What can you do with the result?
Keep the screen and assumptions fixed, then change only the seating distance. A closer seat makes the image occupy a larger angle and increases this model’s pixel estimate. The raster comparison shows how many horizontal pixels Full HD, 4K UHD or 8K UHD provide relative to that estimate. Its percentage describes pixel count, not a picture-quality score.
Understand the calculation
Why acuity alone is only the starting point
The familiar estimate asks how fine a detail the eye can distinguish at a given image angle. With the reference setting visus 1.0, the model uses one arcminute, or 1/60 of a degree. An image spanning 44.4° therefore gives 2,664 horizontal detail positions in this simple estimate. This is the acuity-only value shown beside your current FGRD result.
Fine patterns need enough sampling points
A bright and a dark line form a repeating detail. Sampling such patterns too coarsely can produce false structures: moiré, jagged edges or flicker as the image moves. These effects can draw attention even when individual pixels are hard to distinguish. The Nyquist–Shannon sampling theorem relates recoverable detail to sample spacing for a band-limited signal; in practice, filtering and sampling density work together.
FGRD brings this sampling consideration into the acuity estimate through Ni. At Ni = 0.5 the model doubles the sample count. It then adds the diagonal factor: the diagonal-neighbour spacing on a square grid is √2, about 1.41, times the horizontal spacing. FGRD uses 1.41 as an adjustable model assumption for diagonal detail. The combined default multiplier is 1.41 / 0.5 = 2.82. At 44.4° and visus 1.0, this gives about 7,512 horizontal pixels instead of 2,664.
Use the resolution to support the viewing experience
Choose the image size and seat that give you the view you enjoy, then compare resolution at that setup. The aim is for the content to hold your attention while pixel structures and sampling artefacts recede. Check demanding material such as diagonal edges, fine textures and moving patterns. FGRD is a planning model: source aliasing, scaling, sharpening and display processing also affect the result, and more display pixels cannot recover detail already lost in the source.
From your setup to the formula
FGRD stands for Florian’s Geeky Resolution Demand. It combines the horizontal field of view with a chosen visual-acuity value and two explicit sampling assumptions. Screen diagonal and aspect ratio determine the physical image width. In distance mode, that width and your seating distance determine the angle. In FOV mode, the angle determines the seating distance.
N = FOV × 60 × Visus × DiaF / Ni
FOV is measured in degrees. The model uses 1/Visus arcminutes as its angular resolution assumption. Ni is a divisor: 0.5 increases the count by 2×. DiaF multiplies the estimate to model diagonal structures. The historical defaults are 0.5 and 1.41. These factors are adjustable model parameters, not universal display requirements.
Interpret the result
The primary result is a horizontal pixel estimate. Vertical pixels and megapixels assume the same sampling density on both axes and the selected aspect ratio. The Full HD, 4K UHD and 8K UHD comparison uses raster width only; it does not rank displays. Optical blur, anti-aliasing, motion, contrast and viewing conditions can change what is visible.
Original slides from the talks
The original English slides remain available as an illustrated explanation alongside the calculator. Open a slide at full size to read its details.
View original slides
For the distance example on the slides: at 16:9, 2.5 picture heights correspond to about 39.1°, while 44.4° correspond to about 2.18 picture heights. The calculator uses the geometric relation above. Visual-acuity values in arcminutes must be divided by 60 to convert them to degrees.