Help with WIPE

[HDetwiler]

Greetings,

I have read the manual on the WIPE process, but after using it I have a pretty basic question.

What kind of a result am I looking for? I realize that I'm looking at a temporary "stretch," but don't know if my targets are supposed to be all white, or the background is. Should I be adjusting the parameters to try for a dark background? Are the nebulae supposed to look sharp and in focus?

The attached screenshot is of the lagoon and trifed nebulae. 16 photos were taken in a Bortle 2 zone using a Canon 6D with a 300mm F5 lens and 1.4x doubler, stacked in Siril, run through AutoDev once, cropped, and binned to 50%. I've continued my processing using the shot you see here, and the result comes out with a lot of noise.

Thanks for any suggestions!
Henry

Wipe.jpg (359.38 KiB) Viewed 4875 times

[Mike in Rancho]

Hi Henry,

Hard to tell, but I don't see anything necessarily wrong with that Wipe screen. Hit the Color button and review what that is like also.

Wipe takes a bit of practice and experience, but really you are just looking for an even field. Such as vignetting that didn't fully correct with flats, or a light pollution gradient.

Wipe is not made to erase noise, but can erase signal with increased strength. Depends how much you need.

As you noted, what you see is highly stretched. In most cases you will not reproduce such a strong stretch in the post-Wipe AutoDev. AutoDev takes a little practice also, but it is there that you will choose an optimized global stretch for what the data can support.

I see a little bit of egginess in your stars - unsure if from tracking or field curvature.

You also mentioned this is a stack of 16 lights -- how long was each exposure? The key to signal versus noise will be the total amount of integration. For unfiltered broadband with a stock DSLR, but of a nebulous target, I would start leaning towards this needing more time. For the noise that you noted, that is.

How to divide up that total time, and the settings to use, will depend on the "best" ISO for your camera, and how many star cores you might be willing to saturate in the given subexposures.

Hopefully some of that makes sense?

[hixx]

Hi Henry,
as Mike suggested, this is tot a really bad WIPE result. In a bortle 2 area, You'll get a lot of Shot noise, as this is correlating with signal strength. Note that WIPE will remove the gradient but NOT the gradient's shot noise!
So more background light means more shot noise. The SNR however will increase with integration time so 64 lights or more will surely help. I usually stack more than 100 lights. Also You'll benefit from creating a luminance stack using a narrowband / duoband filter, especially for large-scale nebula fields. As it will reduce background illumination, but not the nebula, it will mitigate the LP shot noise portion as well. Be sure to create a second stack for the color signal using a Clear or UV/IR filter....

Clear skies,
Jochen

[HDetwiler]

Thanks Mike & Jochen, for your analysis and advice. These were 2-minute exposures, at 1600 ISO. 1600 ISO is supposed to be the sweet spot for my 6D camera, although I've tried 3200 and I don't see any difference. I'll try more lights, but worry about falling asleep at the wheel--at some point I either have to camp in the desert or go home. Is there a magic number of light frames which I should shoot for? You mentioned 64, Jochen--is that irregardless of exposure time? If I went to a Bortle 1 zone, could I cut that down to 40 lights? Because I'm using a 300mm lens on a DSLR, I haven't found any high-quality filters that will work. I did try a clip-on filter--but it keeps the mirror locked up, forcing me to look for and focus on stars in "LIVE" mode--an impossible proposition.

Clear Skies!

[Mike in Rancho]

Hi Henry,

So sounds like you are just sort of getting your feet wet here? Although it does seem you found that 1600 is listed as a good ISO for that Canon.

Total integration time is what will count. That will increase your SNR. Doesn't matter too much how it is split up. However, you generally would like to choose subexposure time so that your skyfog peak is just off the left side of the (linear) histogram, but also not oversaturating stars. Or at least not too many. At a very dark site, I suppose that could prove troublesome, as background skyfog would be minimal. Anyway, the gist is to get your shot noise above your read noise.

At a dark site, 2 minute subexposures may be fine. I've never been to such a site! But...check your star cores, or decide if you are okay sacrificing some.

Your total time here, however, is only about 30 minutes, which is rather brief, and thus stands the chance of being pretty noisy.

For broadband shooting, I would avoid LP filters. They may help with the nebulosity, but colors will suffer, and color balancing will be difficult. Dual narrowband filters are quite good, and popular, with OSC and DSLR for nebulas.

For now, you might want to consider ditching the 1.4x tele-extender. I think that's pushing you up to 420mm, and slowing you down to f/7? So maybe go a little wider and faster, and I think there's stuff here around this target that would allow that just fine.

Live view focusing can be difficult through lenses. In fact, focusing on stars in general can be. What I would do, when using one of my D5300's and a Tamron 300mm zoom (aside - zooms also aren't great for astro), would be to do my focusing on some brighter stars ahead of time. Once I got close, I would then use the viewscreen zoom to get in close and dial it in sharp. You may also be able to see some fainter stars nearby suddenly "pop" into view, and those are the ones that let me know I was at good focus. Then, unless your lens has other options, always good to use a piece of tape to hold both the focus ring and zoom in place!

I did my first M31 with that lens, and it was pretty bad. I think the only target I halfway like with a camera lens was a NA + Pelican that I did a year ago. I'm not sure I ever updated that one with improved processing, so maybe I'll go put that up now. Anyway, that was done with a full spectrum D5300, the 300mm f/5.6 zoom, an L-eNhance dual narrowband filter, and was 3 hours of total integration using 240s subexposures. That region is pretty bright, although I think the Lagoon/Trifid is also. But an unmodified camera in broadband will take some bit of time to capture that nebulosity, at least with decent SNR.

Oh and I didn't use any clip-ins, but instead have always bought 2" filters. Good for threading onto T-ring noisepieces, coma correctors, focal reducers/field flatterners, and so on, for telescopes. To use it on my 300mm zoom (which I think uses 62mm filters), I just bought a whole mess of step-up and step-down adapters.

Here's the post with my only camera lens nebula. viewtopic.php?f=11&t=2386

[hixx]

Hi Henry,
First: try to find a cheap but decent fixed focus lens instead of the zoom. Old manual focus lenses are pretty cheap and may be fitted with a Bathinov mask to focus using Life mode & focus assist magnification. I own a Pentax 300 mm f4 which is not APO grade but OK for wide fields. A good one I own is the Samyang/Walimex 135mm f2 manual focus. It is tack sharp, cheap and nearly APO grade, yet affordable. It is great for wide fields and collects tons of photons.
Second: Limit exposure time to what Your mount / guider can handle I started using short exposures (15 - 30s @ ISO 3200 but shot hundreds of lights. 260 lights is a magic number as it brings down read noise by a factor of 16 (on paper) which means the ISO 3200 stack will be equivalent to an ISO200 shot, which brings you into shot noise limited territory.
Third: As Mike said the only thing that helps against shot noise is total integration time, and yes, I used to camp in the countryside at an Astrocamp in our area and let mount/scope/cam do their job (no falling asleep at the wheel) With 280 lights @ 30s You'll get 140 min integration time. You'll need an overhead of 20-30 lights though to exclude bad lights /tracking error, wind, plane trails,.... In a full night You might capture 3-5 hours of integration time (>400 lights) which will give You great images. This is how I started
Some folks even stack multiple nights...

Cheers,
Jochen

[jhart]

I'm with Henry back on his original post -- I still can't get when Wipe had done its job as best as it can with the data it has to work with. I understand its aggressive stretch to identify problems that need to be masked out. But I image in a Bottle 8-9 area and have a lot of light pollution and gradient to deal with. I don't understand how the overly aggressive stretch helps with identifying the best parameter settings to deal with that problem. I have read all of the manuals and as many posts as I can find. (For what it is worth I have tried to create a cheat sheet for my reference -- see attached). I read about the goal of a uniform background, but sometimes striving for that uniformity (which usually looks really bad otherwise due to the stretch) results in a lot of lost data for the main object; and sometimes with a lot of gradient in the Wipe background the image turns out not too bad. Basically, to me the image is stretched so much I can't tell if the gradient is removed or reduced as much as possible at that stage. Given the modular approach of StarTools I would like to be able to see if I have used Wipe correctly before moving on and not having to wait until I see results in AutoDev or in some cases many modules later such as Color-- and even then I am not sure what I need to go back and change in Wipe. I am also completely clueless on what the color image means that Wipe requires me to check. Do the different colors mean a different problem? E.g., is a blue background better than a bright orange one? Does red mean really bad?

Jeff

[HDetwiler]

Thanks gentlemen, for the additional advice! I like your cheat sheet, Jeff--I will make use of that. I really like what you did with your 300mm on the NA nebula, Mike. That gives me a goal to shoot for! I'm using a 300mm prime lens (Canon L), which is 77mm, or about 3". I don't see how those 2" filters are going to work with that. But I'll definitely be increasing the total time on my subs. I see I would be well served to move up from 30 or 45 minutes to a couple of hours.

Yes, I am just a newbie. I've been out shooting the night skies on and off for a few years with my camera and Sky-Watcher Star Adventurer, but my attempts to capture detailed nebulosity are dismal. Only the brightest objects have come out well, e.g. Andromeda and Orion. This is a shot of Andromeda from last week--and it can also use improvement. So I'll try out your suggestions and report back as I get better results.

I do have one follow-up question. My 300mm prime is F4 (not F5 as I stated earlier), and with my 1.4x doubler/extender I do lose a stop (down to F5.6). Is the 1.4x increase in magnification worth the loss of light to pick up more of the detailed nebulosity? And perhaps smaller, fainter objects?

Cheers,
Henry

Andromeda_101722.jpg (471.14 KiB) Viewed 4782 times

[hixx]

Hi Jeff,
the Color view in Wipe lets You check the RGB (color) channel. As luminance and color data gets processed in parallel, both datasets will have different WIPE results which may be checked using this feature. For example, the color view can tell you if there is remaining color shift around the blackpoint. And Yes, at WIPE stage, the output should be uniform, which means vignetting and gradient removed. You will still see the shot noise of the gradient, so You can judge on flatness of the image by looking at uniformity of the noise.....
The WIPE autostretch is something one needs to get used to, I think it is the same stretch applied as the initial AutoDev diagnostic stretch, so a difference may be viewed before/after WIPE using the same stretch.
Wipe should not eat object details, I think it looks for large-scale gradients (the object has much smaller detail size usually). If You still see it eating details, try to play around with the Gradient edge behavior and gradient falloff parameters. As a last resort, reduce aggressiveness a bit.

Cheers,
Jochen

[Mike in Rancho]

Thanks gentlemen, for the additional advice! I like your cheat sheet, Jeff--I will make use of that. I really like what you did with your 300mm on the NA nebula, Mike. That gives me a goal to shoot for! I'm using a 300mm prime lens (Canon L), which is 77mm, or about 3". I don't see how those 2" filters are going to work with that. But I'll definitely be increasing the total time on my subs. I see I would be well served to move up from 30 or 45 minutes to a couple of hours.

Yes, I am just a newbie. I've been out shooting the night skies on and off for a few years with my camera and Sky-Watcher Star Adventurer, but my attempts to capture detailed nebulosity are dismal. Only the brightest objects have come out well, e.g. Andromeda and Orion. This is a shot of Andromeda from last week--and it can also use improvement. So I'll try out your suggestions and report back as I get better results.

I do have one follow-up question. My 300mm prime is F4 (not F5 as I stated earlier), and with my 1.4x doubler/extender I do lose a stop (down to F5.6). Is the 1.4x increase in magnification worth the loss of light to pick up more of the detailed nebulosity? And perhaps smaller, fainter objects?

Cheers,
Henry

Pretty good M31 there Henry, with what I assume is fairly short integration again? Yes, bump things up from 30 minutes to a couple hours or more, and your images should increase in quality significantly. f/4 is pretty fast. My Newt with coma corrector is f/3.8, but I still usually want at least a couple hours on targets. And that's with either an astrocam or, before I had that, a modified DSLR. Unmodified means you will want even more integration, especially nebula. And though Canons tend to be "not terrible" as far as how much Ha gets passed through, it's still going to be pretty low. Like maybe 25% gets through.

A barlow or tele-extender tends to be frowned on for DSO, unless absolutely necessary. And many will absolutely have a cow if you use the term magnification. 1.4x isn't a whole lot I guess, but it is extra glass in the way also, with an increased chance of chromatic aberration, and I'm not sure the extra reach is worth it here. Plus, smaller brighter would likely be more amenable than fainter. Remember with more reach and the tighter field of view you are spreading what light you capture across more pixels, so it takes even longer to increase the SNR. Have you tried some experiments? You can always crop in (to an extent). Your M31 might fit in better with that 300mm prime, would be captured faster, and would also be floating there against a nice background of space and stars.

Oh, before I forget, note also that M31 is a tricky object that sometimes requires a mask in Wipe. Not always, but I've seen plenty of Andromeda data where you really just have to use it.