Talk to me about resolution

[Cath]

Ah, well, that depends on what you mean by "sharpness", I guess

I've not yet looked into the algorithms used in astro photography but I'm currently writing some capture software for canon cameras and in it I'm measuring the absolute amplitude in the change of brightness (the steepness of the edges) on none saturated stars, the larger the amplitude (steeper the edge) the more in focus you are.

[JamesF]

The "ideal" method that appears to be used for astrophotography appears to be based on the idea that an "ideal" star will produce an airy disc and a set of diffraction rings on the image plane. The FWHM figure is the measurement from one side of the airy disc where the light intensity is at half that at the centre to the opposite side where the same applies. That is, if you draw a circle centred on the airy disc and with a radius such that the circle lies at the point where the light intensity is half that at the centre, FWHM is the diameter of the circle. Focus is best when FWHM is minimised.

It strikes me that this method is in many respects measuring the steepness of the edges just as you're doing. I don't know why it is used as opposed to any other way of gauging the rate of change of brightness, though it does seem to be a fairly standard signal-processing concept. Even in tHoAIP Berry and Burnell just say "this is how it's done" rather than "this is why it's done this way" and they work through an awful lot of stuff from first principles. It might be on the chapter on photometry, but at that stage I was struggling to stop my brain dribbling out of my ears. Perhaps other methods struggle to ignore the diffraction effects towards the edge of the star image? I'm guessing.

James

[JamesF]

Ugh. Re-re-re-reading that, it wasn't written very well. Hopefully it's not unclear.

James

[Earl]

From what I have been reading recently sub 2arcseconds is very demanding to get seeing normally been limited?

[swag72]

Makes sense James - Thanks for taking the time to answer.

My bag at the moment Earl is that with one scope I am getting 1.4-1.7 FWHM on 20 minute narrowband subs. When I put on the bigger scope, I was shocked at the difference in FWHM between the two scopes. I know now that my focusing technique is fine, which takes one thing out of the equation. So that leaves me thinking that with the 900mm focal length and the 314L+ I'm going to end up with permanent soft looking subs.

Having a bit of a think at the moment whether I even like 900mm focal length!!!

[NickK]

Seeing, to me, is a visual thing.. If you know something is a point source, you know it's basic airy disc as the 'correct' final image.. you can create a function that corrects and therefore you can undo the seeing. Probably not going to be perfect but it works.

[JamesF]

If you have time Sara, and feel the inclination, you could try taking an image of the same target with each scope. Scale the ST120 image down to match the size of the Pentax image, crop them to the same size and post the two without saying which is which. See which one people think is better

James

[Cath]

No it's fine James, I got exactly what you are saying, it makes perfect sense.

I guess doing the same as everyone else means you end up with a figure that can be compared no matter what software your using.

Here's a snap shot (had to resize it down so have lost detail) from the image/video capture stuff I'm doing showing the raw images after edge differentiation (the sharper the edges the brighter the white edges are).

The aim of it all is of cause to gain optimum focus no matter what method is used.

[JamesF]

Hmm. Actually, I'm not sure that's the right way to do it. I was trying to think of a way to avoid making it immediately obvious which image came from which scope. Perhaps that's not possible.

James

[swag72]

Or conversely, I coule scale up an existing Pentax image to the same scale as the 120ED I suppose, just to get an idea.

How would I go about knowing what size to scale it up to?

[JamesF]

That's very interesting, Cath. It will be interesting to see how it works out.

From what I've read today I'm getting the impression that the FWHM method is kind of attempting indirectly to measure the standard deviation of the point spread function, the idea being that the sd will be minimised when the focus is best.

James

[JamesF]

Or conversely, I coule scale up an existing Pentax image to the same scale as the 120ED I suppose, just to get an idea.

How would I go about knowing what size to scale it up to?

Well, given the figures that CCD Calc came up with for the image scale, I'd try scaling it up by a factor of 2.64/1.48, or about 1.78. I'm not sure how well it will work, but it may be informative. You'll never know without trying

James

[JamesF]

Here's another random thought. The maximum resolution of the ST120 is, I think, 1.12 arcseconds, whereas the Pentax is 1.69 arcseconds. At the calculated image scales that means you have a tiny bit of oversampling (1.32x) with the ST120, but more with the Pentax (1.56x). Perhaps that's why the Pentax image looks better.

James

[NickK]

There's also one other point here - quality of optics..

I think you'll be able to push the 75 far harder than the 120 if my experience with the 105 is anything to go by.

[swag72]

I've always wondered about the optics of the SW120ED if I'm honest. I have spent MANY hours trying to find something to replace it with and usually draw a blank or need a mortgage!

[NickK]

Here's another random thought. The maximum resolution of the ST120 is, I think, 1.12 arcseconds, whereas the Pentax is 1.69 arcseconds. At the calculated image scales that means you have a tiny bit of oversampling (1.32x) with the ST120, but more with the Pentax (1.56x). Perhaps that's why the Pentax image looks better.

I agree, However I don't think you loose data as such with the sub-dawes limit but you have to undo the effect to recover it. Obviously there's a point where the convolution is so bad that you loose information to noise..

[Earl]

What reducer / Flattener are you using with the 120?

[swag72]

I'm not using one at the moment as I want to image at the native 900mm. I do have the SW x0.85 reducer, and while the speed of f6.7 is nice, the focal length of 765mm just isn't long enough for me. I want something very different to my small scope.

[ollypenrice]

If you are trying to image at a resolution below that allowed by the seeing (ballpark 2 arcseconds per pixel?) then you are not gaining anything and would be better with binned pixels, or bigger pixels, both of which are more sensitive. However, you are not really losing anything either, you are just not at maximum effiency.

Whatever the FWHM reading, the resolution is better in comparable larger scopes - at least from personal experience. The TEC 140 mightily out resolves the Tak 85 even though the Tak is a better astrograph. If you resize a TEC image of, say, the Trifid and apply it over the top of a Tak widefield including the Trifid, you cannot apply it at anything like 100% because the stars are so much smaller and the detail so much finer. You have to apply the long FL image gently and at a limited opacity. If you have a widefield in the Pentax and a close up of part of that image in the 120 send me them both and I'll Registar the close up onto the wide and send them back.

Olly

[FraserClarke]

One thing to be aware of is that a measure of FWHM < 2 pixels is not very reliable. The smooth shape of the PSF is very undersampled by the CCD, so the measurement is not very sensitive to small changes in the size of the PSF. I wouldn't really believe a FWHM of 0.8 pixels vs 1.2 pixels as a real difference, for example. If you really want to compare which *scope* gives the sharpest image, you want to sample the images so that you a FWHMs of 4 pixels or more -- then you can measure small differences in the FWHM (though that's not the optimal set-up for taking pretty pictures as Olly says).

To compare scopes, you also need to be sure you're using a common unit like FWHM in arcseconds, rather than pixels (which is linked to focal length). Going back to your original post Sara, assuming the FWHM max values you give are in pixels, the FWHM=2.0 pixels on the Pentax is equivalent to 3.6 pixels on the SW (to correct for the relative focal lengths; 900mm / 500mm). So, if you're measuring FWHM=3.0 pixels on the SW, it's actually giving you a *sharper* intrinsic image than the Pentax, but it's just that image is "scaled up" by the longer focal length so sampled at more pixels per arcsecond by the camera.

[swag72]

Thanks FraserClarke.

Olly, I've sent you a couple of images

[ollypenrice]

... and I've put one on top of the other in Registar and sent 'em back! If ever you want to sell that Pentax...

Olly

[swag72]

If ever you want to sell that Pentax...

Never Olly!!!

[ollypenrice]

Never Olly!!!

Surprize surprize... Rather a convncing image, I thought.

Olly