it will look better, or that I should have shot dark and flat frames at the time for better results?

[zief]

On another note, is there no current 'fairly' easy to use UI based program that might get some better results, or a better way to have Deep Sky Stacker align and process, or will I have to hit the books with Iris as it were?

[rfdesigner]

Cheers for this mate, I'll have more of a play with it! Sounds like I might need to set a day aside!

Do you mean I should still create the dark and flat frames even though I don't have any actual ones from the shoot, and it will look better, or that I should have shot dark and flat frames at the time for better results?

No..

I mean you can stack uncorrected frames and still get something worth playing with.

i.e.

work c:\your working directory path

load rawframes_1

setfindstar 20

setregister 2

coregister2 rawframes_ @coreg_ 10

composit @coreg_ 2 1 1 10

assuming your raw frames are called 'rawframes_1, rawframes_2... rawframes_10' then you will end up with a composit image made from all your raw frames. Because I've pulled up the 'sigma clipping' routine and set sigma to 2 it will ignore wildly outlying values for any one pixel but keep the majority, this takes out unwanted artifacts. If you are using more images you need to increase this number a little.

If coregister is failing it's usually a good idea to increase the 'setfindstar' value, you want it to be picking up maybe 30 stars, not 3000.

If you type any one of the commands above just on its own IRIS tells you what the parameters you need to fill in are and what they do.

I've chosen to give the 'coregistered' frames @coreg_ because there is a rREMOVE@ function that can then clean up your hard disk when you're done.

Derek