Drizzle and scale change - Forums

Topic starter

Sep 18, 2022

Hi.

I’m wondering how drizzle will affect my scale.

my setup is Askar FRA400 paired with ASI294MC pro. That gives me a scale of 2.39”/pix

Now, when I drizzle, the scales goes down to 1.2”/pix? Or there’s something wrong with my logic and lack of knowledge?

Now I attach 0.7x reducer. My scale is 3.41”/pix. Then I drizzle this. Would that take my scale down to 1.7”/pix?
That would mean I’d get wider FOV and faster optics (f3.8 vs f5.6) and the most important better details? Or it only somehow affects stars and not signal?

So by using the reducer and drizzle 2x I’d get benefits mentioned above and detail almost as good as with no reducer and no drizzle?

All this assuming my seeing is good.

andrea tasselli

Frequent contributor

Sep 18, 2022

Drizzling (at 2x) would certainly decrease the image scale by a factor of 2.What it doesn't do is to increase the actual resolution by the same factor nor make any optics faster. So, no, you won't get a faster scope and on top of that the same resolution of the unreduced scope. Sorry…

Helpful Insightful Concise

Astro Snapshot

Topic starter

Sep 18, 2022 · Edited: Sep 18, 2022

andrea tasselli:
Drizzling (at 2x) would certainly decrease the image scale by a factor of 2.What it doesn't do is to increase the actual resolution by the same factor nor make any optics faster. So, no, you won't get a faster scope and on top of that the same resolution of the unreduced scope. Sorry...

Ah, that’s obvious it won’t make my optics faster

maybe my wording is confusing.

what I mean is if I use my 0.7x reducer and drizzle the data (3.4”/pix scale) that would take the scale down to 1.7”/pix. That would make the image 4x bigger as it increases its resolution 2x width and 2x height.

so I’m the end I would get the same level of detail as with native 400mm with no reducer (2.39”/pix) or even better. Or at least detail won’t be worse despite the shorter focal length (as the scale is bigger and so the resolution)

does that make sense?

Or to put it in different words: would the detail be similar if I used non-drizzled native FL vs drizzled reduced FL.

Kevin Morefield

IOTD Staff

Sep 18, 2022

Hi.

I’m wondering how drizzle will affect my scale.

my setup is Askar FRA400 paired with ASI294MC pro. That gives me a scale of 2.39”/pix

Now, when I drizzle, the scales goes down to 1.2”/pix? Or there’s something wrong with my logic and lack of knowledge?

Now I attach 0.7x reducer. My scale is 3.41”/pix. Then I drizzle this. Would that take my scale down to 1.7”/pix?
That would mean I’d get wider FOV and faster optics (f3.8 vs f5.6) and the most important better details? Or it only somehow affects stars and not signal?

So by using the reducer and drizzle 2x I’d get benefits mentioned above and detail almost as good as with no reducer and no drizzle?

All this assuming my seeing is good.

Drizzling is merely a way to recapture some of the resolution lost to undersampling (having a larger image scale than your optical system can resolve). It won't impact your FOV or focal ratio.

With a large 2.39" image scale, assuming your seeing, optics, focus and tracking are good enough, you would wind up with squarish stars and pixelized details. Drizzling divides each pixel in your sub exposure into four pixels before alignment and resampling. Becuase the star or object detail is in a slightly different position on the sensor with each sub (assuming you are dithering here), those sub-divided pixels will be stacked with other subdivided pixels and smooth out the pixelization. This should reclaim some of the lost resolution, though it will not approximate actually having a 1.2" pixel scale.

Due to your large image scale you should see benefits from drizzling either with our without the reducer. Remember that with 2" seeing and all other factors in line, you need 1" or smaller to fully sample that seeing. So drizzling should always help you.

I use a 1.47" image scale in my backyard and drizzling definitely helps with my square star problem.

Kevin

Helpful Respectful Engaging Supportive

andrea tasselli

Frequent contributor

Sep 18, 2022

Or to put it in different words: would the detail be similar if I used non-drizzled native FL vs drizzled reduced FL.

I would go with not very likely, in fact very unlikely. You'll recoup some lost resolution for sure but it'll never get to the point you'll be with the unreduced telescope. And then you'll have to dither in spiral or random walk fashion with some frequency and the seeing must alway been undersampled by a large amount.

Astro Snapshot

Topic starter

Sep 18, 2022

@Kevin Morefield that makes sense.

Do I need to dither each frame? Or every 3-5 is ok? Also 5 pixel dither is fine?

My seeing is usually above 2”. I consider 1.9” as great

30m above sea level and very humid, 90% is a norm

Astro Snapshot

Topic starter

Sep 18, 2022

andrea tasselli:
Or to put it in different words: would the detail be similar if I used non-drizzled native FL vs drizzled reduced FL.

I would go with not very likely, in fact very unlikely. You'll recoup some lost resolution for sure but it'll never get to the point you'll be with the unreduced telescope. And then you'll have to dither in spiral or random walk fashion with some frequency and the seeing must alway been undersampled by a large amount.

Mmm I thought that when drizzling and making the scale bigger it would also fill the gaps everywhere. So increasing the resolution would not only make stars less blocky but also make the detail to be in that scale which is bigger now after drizzle

kuechlew

Sep 18, 2022

You may have a look at the drizzled vs. undrizzled example in this small thread I posted to the beginner AP forum: https://www.astrobin.com/forum/c/group-forums/beginners-ap/drizzle-integration/

I usually dither every 5 images.

Clear skies
Wolfgang

Astro Snapshot

Topic starter

Sep 18, 2022

You may have a look at the drizzled vs. undrizzled example in this small thread I posted to the beginner AP forum: https://www.astrobin.com/forum/c/group-forums/beginners-ap/drizzle-integration/

I usually dither every 5 images.

Clear skies
Wolfgang

Nice! Would you be able to provide these 2 images in the same scale 1:1 to compare them? I can see much nicer stars but also a bit better detail?

Chris White- Overcast Observatory

Sep 18, 2022

Drizzling 2x with well dithered and many subs will allow you to capitalize on some extra resolution with undersampled data, but keep in mind that most all reducers will degrade your image off-axis. So if you dont get a well corrected field with your reducer, it's not worth using it IMO unless you are striving for a larger FOV. Drizzling undersampled data can make your star edges better and less pixelated.

Keep in mind that you dont need to drizzle 2x though, you can still get benefits from the interpolation process even if you only drizzle 1x. It really depends on how undersampled your data is. Just make sure you get enough subs and dither aggressively to get the benefits of Drizzling.

Helpful

Kevin Morefield

IOTD Staff

Sep 18, 2022

@Kevin Morefield that makes sense.

Do I need to dither each frame? Or every 3-5 is ok? Also 5 pixel dither is fine?

My seeing is usually above 2”. I consider 1.9” as great

30m above sea level and very humid, 90% is a norm

Large dithers are good for minimizing the fixed pattern noise we seeing CMOS sensors these days but I think you would get all the benefit of drizzling with a half pixel dither or more. That said, you need a larger dither to resolve noise rejection so go with as much as your process allows. Similarly, the more frequent the dither the better. But if the dither is creating too much overhead for you, every other sub or every third sub should get the job done.

Even with 2-2.5" seeing my goal would be to get subs at 3" or below. So with your image scale the drizzle should help.

It's not magic though - the biggest benefit I get from drizzling is curing blocky stars rather than large increases in resolution. That's at 1.47" image scale. But the more undersampled you are the more benefit you will get.

Helpful Insightful Concise

kuechlew

Sep 18, 2022

You may have a look at the drizzled vs. undrizzled example in this small thread I posted to the beginner AP forum: https://www.astrobin.com/forum/c/group-forums/beginners-ap/drizzle-integration/

I usually dither every 5 images.

Clear skies
Wolfgang

Nice! Would you be able to provide these 2 images in the same scale 1:1 to compare them? I can see much nicer stars but also a bit better detail?

First both at the same image scale (1:1 for the undrizzled version):

1:1 detail comparison:

unfortunately they differ somewhat in the processing but it's very clear that the drizzled version gets more fuzzy at 1:1. You simply can't expect to double the resolution. Admittedly I could work harder on the image by applying deconvolution and additional sharpening techniques.

Original image scale of the undrizzled image was about 3 arcseconds.

Clear skies
Wolfgang

Helpful

Wei-Hao Wang

Sep 19, 2022

Pixel scale is only part (actually, a small part) of the factors that determine you final resolution, as others had already pointed out. There are many situations where you can reduce your pixel scale (drizzle, or just getting a camera with smaller pixels) but do not get more details. Since your case does not involve spending extra money to get more hardware, you may just try with and without drizzle and see what's the difference. This trial is essentially free and there is no reason not to try and see.

I dither every exposure. Unless you expect to get many tens of exposures, I suggest to dither every one of them.

Helpful Concise

Wei-Hao Wang

Sep 19, 2022

but keep in mind that most all reducers will degrade your image off-axis.

Not just off-axis, many reducers even degrade the on-axis resolution (in terms of FWHM in arcsec, not FWHM in pixels). From this point of view, drizzling images taken with a reducer often doesn't help at all.

That being said, SharpStar's FRA is a rare example where a reducer improves image quality, both on-axis and off-axis. This is an interesting behavior, not an exciting one, because its resolution without a reducer isn't excellent to start with. SharpStar made some very interesting design choice here.

Well Written Helpful Insightful Concise

Frédéric Auchère

Sep 19, 2022

For what it's worth, and if that helps, here is a comparison of native / drizzle x4 (with half size square droplets)

https://astrob.in/6d666u/E/

As written before, drizzle will help in case the image is undersampled, which is the case here. The native sampling of my Samyang 135 / ASI178 combination is 3.8"/pixel and the drizzled sampling is 0.96"/pixel. That particular acquisition was drizzled at every frame. You do not necessarily need that many, but you need that the coverage of the image plane by the droplets is uniform enough. In this case, drizzle allowed a correct sampling of the image and thus deconvolution and a gain in resolution.

CS

Frédéric

Helpful Insightful Respectful