DSS and wide field image stacking

[fwm891]

Hi, Been taking a few wide angle shots of the MW lately with a static camera (laid on its back pointing skywards!)

When I try to stack the RAW files in DSS (or AA5) the central area looks reasonable but the outer areas look like I've zoomed the lens while the exposure is underway.

The only way I've got round this is to make each image a layer in photoshop and then use the transform/distort functions on each layer (with lots of minor tweaks for each layer) until I get a null (setting lowest layer as 'normal' and each layer above as a 'difference' blend mode switch one layer on adjust it, turn it off then the next layer etc...

Anyone know what I'm doing wrong

[fwm891]

Image to go with above (edit, edit, edit....)

[rfdesigner]

You need to stack with an algorithum that can: align rotate and stretch.. i.e. highest order of matching.

I'm pretty sure DSS can do this. I use IRIS and I know that it can do it.

Derek

[JamesF]

What camera/lens/aperture/exposure are you using?

James

[fwm891]

rfdesigner: Not used Iris before so may give that a try - meantime play with other settings in DSS

James: Camera Nikon (full frame chip D700), 24mm f2.8 @ f3.5 and 20 sec exposures at 1250 ISO

There are 17 x 20 sec exposures in the stack

[JamesF]

Twenty seconds woudn't be enough to show field rotation at that kind of focal length I'd have thought. I'd have to suspect distortion in the lens initially. My kit Canon 18-55mm lens shows quite noticeable distortion around the edges of the frame that is nowhere near as noticeable when I use the 50mm f/1.8 lens even at exposures twice as long.

James

[Digz]

If the lens is the Nikon 24mm f2.8 I think distortion at the edges will be minimal, although shooting wide open will highlight any distortion thats present. I think Derek might have hit the nail on the head to me it looks like an issue with the stacking.

Can you post a single sub for us?

[fwm891]

Hi James, No, individual exposures are pretty good right across the frame (every frame). Its only when I try to combine them in DSS that the 'zoom' effect is seen.

Single frame 1 x 20s from this set:

[fwm891]

Hi Lee - posts crossed see above - sat trails as well!

[Black Knight]

I've had this before. During troubleshooting, I even went and re-processed a known-good stack to check the output: Same zoom effect. The conclusion I came to was that either the DSS settings had become corrupted, or a file had become corrupted. I removed and reinstalled DSS, and everything was fine from that point on.

[Black Knight]

... or a file had become corrupted. I removed and reinstalled DSS, and everything was fine from that point on.

What I meant to say was: "or a DSS installation file had become corrupted".

[anthony]

I presume all the subframes are of similar quality, and it is not the case that one or two dodgy ones have sneaked in! I wondered that the camera just lying on its back might not be the most stable of arrangements :-)

Although a single 20 sec sub won't show rotation, there will be some over that 20sec x 17 series of images. I have not been actively imaging for a little while and so I don't know what super clever stuff the software can do with things like this nowadays so I am probably wrong!

[Digz]

Cheers Francis.

Nice sat trails. I would be tempted to uninstall and reinstall DSS to see if that solves it, maybe after a thorough check of the subs for any obvious issues.

[fwm891]

Lee - This is a fresh download as I had similar thoughts

Ant - all the subs match as they were shot on a programmable timer: 20 second exposure with a minute between the start of each frame. The camera is quite heavy and was placed on a stable surface. There is some rotation over the period so I was expecting some edge/corner problems due to rotation as stars enter and leave the frame.

I'm currently taking a few frames from the session and applying different stacking/alignment settings so I have some comparison between the various options

[dph1nm]

My suspicion is that DSS has detected too many stars, or not the same stars in the different images. So DSS is then having trouble doing its alignment routine. Try cutting down the number of detected stars and/or using the manual star edit function.

NigelM