rayleigh criterion vs image sharpness

[pipnina]

I've been taking snaps with my 250PX recently and found that the telescope is possibly not as sharp as the rayleigh criterion would have me believe. At 254mm and a 6000x4000 APS-C sized sensor I have a pixel scale of  0.69" and a rayleigh criterion of 0.66" (at hydrogen alpha line, to give a pessimistic estimate, 0.44" at far blue for optimistic estimate). I have found that the camera is observing a rather large blur around objects (not the somewhat sharp (or possibly under-sampled) image I was expecting.

This image was taken from the top of my driveway, it shows the top of a communications mast some miles away. There is a rather large level of "blur". The telescope was stable and I used 1/250th shutter so I doubt that it is wobble. I spent some time trying to focus. It also shows the center of the frame, with the telescope collimated (primary only) before hand.

Bottom of the frame (below) seems to show yet more blur (i think) possibly coma or perhaps mis-aligned secondary?

 

On astronomical terms, this image is of Mizar A&B:

Some clear atmospheric effects seem to be at play, making it hard to identify whether it is the same problem as before or just turbulence. 1/6th of a second exposure. [EDIT: It is 1/60th not 1/6th]

And Algieba: same exposure

Finally Jupiter. I think the atmosphere's effect is easily visible here but still hard to assess for me.

 

Do you guys think these images look right considering the telescope's properties? I'm not using a coma corrector but these photos all show roughly the center of the frame. There are rather large "blurs" around objects which is my main concern.

 

Cheers,

[beka]

Hi pipnina,

You should be able to get better pictures with your scope. In the terrestrial shot I think camera shake can still be a factor even at 1/250th of a second. Your pictures of the stars and Jupiter could be focusing accuracy. How did you attach the camera to the scope?

Cheers!

 

[pipnina]

23 minutes ago, beka said:

Hi pipnina,

You should be able to get better pictures with your scope. In the terrestrial shot I think camera shake can still be a factor even at 1/250th of a second. Your pictures of the stars and Jupiter could be focusing accuracy. How did you attach the camera to the scope?

Cheers!

I'm using a T2 adaptor for my camera's mount that screws in to the base of my 2" to 1.25" adapter. So it's Nikon D3200 -> T2 adaptor -> 2"/1.25" adaptor base -> focusser. I have excess focus inwards and outwards.

I've noticed my focusing not being perfect. I took a 30-second exposure of polaris which showed two diffraction spikes per side instead of one but I focused individually for Algieba, Mizar and Jupiter.

With bad wobble one night I took a 1/3 second exposure of M42. It wobbled and made a very long spaghetti star trail. Dividing that length by nearly 100 might be close to that level of clarity, but even zoomed in through live view on the communications tower showed the same blur, with little wind and no visible wobble.

[vlaiv]

Welcome to the world of high res imaging

Forget Reyleigh criterion when working with imaging. There are two types of resolutions that are important in imaging. One is so called critical resolution for planetary work, and that is usually quoted as 1/3 of Airy disk radius. I've done some research and found that actual figure is closer to 1/2.4. For your scope that would be around 0.22"/pixel.

For this type of imaging you need very stable atmosphere (even if you shoot distant terrestrial object, there is seeing like turbulence), and exposure times of order of 5-20ms (or 1/200 - 1/50). You need a lot of them stacked, and result will still be blurry! Then you proceed to restore frequencies (because telescope optics allow to capture frequencies up to a point) - this is usually done by deconvolution and / or wavelet sharpening.

For long exposure imaging, achieved resolution depends on 3 factors - seeing, scope aperture (and quality of optics in general) and quality of your mount (tracking quality). For this type of imaging, resolution that is really achievable in 95% of cases for amateur equipment is somewhere around 1"/pixel - if you plan to go below, you'll need exceptional sky and exceptional mount (and guiding) alongside your 10" scope.

These should be considered theoretical maxima. In reality you can use "finer" resolutions, but blur will be consequence - you simply will not capture additional detail. You will often work with conditions that are less then perfect and even these max resolutions will be too much - atmosphere will not cooperate, scope does not have perfect optics ....

On top of all of that - focus is critical! In all above images, except Jupiter, I would say that focus is off (next to other things like seeing). Jupiter is as good as it gets in average - good seeing. To do better than that in single frame, you really need very good / exceptional seeing (and perhaps Jupiter being a bit higher to lessen atmospheric dispersion evident in frame).

 

[pipnina]

3 minutes ago, vlaiv said:

Forget Reyleigh criterion when working with imaging. There are two types of resolutions that are important in imaging. One is so called critical resolution for planetary work, and that is usually quoted as 1/3 of Airy disk radius. I've done some research and found that actual figure is closer to 1/2.4. For your scope that would be around 0.22"/pixel.

For this type of imaging you need very stable atmosphere (even if you shoot distant terrestrial object, there is seeing like turbulence), and exposure times of order of 5-20ms (or 1/200 - 1/50). You need a lot of them stacked, and result will still be blurry! Then you proceed to restore frequencies (because telescope optics allow to capture frequencies up to a point) - this is usually done by deconvolution and / or wavelet sharpening.

I could attempt video capture, but my camera is limited to 1920x1080 @ 24fps. The frame rate isn't so bad but the 3x shrinking of resolution would be a killer. I used some scale-aware contrast boosts on the jupiter image (it basically just brought the bands out a bit better) i think this is somewhat similar to wavelets.

50 minutes ago, vlaiv said:

On top of all of that - focus is critical! In all above images, except Jupiter, I would say that focus is off (next to other things like seeing). Jupiter is as good as it gets in average - good seeing. To do better than that in single frame, you really need very good / exceptional seeing (and perhaps Jupiter being a bit higher to lessen atmospheric dispersion evident in frame).

Hmm, It's one thing to be able to focus on stars (there are tools I could use, though bahtinov masks are a bit confusing.) but focusing on the tower proved rather difficult. I kept moving the focus back and forth past where it appeared to become most sharp, but it never got any better than the images attached.

I'm not sure if the Mizar and Algieba shots are out of focus or not. What tips you off to it being a focus problem and not just turbulence?

[vlaiv]

5 hours ago, pipnina said:

I could attempt video capture, but my camera is limited to 1920x1080 @ 24fps. The frame rate isn't so bad but the 3x shrinking of resolution would be a killer. I used some scale-aware contrast boosts on the jupiter image (it basically just brought the bands out a bit better) i think this is somewhat similar to wavelets.

Hmm, It's one thing to be able to focus on stars (there are tools I could use, though bahtinov masks are a bit confusing.) but focusing on the tower proved rather difficult. I kept moving the focus back and forth past where it appeared to become most sharp, but it never got any better than the images attached.

I'm not sure if the Mizar and Algieba shots are out of focus or not. What tips you off to it being a focus problem and not just turbulence?

Couple years ago when I started fiddling with imaging, I took an image of Alcor / Mizar pair with fairly simple equipment (SW 130/900 reflector on Eq2 mount and QHY5IIL color cam). I don't think I have that image any more, but I remember Mizar A/B pair being well separated and stars tight. It looked something like this:

This is from stellarium but looks very much like image I made. So there were at least 2-3 times radius of individual stars between them. Look at your image. There is hardly any space between.

Also count pixels across star radius - it is about 20, and if your resolution is 0.69"/pixel - that would mean that star size is ~14" across. That is huge! No wonder there is no space between components (which have separation of 14.4").

Now extremely poor seeing might be able to cause something like that, but you would have to shoot above a chimney of a house with smoke coming out to get that kind of atmospheric disturbance. Other more plausible explanation is that star is out of focus, and instead of concentric rings being seen - it has been blurred by seeing into blobby shape. You can almost see that it is doughnut shaped in above images.

It might be due to very bad optics of telescope (like severe spherical), but in that case you would not be able to make Jupiter image like you did.

If I'm not mistaken 10" Newtonian (250px) is F/4.7 scope? At that speed focusing is rather critical, so critical zone is only 55 microns - or 1/20 of single mm. Do you have dual speed or single speed focuser fitted to that scope?

 

[beka]

Regarding focusing I have found a Bahtinov mask to be very useful. Also did you use the feature of your DSLR to magnify the live view to make focusing easier? Finally I believe your Nikon has a flipping mirror during exposure so using mirror the mirror lock feature is advised for astrophotography.

Best

[pipnina]

8 hours ago, vlaiv said:

If I'm not mistaken 10" Newtonian (250px) is F/4.7 scope? At that speed focusing is rather critical, so critical zone is only 55 microns - or 1/20 of single mm. Do you have dual speed or single speed focuser fitted to that scope?

Yep that's my scope. That's also a very narrow focusing margin! Unfortunately I only have a single-speed focuser fitted. It seems to manage about 12 mm of travel per turn. So that would be 0.01~mm of travel per mm of focus dial rotation (measuring the dial as 21mm radius). So I would have to get the focus to within 5mm rotation if I've measured everything correctly (which I probably haven't).

[vlaiv]

19 minutes ago, pipnina said:

Yep that's my scope. That's also a very narrow focusing margin! Unfortunately I only have a single-speed focuser fitted. It seems to manage about 12 mm of travel per turn. So that would be 0.01~mm of travel per mm of focus dial rotation (measuring the dial as 21mm radius). So I would have to get the focus to within 5mm rotation if I've measured everything correctly (which I probably haven't).

Not sure if you have that right, I did not do the math, but I have 2 speed (1/10 reduction) fitted to my 8" F/6 dob. I was observing last night (after almost half a year of lousy weather) and I noticed that even about 1/10 of a turn (that roughly being 30 degs) of reduction knob would throw good focus out (was observing Jupiter among other things).

Let's do the math. 12mm of travel per turn would make focuser shaft diameter ~3.82mm - that sounds about right for standard crayford.

Plain ratio of travel to radius should do so 1.9 mm : 0.055 mm = 21 mm : X

If we solve this for X, result is 0.6 mm so you missed by order of 10 (if I'm not mistaken, but that kind of agrees with what I remember about precise focus prior to fitting 1:10 reduction on my - it was literally a touch to get it to right place).