Pages: 1 [2]   Go Down

Author Topic: Help me understand resolving limits  (Read 3249 times)

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: Help me understand resolving limits
« Reply #20 on: February 09, 2012, 10:14:04 pm »

Hi,

Thanks for a lot of good explanations!

What is pixel pitch is need to correctly sample a given PSF in your view? Reason I'm asking is partly that I see quite a lot of aliasing with my Sony A55 SLT which has 4.77 micron pixels.

Best regards
Erik


BTW if you are getting a PSF width of 5.5 microns, then 4.7 micron pixels are undersampling the PSF. But you'll probably find that atmospheric turbulence and tracking errors broaden the PSF a lot anyway, before it reaches the camera.

Ray
Logged
Erik Kaffehr
 

Fine_Art

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1172
Re: Help me understand resolving limits
« Reply #21 on: February 09, 2012, 11:11:05 pm »

Thanks for the help. It's appreciated.
Logged

ondebanks

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 858
Re: Help me understand resolving limits
« Reply #22 on: February 10, 2012, 05:04:42 am »

Hi,

Thanks for a lot of good explanations!

What is pixel pitch is need to correctly sample a given PSF in your view? Reason I'm asking is partly that I see quite a lot of aliasing with my Sony A55 SLT which has 4.77 micron pixels.

Best regards
Erik


On a monochrome sensor, you'd aim for the usual Nyquist sampling, 2 pixels across the PSF FWHM. Some cameras sample less than this (e.g. the Hubble Space Telescope imagers), the designers having decided that PSF undersampling is a worthwhile trade-off, in order to gain a larger field of view on the fixed sensor size. On all but the smallest ground-based telescopes, turbulent atmospheric seeing makes the PSF size unpredictable, so the design would probably aim for Nyquist sampling in typical or slightly-better-than-average seeing. On nights of bad seeing it will oversample, on nights of exceptional steadiness it will undersample (but undersampled good seeing beats correctly-sampled mediocre seeing any day...or night!)

On a colour sensor with a CFA, there's no right answer to completely avoid colour aliasing. Nomatter how much you oversample the PSF, different parts/intensities of it are landing on different coloured pixels. It's down to the interpolation algorithms to guess the "right" RGB for each pixel. My guess is that with no AA filter and sampling of around 3-4 pixels across the PSF FWHM, the algorithms would do a good job of this, and colour aliasing would be subdued. It's just a guesstimate, based in part on my experience with my MFDB: for example I have images of tree branches shot with my 55-110mm zoom, with stark colour aliasing at 55mm, and none at 110mm from the same tripod position. Doubling the focal length doubled the number of pixels that each branch fell on - doubled the sampling of the "branch spread function"(!) if you will - and that was enough to eliminate aliasing. The branches at 55mm were only about 1.5-2 pixels across, which is how I end up with the 3-4 pixels estimate for the sampling to avoid moire. It's rough since there may have been other variables at play as well - focus, wind, lens aberrations.

Ray
« Last Edit: February 10, 2012, 06:37:56 am by ondebanks »
Logged
Pages: 1 [2]   Go Up