Imaging scope dilema - GSO/SN shootout

[adamsp123]

Here is the first comparison as threatened, 2 images of M1 - taken with a F/4 GSO 200mm Newt and my Meade 10" Schmidt Newt, both images have 9x5min subs, MPCC and CLS LP filter, canon modded 1000D ISO 800 used on both scopes.

I am very interested in you opions/comments

Excuse my lack of processing skills I haven't matched the colours and in both cases I kept post processing down to a minimum.

First pic GSO second the SNT.

[David Pavlich]

The SN has more detail and looks closer to what M1 looks like. There is green in M1, however, so you may have done a bit better stretch in the Newt image.

Stars look good in both and I admit, I like diffraction spikes. But I have to go with the SN image. I would try stretching it differntly to see if you can bring out any green.

David

[Beyond_Vision]

The SN image is better but I think the focus is a litle bit out on the GSO image

Regards

Kevin

[omneferuss]

The Newt' for mine.. as presented. Matching scale and orienting the same way up would make comparison easier, as would removing differences in process-result...

..but that just goes to show that there's variability in each and every aspect of every image, and there is no absolute benchmark for anything in this game

[adamsp123]

Here is a quick rehash of the SN taken image matching the colour more of the Newt image and rotated.

Its looks to my sorry eyes that the SN shows a bit more detail, I do lile spikes but in choosing between the 2 scopes that is not going to sway my decision as there are ways of adding spikes artificially.

I have yet to do a comparison image of M1 with my 120ED, waitng for a hole in the clouds

[step_hen]

SN looks better to me, but then i am biased!

Stephen

[Deneb]

The SN looks a lot better.

[Whippy]

The SN looks like the better image but I agree with Kevin that the GSO image looks a little out of focus.

Tony..

[adamsp123]

Regarding the focus I think what happened I used my Bahtinov mask but the image was done on one of those very cold (-15C) nights and I think as the tube contracted the focus went out, darn it, looks like I need to repeat the GSO image to get a proper comparison.

[michael.h.f.wilkinson]

Apart from focus issues, the 10" SN will tend to show more detail, simply because it is bigger. I would certainly go with the SN

[Whippy]

Apart from focus issues, the 10" SN will tend to show more detail, simply because it is bigger.

I'm curious, how does aperture rule for imaging?

Tony..

[adamsp123]

On this topic of aperture, will the SN 10" when used for imaging show noticable better resolution/detail say than my 8" GSO or 120ED refractor?

[michael.h.f.wilkinson]

Higher aperture gives better resolution regardless of whether you are imaging or using it visually (you look at the planetary results for a C8 vs a C11 or a C14). If you are sampling the image in the image plane at the Nyquist frequency, you should get more detail out of a larger aperture (because of the diffraction limit, you optic's PSF diameter will be proportional to the ratio of the wavelength to the aperture).

Of course in DSO imaging we are not always at the Nyquist frequency, but the modulation transfer function (MTF) of optics (which governs general crispness of the image) at lower spatial frequencies will tend to be better with increasing aperture, if you compare optics with the same kind of secondary obstruction and overall correction. A scope without secondary obstruction (like an APO) will have better MTF at lower spatial frequencies, leading to the increased apparent crispness, even though the resolution of a 120mm will be about half that of a 10" SN.

Bigger diameter also can lead to better signal-to-noise (S/N) ratio (at the same image scale), because you have a larger "flux bucket": 2x aperture -> 4x the number of photons -> 2x better S/N in photon-noise limited imaging

[catburglar]

The largest aperture will always have the highest theoretical resolution limit- nothing beats the laws of physics on that one.

However for most DSO imaging, the aim is to sample at around 1-2 arcsec/pixel (or sometimes even lower) in order to get good S/N ratio.

Image scale is dependent on the effective focal length and the pixel size of the imaging camera. Given that your scopes have focal lengths of 800, 900 & 1000 mm- you can calculate the image scale from the formula: Scale (arcsec/pixel)= 206325*pixel size (micrometers)/focal length (mm)

The comments about S/N ratio don't quite hold true however. For visual use increasing aperture gives brighter images (at the same magnification), but for imaging, brightness is governed by the F/ratio of the optics (excluding light losses due to differing optical designs), but this can be compensated for (to some degree) by increasing the exposure times, which obviously then makes guiding more important... the list of compromises goes on.

[michael.h.f.wilkinson]

The comments about S/N ratio don't quite hold true however. For visual use increasing aperture gives brighter images (at the same magnification), but for imaging, brightness is governed by the F/ratio of the optics (excluding light losses due to differing optical designs), but this can be compensated for (to some degree) by increasing the exposure times, which obviously then makes guiding more important... the list of compromises goes on.

This is why I stated: at the same image scale (i.e. arcsec/pixel)

[Whippy]

Higher aperture gives better resolution regardless of whether you are imaging or using it visually (you look at the planetary results for a C8 vs a C11 or a C14). If you are sampling the image in the image plane at the Nyquist frequency, you should get more detail out of a larger aperture (because of the diffraction limit, you optic's PSF diameter will be proportional to the ratio of the wavelength to the aperture).

Of course in DSO imaging we are not always at the Nyquist frequency, but the modulation transfer function (MTF) of optics (which governs general crispness of the image) at lower spatial frequencies will tend to be better with increasing aperture, if you compare optics with the same kind of secondary obstruction and overall correction. A scope without secondary obstruction (like an APO) will have better MTF at lower spatial frequencies, leading to the increased apparent crispness, even though the resolution of a 120mm will be about half that of a 10" SN.

Bigger diameter also can lead to better signal-to-noise (S/N) ratio (at the same image scale), because you have a larger "flux bucket": 2x aperture -> 4x the number of photons -> 2x better S/N in photon-noise limited imaging

Whilst I'd agree that aperture rules for observing (who wouldn't?), it doesn't quite work like that for imaging IMO. If it was the case then a) everyone using small scopes (like my ZS66 & ED80) or camera lenses are essentially wasting their time and Newtonians would be the weapon of choice which they are not.

Focal ratio (speed) and focal length (FOV) are important which of course are related to aperture.

Tony..

[MartinB]

Now you've started something Adam! I agree with other comments that the GSO didn't appear focused but the star shapes were odd as well. The SN10 was a clear winner on this comparison.

Of course in DSO imaging we are not always at the Nyquist frequency, but the modulation transfer function (MTF) of optics (which governs general crispness of the image) at lower spatial frequencies will tend to be better with increasing aperture, if you compare optics with the same kind of secondary obstruction and overall correction. A scope without secondary obstruction (like an APO) will have better MTF at lower spatial frequencies, leading to the increased apparent crispness, even though the resolution of a 120mm will be about half that of a 10" SN.

Bigger diameter also can lead to better signal-to-noise (S/N) ratio (at the same image scale), because you have a larger "flux bucket": 2x aperture -> 4x the number of photons -> 2x better S/N in photon-noise limited imaging

Michael, I'm afraid I'm not up to speed with the optical theories you have expounded here. I suspect that my state of ignorance is matched by over 90% of SGL members.

I do know that the central obstruction does reduce contrast and that edge sharpening filters work largely by increasing contrast along bright and dark borders. So I can accept that higher contrast can increase apparent sharpness. I also know, from practical experience using SCTs, refractors and newtonians, that, when deep sky imaging, atmospheric seeing knocks any resolution gains from aperture into a cocked hat once you start to move above 1000mm focal length.

As far as s/n goes, what a vexed question that is! My take on this is - if you can get over the read noise hump (which can be a very big if) then signal to noise for a given area of sky is determined by aperture. However, the overall light falling on the chip is determined by focal ratio...assuming the effiency to 2 optical systems is identical (central obstruction, transmission, reflectivity etc etc).

An M1 taken with an ED80 is going to look noisy when blown up to match the size of an M1 through an ED120 using same exposure times. However, if you take an image of an extended area of nebulosity the ED80 image will appear as clean as one through an ED120.

[adamsp123]

Now you've started something Adam!

Sorry about that, actually I am not! this is a very interesting topic as it turns out. Like most things in life, and engineering, it is not as simple as one might expect.

Michael, Like others I really don't understand the more complex parts of the explaination you gave, could you please give a "Carl Sagan" simplified version for the less educated of us on this topic, I understood some of it but several important points went straight over my head.

[michael.h.f.wilkinson]

I am not Carl Sagan, but I will give it a shot:

First of all, when I talk about spatial frequencies, I mean the following: any pattern in any image can be considered the superposition of sine and cosine waves of different frequencies and orientations in the image plane. Wherever they interfere constructively, the image is bright, elsewhere it is dark(er). For simple, grid-like patterns this is easiest to understand, but the principle holds for any image. This principle is used in filters based on Fourier analysis: we speak of a low-pass filter, if it removes the high frequency sines and cosines, thus smoothing the image. A high-pass filter removes or suppresses the lower frequency sines and cosines, giving the image a sharper appearance, but lowering the overall dynamic range.

Any optical system modulates the frequencies in the image it records, and acts as a low-pass filter. The resolution limit is in fact equivalent to the highest frequency waves (in the image plane) which are passed by the "filter." Any finer detail cannot be detected.

Suppose I am imaging sets black and white lines, ranging from very thick, broadly spaced, to very fine, closely spaced lines (these are often used to test optics). Suppose their contrast is 255 (pure white, pure black). For very broad lines (spatial frequency close to zero) the contrast will be preserved almost perfectly. By contrast, detail beyond the resolution limit is imaged as an even grey area, i.e. the contrast in the image is zero. However, just below the resolution limit, the contrast tends to be poor. There is a more-or-less smooth transition from perfect preservation of contrast at low frequencies, to zero contrast preservation beyond the limit of resolution.

To quantify how contrast of line patterns or sine waves is changed by the optics, we use the MTF, or modulation transfer function. This function simply gives the ratio of the image contrast to the actual contrast as a function of the spatial frequency (often specified as lines/mm in the image plane). So if the MTF is 0.5 for 30 lines/mm, this means that lines with true contrast of 255 will appear at a contrast of 127.5 grey levels.

Any optical system has a "comfort zone" in which MTF is high. The broader this comfort zone, the finer the detail that appears crisp in the image. Visually, high MTF at medium frequencies is generally more important than a high resolution limit (though the two are often correlated). An MTF of 0.8 at 30 lines/mm and a resolution limit of 100 lines/mm is generally better than an MTF of 0.5 at 30 lines/mm and a resolution limit of 180 lines/mm.

Many factors influence the MTF. In the ideal case, with perfect correction of all optical errors (we can dream:)), the MTF depends on the focal ratio of the optics, and the secondary obstruction. Given two similar systems in terms of secondary obstruction, focal ratio, and level of correction, the MTF in the image plane will be roughly the same.

So in the case of an 8" vs 10" Newtonian design, both at F/5 with the same central obstruction (in %), both will have the same MTF and reslotuion in the image plane. The difference in the case above stems from the different resolutions not in the image plane but in the sky. Filaments in the 10" image will appear larger, and therefore lie more in the "comfort zone" of the optics, because they appear at lower spatial frequencies.

By contrast, if the 8" and the 10" are F/5 and F/4 respectively (i.e. they have the same focal length) the MTF and resolution in the image plane of the 10" will tend to be higher.

APOs have no central obstruction which leads to the best possible MTF in mid frequencies, even though their resolution limit is lower.

I hope this helps. I will ramble about S/N later;)

[David Pavlich]

I knew there was a reason why research telescopes keep getting bigger!

David

[michael.h.f.wilkinson]

I knew there was a reason why research telescopes keep getting bigger!

David

and they have the advantage of MASSIVE, seriously cool CCD chips. Check this beast ( http://www.astro.rug.nl/~omegacam/ ) out. One square degree on the VLT! Built at my uni!