LRGB Star Overlap

[groberts]

I am just embarking on LRGB imaging using a ZWO 1600-Cool and have much to learn. I did get a few frames recently with a number of obvious problems, one being focus which, if we ever get clear skies here again, I hope to rectify soon.  However, I have manged to process, merge and obtain my first LRGB image which is very poor but, notwithstanding the other problems, indicates that the channels are not properly aligned, with the respective colours clearly shifted (see enlarged image section below) - I stacked in DSS and process in Photoshop.

What could be causing this / is it processing or image capture?

      

[Freddie]

Is the mis alignment the same everywhere across the image?

[Davey-T]

If it's the same all over you can simply align the channels in Photoshop.

Dave

Never tried it but I thought DSS was supposed to align the channels

Edited February 8, 2017 by Davey-T

[carastro]

DSS will align the frames but I have never heard of it aligning the channels.  If the overlap is the same all over, you can just slide the individual channels using the move tool in photoshop.  If there is any rotation or it is not the same all over you will need something like Registar to line them all up.

 

Carole  

[Xplode]

All images should be registered to the same image, i guess you chose 1 image from each channel to register against?

[groberts]

18 hours ago, Xplode said:

All images should be registered to the same image, i guess you chose 1 image from each channel to register against?

Thanks everyone, I was expecting the respective colour channels to automatically align when pasted into Photoshop not to require manual alignment. What I did was this:

1. Opened each colour image set one at a time for L/R/G/B separately in DSS + calibration files.  Run Register once then select the picture with the highest score as the reference frame and registered again,  then Stack and repeat for each R/G/B/L set of images;

2.      Next, for each L-R-G-B stacked master, stretched it a bit then pasted each one into the Red-Green-Blue channels to produce an RGB master + created a Luminance layer to combine with the RGB master and subsequent post processing.

 

I have a feeling the registration might be the problem but not quite sure what is meant by Ole’s comment.  Are you saying:

(i)                All the LRGB subs + calibration are first registered all together, rather than for each colour?

(ii)              Then select the one with the highest score from all these as the Reference Frame, then register all together again?

(iii)             Then Stack each registered LRGB + calibration group separately to produce RGBL stacked masters for subsequent combining (as previously dine) and post processing in Photoshop?   

 Thanks, Graham

[cfpendock]

What I do in DSS is to select my first Luminance image as the reference frame.  So, whenever I am stacking (for the particular target) I always load that image (usually L1), together with darks and/or flats etc, as well as the other images I wish to stack - for example all the reds.  Then select L1 from the list, right click and "use as reference frame".  Then "check all", followed by unchecking the L1 on the list.  It will still be used as the reference frame, but will not be stacked.  Then click stack or register (depending on what you want to do.  The final resulting stack will all be aligned with L1.  There are also ways of loading all the colours at once into DSS, keeping them separate for stacking, but always using the same reference frame.  It's all in the DSS help section.

Hope this helps

Chris

[groberts]

Thanks Chris, when you say you use the first luminance image as the reference frame, do you mean the stacked luminance subs or the 'best' luminance sub?   

[cfpendock]

Neither the best nor the stack.  I tend to use just the first luminance sub simply so that I can remember it. It could as well be any other (acceptable) sub, as long as it is the same sub used as the reference frame for all further stacking.  For example, I might do a session with LRGB, and then another time, I might have the opportunity to add some Ha.  So then I load my Ha subs, darks etc, as well as the original L1 sub.  But I leave L1 unchecked for stacking, just clicking on it to use as the reference frame.

Chris

[sharkmelley]

Agreed that you must use the same reference image for all channels.  Your data has one additional problem - it is unlikely that DSS will recognise those coloured discs as stars, so the "stars" it finds are more likely to be hot pixels.  If it stacks on the hot pixels then whatever reference frame you use will make no difference at all - the channels still won't line up.

Shoot some decently focused data and use that for your DSS experiments otherwise you'll end up going round in circles.

Mark

Edited February 10, 2017 by sharkmelley

[groberts]

Thanks guys - understood now.

While on the matter of LRGB DSS processing, is it necessary to use relevant calibration images with every colour?  

Yes, I am aware that focus is a problem and whenever (if ever) we get clear skies intend to hopefully sort it out; coming from DSLR I have been surprised how much more difficult this with CCD type sensors.

Graham  

Edited February 10, 2017 by groberts

[Freddie]

4 hours ago, sharkmelley said:

Agreed that you must use the same reference image for all channels.  Your data has one additional problem - it is unlikely that DSS will recognise those coloured discs as stars, so the "stars" it finds are more likely to be hot pixels.  If it stacks on the hot pixels then whatever reference frame you use will make no difference at all - the channels still won't line up.

Shoot some decently focused data and use that for your DSS experiments otherwise you'll end up going round in circles.

Mark

The OP says the pic is enlarged just to show the misalignment. As he has managed to "process" the subs I think DSS is seeing the stars as they no doubt look sharper at normal scale. Always room for improvement with focus though.

[cfpendock]

4 hours ago, groberts said:

is it necessary to use relevant calibration images with every colour?  

If you are likely to have dust on the filters, yes, you should use different flats for each filter.  I try to make sure my filters are clean so I only take one set of flats which I then use for all my filters for a particular image. Darks should only relate to the image exposure time - most people have a library of darks - e.g. 5, 10, 15, 20 minutes etc., and use the appropriate dark for the image.  Bias are always the same.

Chris