Processing process for deep sky imagers: Page 2

Processing process for deep sky imagers

bdeclerc

JohnK wrote:
Great information sharing guys - thanks all.

One of the issues I see in capturing several thousand frames over several hours (the whole night form me) is with a non permanent set up (again me!) some field rotation and the somewhat varying quality of the frames to to the effective focal length.

The beauty of AS2 is it's ability to sort the frames in order of quality and then stack. But it will struggle to stack with field rotation. Is this correct?

It won't just struggle to stack, it will fail to stack decently with field rotation - there is an experimental feature (under "advanced - Experimental features and parameter tuning") for compensating for field rotation - it allows you to set observer latitude, object altitude & azimuth at the start of the recording and recording duration in seconds - this supposedly compensates for the field rotation but I'm not sure how good it works in practice - it doesn't fit your use case (since I assume your "some field rotation" is more due to imperfect polar alignment, not due to using an Alt-Az mount).

But you might be able to manually measure the field rotation betwean first frame and last frame, then tweak the rotation parameters directly to get the desired result.

johnk

So for my set up, the process thus far looks like this:

1. Capture with latest FC software in separate LRGB sets in BPM or TIFF format.

2. Take Darks, Bias and Flats

3. PIPP pre-sorting best frames (say 3,000 per set)

4. Deep Sky Stacker staking per LRGB set

5. Combine and process in Photoshop

The only issue remaining is the vertical banding issue being reported in terms of gradients - so perhaps a Gradient Removal Plug In or Action for Photoshop??

firstc8

maurizio wrote:
and here you are the Luminance too

Thank you for the images.

I struggle to learn merge. Went through several photoshop help sessions but none of them were astroimaging specific. Anyone has a link to some youtube tutorials to recommend?

I want to be proficient with mono processing before getting a mono camera and filter wheel.

landbo

JohnK wrote:
Great information sharing guys - thanks all.

One of the issues I see in capturing several thousand frames over several hours (the whole night form me) is with a non permanent set up (again me!) some field rotation and the somewhat varying quality of the frames to to the effective focal length.

The beauty of AS2 is it's ability to sort the frames in order of quality and then stack. But it will struggle to stack with field rotation. Is this correct?

So, if this is the case, the only way to then overcome field rotation is either for Emil to add this as a feature of AS2 in the future, or to pre-sort the frames in PIPP, choose the best ones and then use DSS as per normal (for me) to Stack, add Darks, Bias and Flats. So SER, AVI files are out of the picture here, so need frames in BMP, TIFF or FITS for DSS. However, the banding issues that has been discussed with the CMOS cameras then needs to be sorted afterwards.

I am not sure to what extent the new cooled cameras from ZWO will have set point cooling, but my thinking will be to build a set of dark frames with pre-set gain/exposure levels which reduce noise and minimise the banding issue that was being discussed with CMOS chips (if possible)

Any more thoughts from anyone? and comments on my logic regarding the above?

Thanks again.

John K.

Yes John. If one has field rotation in the recordings, it can be a major problem. Even the I use AA5 (AstroArt5) for stacking that can handle the rotation automatically. And AA5 can also handle discarding of frames automatically.

Have tried the new AS! 2 beta once without much luck. Hopefully the program will be better some day.

The problem with banding, I have not encountered ........... yet.

Has also played a bit with cooling (ice in a plastic bag behind the camera) of my ASI174MM. Have found the amount of amp glow is greatly depending on the camera's chip temperature. Therefore it is a bit difficult and compensate. How well the cooling works on the new refrigerated version, I do not know.

Bedst regards, Leif.

johnk

Thanks Leif,

Another question I have is regarding exposure levels for LRGB.

With planetary using my ASI120mm, my gain/exposure level is based on the Histogram and FPS numbers.

I gather this is the same for Deep Sky when using these cameras?

I need to determine how to set the exposure range for the 174mm that I am buying and will use LRGB filters.

Many thanks,

John.

landbo

Sorry, we encountered an error while displaying this content. If you're a user, please try again later. If you're an administrator, take a look in your Flarum log files for more information.

johnk

Just one more question - and maybe this is for the creator of Firecapture - I know that obviously taking darks separately for each LRGB filter is the way to go - BUT firecapture can do this on the run correct?

Just wondering for Lucky type imaging if this is a real time saver so that captured images are then already dark subtracted?

riklaunim

You need a dark frame for every exposure time/gain. If you keep the same settings for two or more filters - you can re-use it. You need to apply the dark frame on your own when processing (and probably a flat frame too).

johnk

So the latest version of Firecapture now allows up to 50 flats and 50 darks to be taken, averaged, stored and used on the fly. This is a real innovation by Torsten Edelmann.

Just wonder what the optimal would be so that this does not necessitate any post processing flat or dark subtraction and can be done on the fly - as the new cooled ASI cameras are set point, this to me suggests that darks and flats are simply taken and stacked/subtracted etc with each filter change during an imaging run.

riklaunim

You can't make flat or dark frames just like that. For a dark frame you have to cut the light and for a flat frame you have to have evenly illuminated surface. Most applications that offer such functionality store flat/dark/bias that have been done semi-manualy before imaging.

The dark frame/flat frame could be handy for some planetary imaging like CH4/UV where FPN and hot pixels could be visible due to high gain, longer exposures and low object brightness. For DS imaging I would rather like to process frames with my own flat/dark and hot pixel map frames. But still it simplifies processing for someone not wanting all of that processing options.

firstc8

maurizio wrote:
and here you are the Luminance too

I finally had the time to play with your LRGB channels on Photoshop. RGB image looked fine, but when I added the lum channel it was all red. What am I missing?

Thanks for the images for practice.

maurizio

I finally had the time to play with your LRGB channels on Photoshop. RGB image looked fine, but when I added the lum channel it was all red. What am I missing?

I don't know. I'm using Paintshop. Did you align succesfully the rgb channels? When you're done with rgb and you get a single layer, merged rgb picture, then theorically addling the new layer for luminance you should only choose a particular "blend" mode for that layer named "luminance" (or something like that) into the Photoshop interface.

firstc8

Yes every channel aligned perfectly well. The RGB image looked quite well already, but I understand adding the luma channel will make the image pop more. Will try luma again, thanks.

firstc8

maurizio wrote:
I finally had the time to play with your LRGB channels on Photoshop. RGB image looked fine, but when I added the lum channel it was all red. What am I missing?

I don't know. I'm using Paintshop. Did you align succesfully the rgb channels? When you're done with rgb and you get a single layer, merged rgb picture, then theorically addling the new layer for luminance you should only choose a particular "blend" mode for that layer named "luminance" (or something like that) into the Photoshop interface.

Yes after selecting "Luminance" the combined image came out well and much more pleasing than just RGB, thank you!

maurizio

One more simple trick to better align stacked RGBL pictures.

It is normally possible that you cannot align perfecly for not being able to be pixel exact in overlapping each component. This results in colored fringes around the stars discs.

But if you previously double the size of each RGBL picture the following alignign process will be two times more precise and will display smaller or even no colored firinges around the stars and a more sharp image. This is true also for planetary pics. Next you can resize to the original with virtually no loss of quality.

firstc8

maurizio wrote:
One more simple trick to better align stacked RGBL pictures.

It is normally possible that you cannot align perfecly for not being able to be pixel exact in overlapping each component. This results in colored fringes around the stars discs.

But if you previously double the size of each RGBL picture the following alignign process will be two times more precise and will display smaller or even no colored firinges around the stars and a more sharp image. This is true also for planetary pics. Next you can resize to the original with virtually no loss of quality.

I did that yes zooming in helps but there seems to be a limit. Photoshop does not seem to allow you to align down to pixel level.

maurizio

I did not talk about zooming, I wrote "double the size" that means if the picture's size are 1280x960, then doubled (resized 200%) will be 2560x1920 pixels :

Of course you can zoom to see better the details during alignment but the trick is based on resize and not zoom, resize is a function that any graphic SW can perform, even Photoshop ;) .

You can resize in any percent of the original (124%, 45%,94%... etc) or giving pixel values (640x480, 1456x357.... etc)

Hope I'm clear now

firstc8

Got it, thanks!

« Previous Page