Observatory Tools

How your stars land on pixels

Pair a telescope with a camera and see the image scale it produces — then see the star itself, rendered on the actual pixel grid your sensor would record.

The Setup
mm
Effective focal length: 250 mm at f/4.9
µm
Best FWHM
″
Worst FWHM
″
Read this off a night's subs — ASIAir reports it beside average star size.
The Result
2.39
arcsec
per pixel
Optimal
One square = one pixel

A 3″ star spreads across 1.3 pixels

Stars need roughly two to three pixels across to render as round. Fewer and they turn blocky; many more and you are magnifying blur rather than recording detail.

Sampling Scale
Ideal window 0.67″ – 2.00″
Over-sampled
Well matched
Under-sampled
—
Field of view
—
Frame size
—px
Star FWHM
—×
Nyquist ratio
Good pairing. Stars will sample cleanly at this scale.

Reading the result

Below the window

Over-sampled

Each star is spread over more pixels than the atmosphere justifies. Star shapes stay pleasing, but you are trading away field of view and splitting the same photons across more wells, so every sub needs longer to reach the same signal. Binning or a focal reducer pulls you back.

Inside the window

Well matched

A star covers enough pixels to record as a smooth disc without wasting resolution on seeing blur. This is the range where sub length, field of view, and star shape all sit in balance — and where guiding error stays a small fraction of a pixel.

Above the window

Under-sampled

Stars land on too few pixels and take on square, angular edges. Fine detail is lost for good, since no amount of processing recovers what the grid never separated. This is the worse failure of the two — a longer focal length or smaller pixels is the fix.

image scale = ( pixel size µm ÷ focal length mm ) × 206.265

The ideal window is derived from your seeing: the coarse limit is half the best FWHM you expect, and the fine limit is a third of the worst. Sampling below two pixels per star is where square stars begin.