Expected HFR Matrix

Predicted NINA-style HFR in pixels for each telescope and sensor, from seeing, diffraction, guiding and pixel size. Tap a cell for the full breakdown.

Seeing (FWHM)

2.5″

Typical conditions

Forecast for a US ZIP code

Guiding RMS (total)

0.50″

Filter

HFR definition

Binning

Undersampled (FWHM < 2 px) Well sampled (2–5 px) Oversampled (> 5 px) Big number = HFR px · small = FWHM px · MP in each column header updates with binning

Your equipment

Add telescopes and cameras to the matrix, or remove ones you don't use. Changes are saved in this browser.

Brands

Manufacturer focal lengths, plus the matching reducer or flattener where one is sold. Entries marked ≈ are aperture × focal ratio where the maker doesn't publish a focal length. Plate-solve your own rig for exact numbers.

Telescopes

Cameras

Back up, move or reset your list

Copy this JSON to move your list to another browser or to the Docker version, or paste a saved list and load it.

Model

scale″/px   = 206.265 × pixel_µm / focal_mm
diff″       = 1.03 × λ / D × 206265            (Airy FWHM, obstruction ignored)
guide″      = 2.355 × guiding_RMS              (RMS → FWHM)
pixel″      = 0.68 × scale                     (square-pixel blur, 2.355/√12)
FWHM″       = √(seeing² + diff² + guide² + pixel²)
HFR_true″   = FWHM / 2                         (Gaussian star)
HFR_NINA″   = 1.0645 × FWHM / 2                (flux-weighted mean radius)
HFR_px      = HFR″ / scale

Real stars run larger. Stars have Moffat-like wings, and background noise inflates a weighted-mean radius, so expect measured NINA HFR to come in somewhat above these values. Use the matrix for relative comparisons and sampling, not as a pass/fail target.
Color sensors match mono in this table. Behind a Bayer matrix each red or blue channel is sampled at twice the pixel pitch, so undersampled cells are worse in practice on MC/C cameras.
Fast optics aren't modeled. The RASA 8 (f/2) and Pleiades 68 (f/3.8) have real optical spot sizes and are very sensitive to tilt and backfocus, which typically adds blur beyond this model.
Correctors change focal length. Newtonian coma correctors and RC flatteners shift focal length a few percent, and GSO RC10s plate-solve anywhere from about 1950 to 2030 mm. Plate-solve your own rig and use that scale.
Binning: the Binning control multiplies every pixel size (IMX571 at 2×2 = 7.52 µm). On these CMOS cameras binning is done in software, so it changes sampling and per-pixel SNR but not read noise per binned pixel. The 294MM's default mode is already bin2 (4.63 µm), so its "IMX492 bin1" column at 2×2 matches the 4.63 µm column at 1×1.

What's a good HFR?

There's no single best number, because HFR in pixels depends on your plate scale. Aim for a range in pixels, so stars are sampled well, and then as small as you can get in arcseconds, which is how sharp they are. The pixel ranges below follow the NINA HFR setting above.

HFR (px)	Star FWHM (px)	What it means

In arcseconds, smaller is better, but seeing sets the floor. Multiply HFR in pixels by your plate scale to compare setups. At typical backyard seeing of 2.5–3″, an HFR of about 1.4–1.8″ is very good; over about 2.5″ usually points to focus, guiding or tilt.
Measured HFR runs higher. Expect NINA's numbers to come in 10–30% above this calculator, because real stars have broader wings than the model.
For focus, watch the trend, not the number. Best focus is the lowest HFR you see that night. A rise of 10–15% during the session means it's time to refocus, usually because temperature has shifted the focus.
Compare like with like. HFR also changes with target altitude, filter and star brightness, so only compare frames taken under similar conditions.