Align without combine and move to another software program

[andyo]

Hi,

I have recently moved from dslr imaging to imaging with a mono atik314L+ with filters.I have only had one session with the new camera and am looking at my workflow after receiving advice from other members.I did the usual callibrating and stacking of the channels to make seperate masters of the R,G,B and L channels.I then used AA5 Trichromy and LRGB synthesis to align and combine then finished off in CS6.

The advice I was given was, its best to align but not combine the seperate R,G and B channels in AA5 then combine the channels in photoshop.

However I have a lot to learn regarding processing especially with mono imaging.So my question is how do I align in AA5 then move the aligned images to photoshop.Am i right in thinking that I align in AA5 by selecting" align all" autoalign then just save the three master files by overwriting them and these new saved files can be moved now ready aligned for combining in photoshop( i also know how to merge them).All help appreciated

[steppenwolf]

I don't use AA5 (I use MaxIM DL) but the process will be the same - combine the individual channels (R, G and into three masters in AA5 and save each as a 16bit TIF file with a unique name that includes a reference to the colour, for example:-

m31_210313_r.tif for the red channel file

m31_210313_g.tif for the green channel file

m31_210313_b.tif for the blue channel file

You can then open these three files in PhotoShop and combine them to produce an RGB colour image

[andyo]

Thanks Steve.I use the technique you describe in your book when in photoshop for merging the three master files but I thought I needed to align these 3 channels first before merging them together into an rgb image.I read this in Martin Bs tips and tricks in image processing on this forum where he states:-

Align and combine each channel seperately then align (but don't combine) the combined channel images.

[steppenwolf]

I slightly misread your original post I'm afraid as I thought you were talking about the original individual subs. when you mentioned "its best to align but not combine the seperate R,G and B channels".

My full procedure is to combine each of the R, G and B data sets into masters using a two star alignment method to remove any drift or field rotation from the stack. I then save the Red master as a 16 bit TIF. Next I align (but don't combine) the green and blue masters with the red again using a two star alignment and save them as 16 bit TIFs as well, using the typical file names shown in my example.

[ollypenrice]

I use AA5 and do use their trichromy routine but not their LRGB synthesis. My workflow:

Make calibrated stacked masters of R and G and B separately so I have a final image for each channel. (BTW I make flats using a master bias as the dark-for-flats. I use the hot pixel filter on the options page as well. It's very effective. The option I use for combining the subs is Sigma at default and I align using star pattern translation and rotation. This is most important.

Before using Trichromy I use Image Align All with just the three channels open. Again I use star patttern translation and rotation. This is vital because for some daft reason the default 'auto align' once you are in Trichromy is just x-y translation and will not align the corners properly unless you have supernaturally good polar alignment, which I don't. In Trichromy I check the white balance but don't use auto calibrate. You can try it though.

While in AA5 at this stage you could also align the L channel to the new RGB but I would certainly not apply it to RGB at this stage because processing L is totally different from processing RGB.

Next I do a DBE routine in Pixinsight to colour calibrate the RGB and settle the background, but you could do this in Ps by hand.

I process the RGB for strong colour and low noise but not for detail. I process the L to drag out faint stuff and get the details sharp and contrasty. (Totally different objectives.) I thn apply the processed L to the processed RGB in blend mode Luminosity. Exactly what follows can get complicated but that's basically it.

AA5's trichromy is good. Why not use it?

Why don't I use PI for all this aligning? Because it drives me nuts and because AA5 is so beautifully laid out.

Olly

[andyo]

Thanks a lot for your help Steve and Olly that really helps me.Olly what is a DBE routine

[ollypenrice]

Thanks a lot for your help Steve and Olly that really helps me.Olly what is a DBE routine

Dynamic Background Extraction. It has to be the greatest single graphics processing routine in the history of AP. It listens carefully to you when you tell it what, in your image, is really background sky. It then sets these regions to a neutral dark grey weighted 1:1:1 in RGB. If you prefer a different background sky value you can easilty tweak it later but a sodium street light in the upper right hand corner will lose it's dominion over your picture!! It will also do a very good job of setting a realistic RGB balance. I find it far more accurate than auto calibration in AA5. In fact, for me, auto calibration in AA4 was better than in AA5 but this will depend on the filters you use and the camera's response in different wavelengths.

PI enthusiasts will insist that PI has other virtues beyond DBE and it does. SCNR green is another 'must have.' But after a while I part company with PI because I find the PI guys produce images which are a long way removed from what I'm trying to do. A matter of taste. I will have nothing to do with HDR wavelets, for instance, because I can see them a mile away.

Olly

[andyo]

Dynamic Background Extraction. It has to be the greatest single graphics processing routine in the history of AP. It listens carefully to you when you tell it what, in your image, is really background sky. It then sets these regions to a neutral dark grey weighted 1:1:1 in RGB. If you prefer a different background sky value you can easilty tweak it later but a sodium street light in the upper right hand corner will lose it's dominion over your picture!! It will also do a very good job of setting a realistic RGB balance. I find it far more accurate than auto calibration in AA5. In fact, for me, auto calibration in AA4 was better than in AA5 but this will depend on the filters you use and the camera's response in different wavelengths.

PI enthusiasts will insist that PI has other virtues beyond DBE and it does. SCNR green is another 'must have.' But after a while I part company with PI because I find the PI guys produce images which are a long way removed from what I'm trying to do. A matter of taste. I will have nothing to do with HDR wavelets, for instance, because I can see them a mile away.

Olly

This is all fairly new stuff to me but DBE sound like a more complex versiion of gradient Xterminator or am I way of the mark.Also SCNR green is that like the HLGV plug in for photoshop.

I tried balancing the rgb in cs6 by adjusting the black point nudging it up to the start of the histogram for each individual colour (R,G&.

[ollypenrice]

Yes, HLV Green is based on the PI routine as Rogelio explains on his website. And, yes, there is a similarity between GradX and DBE but DBE is far more sophisticated and effective in my experience. GradX is still good though.

If balancing colour by using the histograms for each channel in Ps I'd try to get each channel aligned at the top left hand side of the peak rather than at the foot. I'd be sure to treat this as a starting point for the colour balance though. I wouldn't get too rigid about it. In this screen shot I've arrowed the point I would align for the red channel.

I wouldn't get too close to the foot of the peak with the black point. There can be faint data in the flat black line to the left of the peak.

Olly

[andyo]

Thanks olly,that very useful.I always aligned the foot of the histogram from now It will be the top left hand side of the peak.Gradient xterminator will have to do for now I might look into PI in the future

[ollypenrice]

If you think about a diffuse emission nebula there will be a lot of faint signal from Ha in the red channel so its peak will rise slowly. If there's no reflection nebulosity, though, the blue will only really come from the stars so that peak will rise from a flat base very steeply. The sky itelf should be roughly 1:1:1 for most people so the far left of the histogram should be at parity and just a think line. Well, so much of the theory...

Olly

[fwm891]

There is an option in PS using levels. If open levels and you double click on the black eye dropper it will open up a colour pallet where you can select a dark grey of your own choice (make sure web only is not checked). Then if you open each of the channels you can apply that grey to the RGB in turn.

I use AA5 and find it very good combining RGB images - not keen on the LRGB synth routine.

[andyo]

Thanks Olly Steve and Francis your help has been fantastic and much appreciated .Been experimenting most of the day and applying the tips given here.I have found that the rgb align in AA5 works great.The eyedropper tip is useful to know

and has also worked well.I just need more data now to try it on.That could be a while looking at the forcast though