Lens distortion stopping DSS finding star patterns?- Distortion-correcting frames

[pipnina]

When trying to get my 81 images to stack in DSS, I can't help but think some lens distortion is stopping DSS from picking up the star patterns in some of my subs.

The list starts out fine, http://i.imgur.com/O5ZyOJs.png

But a lot of the images have no offsets detected... http://i.imgur.com/9CxUTe8.png

I was wondering if this could be because of the same reason the final image (43/81 light) loses sharpness dramatically away from the stacking centre, (http://i.imgur.com/zipxuH3.jpg) lens distortion.

I've already taken measures to prevent this distortion (I found it is least present on my lens at 35-55mm) but it is clearly still having an effect on the stacked images.

This sort of ties in to another question I wanted to ask: I'm planning on making my own stacking software (unsatisfied with DSS and the eye-watering prices (well, it is to me, I'm 17) of alternate software. I kinda want to make my own)

I'd like to know if It would be possible to make distortion-correcting frames? I.e. point the scope/camera at a peice of paper with a pattern of black squares such that it filled the entire FOV and use the didigtal version to compare and create a map of how the image gets distorted. Would this work? I have some ideas on some new ways of making stacking (hopefully) easier and produce nicer results on lower-end kit and will talk about this idea with my software design tutor once I have some details sorted.

    Help on Q1 and opinions on Q2 appreciated!

        ~pip

[D4N]

Is it field curvature that you are referring to as distortion?

I imagine that if it was that easy to process out then they would not be able to sell flatteners at the prices they currently do!

Sent from my iPad using Tapatalk

[FunkyKoval35]

Hi pipnina

Yes, DSS "do not like" distorted stars, and I think this is a problem. Also, a Moon(?) is not helping.

You can try Sequator. It is a new stacking software for short exposure wide-field images. Sequator has "reduce distortion:lens or field effects" - which may be useful for You. 

[pipnina]

Is it field curvature that you are referring to as distortion? I imagine that if it was that easy to process out then they would not be able to sell flatteners at the prices they currently do! Sent from my iPad using Tapatalk

Yes, it is field curvature. Different name, same illness. I guess you're probably right, although it sounds in my head like my idea for flattening the field artificially *should* work.

Hi pipnina

Yes, DSS "do not like" distorted stars, and I think this is a problem. Also, a Moon(?) is not helping.

You can try Sequator. It is a new stacking software for short exposure wide-field images. Sequator has "reduce distortion:lens or field effects" - which may be useful for You. 

I'll have a look at that software. Thanks.

The moon in that image is nowhere near full btw. It was below 20% at the time (22nd april) I just boosted the poop out of the contrast to make more stars show themselves, hence why it looks full.

Admittedly, it probably still isn't helping...

[lensman57]

When trying to get my 81 images to stack in DSS, I can't help but think some lens distortion is stopping DSS from picking up the star patterns in some of my subs.

The list starts out fine, http://i.imgur.com/O5ZyOJs.png

But a lot of the images have no offsets detected... http://i.imgur.com/9CxUTe8.png

I was wondering if this could be because of the same reason the final image (43/81 light) loses sharpness dramatically away from the stacking centre, (http://i.imgur.com/zipxuH3.jpg) lens distortion.

I've already taken measures to prevent this distortion (I found it is least present on my lens at 35-55mm) but it is clearly still having an effect on the stacked images.

This sort of ties in to another question I wanted to ask: I'm planning on making my own stacking software (unsatisfied with DSS and the eye-watering prices (well, it is to me, I'm 17) of alternate software. I kinda want to make my own)

I'd like to know if It would be possible to make distortion-correcting frames? I.e. point the scope/camera at a peice of paper with a pattern of black squares such that it filled the entire FOV and use the didigtal version to compare and create a map of how the image gets distorted. Would this work? I have some ideas on some new ways of making stacking (hopefully) easier and produce nicer results on lower-end kit and will talk about this idea with my software design tutor once I have some details sorted.

    Help on Q1 and opinions on Q2 appreciated!

        ~pip

I admire your drive and enthusiasm but I doubt very much if writing codes to correct lens distortion in a stacking software would be so easily done. DSS is a competent piece of software so long as you feed it stars that more or less round if not so then it will ignore them in the registration. Some camera software do allow for lens correction but not on third party lenses, you may wish to correct the subs prior to stacking. It is worth having a go.

A.G

[kalasinman]

Image processing software can have plug ins fo lens distortion the data are available online for many lenses and can be easily entered

[wxsatuser]

I'm not surprised DSS struggles if your stars are that shape in that image.

IMO the camera moved or the stars have trailed.

All lenses have some distortion or abberations but imo I do not think it will stop the stacking.

You need to get the stars round and better focused, your fwhm is over 5 and your scores are quite low.

[Zakalwe]

You could try stacking in Photoshop (assuming you have access to it)? That can work well on widefield shots like this.

Have a browse through the tutorials on Lonelyspeck.com...there's a few that deal with this.

[graemlourens]

Hey.

Whats your star detection level?

Try setting it higher, but that it still detects around 30-60 stars. like that it will only take the 'best' stars, and ignore the distorted ones at the outside.

Regards, Graem

[michael.h.f.wilkinson]

Rather than field curvature, I think coma and astigmatism are playing a role. Stopping the lens down reduces that a lot (also affects field curvature as depth of field is increased). True distortion can be measured and corrected for.

[peroni]

Those stars look like the beginning of trails. What was the exposure time and focal length of the subs?

I haven't taken wide field for a while but I seem to recall that at 18mm my max exposure was 30 secs and at 50mm it was 12 secs. Any longer and I'd start to see noticeable trailing.

[dph1nm]

As far as I remember, DSS only fits rotational and translational terms when matching the stars, and so cannot cope with distortions. You can certainly do higher order fits which will take this into account.  It only matters of course, if the distortion pattern (relative to the stars) is different between frames. This will happen with images taken on an alt-az mount (or no tracking mount at all), but not on an EQ mount (unless you move the field centre manually).

NigelM

[Psychobilly]

I have run the raws through the lens correction tools that I have at my disposal (DXO Optic Pro 10 Elite or Photoshop CC)  and then stacked the resulting  tiff's in the past... 

Although if you haven't got the hardware to cope then it can be a painful task stacking a load of 100MB+ tiffs......

Peter...

[pipnina]

Rather than field curvature, I think coma and astigmatism are playing a role. Stopping the lens down reduces that a lot (also affects field curvature as depth of field is increased). True distortion can be measured and corrected for.

Funny enough, I'd already stopped the lens down from 3.6 to 8. I read on some online reviews the sharpness peaks at F8 and degrades due to diffraction past that.

Those stars look like the beginning of trails. What was the exposure time and focal length of the subs?

I haven't taken wide field for a while but I seem to recall that at 18mm my max exposure was 30 secs and at 50mm it was 12 secs. Any longer and I'd start to see noticeable trailing.

It was 25s @ 35mm. Possibly a little high based on your stats. I tried at 18-24mm once but the lens distorion caused stars anywhere not in the centre to multiply

I have run the raws through the lens correction tools that I have at my disposal (DXO Optic Pro 10 Elite or Photoshop CC)  and then stacked the resulting  tiff's in the past... 

Although if you haven't got the hardware to cope then it can be a painful task stacking a load of 100MB+ tiffs......

Peter...

I have a computer capable of loading about 130 100MB images in RAM at once without breaking a sweat. Not to brag... my optical equipment is awful

Unfortunately, I don't have photoshop. I do have a program that can do most of the things PS can do, but stacking the images manually would take me.... A long time.

[Psychobilly]

The stacking was done in DSS...  the Raws had lens corrections applied to them (using DXO Optics Pro 10 Elite) and were "developed" into Tiffs which were then stacked in DSS...

Lots of cores, decent software which supports CUDA/Open CL  , 64GB or Ram and a couple of fast HDD /SSD cached pairs helps 

Peter...

[Zakalwe]

Unfortunately, I don't have photoshop. I do have a program that can do most of the things PS can do, but stacking the images manually would take me.... A long time.

You don't need to stack manually in PS. You can load them into a stack and have PS align them. Then apply a median filter which will remove the noise.

See the lonelyspeck.com tutorials.