Pixel size

[jambouk]

Binning increases sensitivity, reduces resolution/detail....

So, for small bright objects, like Jupiter, is there an advantage to having smaller pixels, as there are plenty of photons, and spreading the image over more pixels will improve the amount of detail resolvable?

Jd

[Cornelius Varley]

Smaller pixels would suggest better resolution, the limiting factor, though, is the resolving power of the telescope. A quick play with the numbers suggests that to get the best arcsecond/pixel ratio with the neximage 5 you need to shoot at a shorter focal length than a typical webcam. Typically you need to image at 0.5 to 0.33 of the resolving power of the telescope per pixel to record the maximum detail. Imaging at higher ratios (> 0.33) doesn't give greater detail, just more pixels recording the same level of detail (oversampling)

[jjongmans]

Binning doesn't increase sensitivity, it lowers noise.

[jambouk]

Thanks both.

I found this website (http://www.astro-imaging.com/Tutorial/Calculations.html), and interestingly at the bottom of the page is a spread sheet to download, and plug in the numbers for ones telescope and imaging camera. For the C11 and ASI ZWO 120MC, it generates the following data:

Theoretical resolution (TR) = 0.28 " / pixel

Field of view (FoV) = 5.9 ' x 4.4 '

If I swap the ZWO for a camera with much bigger pixels, say 6.5 um x 6.5 um, it generates the following results:

TR = 0.48 " / pixel

FoV 10.2 ' x 7.7 '

But I don't think the TR for the colour ZWO camera is correct because of the Bayer matrix reducing the resolving power of the underlying sensor.

And as my own experiments and trials have shown, this is all VERY dependent upon the seeing on the night in question.

JD

[ChrisLX200]

Yes local seeing conditions dictate results. If you have 0.5 arc/sec seeing I'll move house to live where you are

ChrisH

[BlueAstra]

Smaller pixels would suggest better resolution, the limiting factor, though, is the resolving power of the telescope. A quick play with the numbers suggests that to get the best arcsecond/pixel ratio with the neximage 5 you need to shoot at a shorter focal length than a typical webcam. Typically you need to image at 0.5 to 0.33 of the resolving power of the telescope per pixel to record the maximum detail. Imaging at higher ratios (> 0.33) doesn't give greater detail, just more pixels recording the same level of detail (oversampling)

I think you also have to factor Seeing into the equation. For my scope (AG10), the theoretical resolution is 0.65 arcsec which would imply a pixel size of 2.4 um at 0.33. However, typical seeing is maybe 2 - 4 arcsec, which means you'd be looking for 0.7 - 1.3 arcsec with the 0.33 factor, which would be 3.25 - 6 um with my scope. The actual camera I'm using has 7.4 um pixels, giving a resolution of 1.6 arcsec, which is perhaps a tad too low for typical conditions.

[JamesF]

Theoretical resolution gets a bit messy once you add the CFA into the equation.  I agree with the basic result though -- the ASI120MM would be 27.6 arcseconds per pixel in the C11.

Oversampling a small amount may well be a good thing for planetary imaging, but smaller pixels are not necessarily better.  A group of four 3um pixels for instance may well not collect as many photos as a single 6um pixel because the actual working photon collection area may be smaller for the group of four small pixels than for the single larger one.

James