TV size and viewing distance: when resolution stops mattering
Resolution stops mattering at the distance where your eye can no longer separate two adjacent pixels. On a 65-inch 4K set that distance is about 4.25 feet (1.3 m), and from any farther back a 1920x1080 panel of the same size shows exactly as much detail as a 3840x2160 one, because the extra pixels fall below the resolving limit of normal eyesight. It follows from a definition in a current ITU recommendation, not from a review site's rule of thumb.
The one arcminute everything is built on
"Normal" vision, meaning 20/20 in US notation or 6/6 in metric, is pinned to an optotype that subtends 5 minutes of arc. A Snellen E or a Landolt C is built so that "the critical gap that needs to be resolved is 1/5 this value, i.e., 1 arc min." At 6 metres, an eye performing at that level "is able to separate contours that are approximately 1.75 mm apart."
NHK's engineers reached for the same threshold while designing Super Hi-Vision, the format that became 8K. Masayuki Sugawara's 2008 Q2 write-up in the EBU Technical Review reports that their experiments landed "not far from what is generally believed to be necessary, i.e. 1 pixel per 1 arc-minute, or 30 cycles per degree (cpd), but a moderately higher resolution of 40~50 cpd is desirable for our criteria." Two things follow. The 1-arcminute figure is the textbook acuity limit, not a broadcast-industry convenience. And the people who specified 8K decided from the start that it was not a demanding enough target.
What ITU-R BT.500-15 actually says
Recommendation ITU-R BT.500 has existed since 1974 and was last revised in May 2023 as BT.500-15, "Methodologies for the subjective assessment of the quality of television images," still the in-force version. Section 2.1.3.2 defines the design viewing distance, or optimal viewing distance, as "the distance at which two adjacent pixels subtend an angle of 1 arc-min at the viewer's eye; and the optimal horizontal viewing angle as the angle under which an image is seen at its optimal viewing distance."
Table 1-1 gives the results per format in picture heights (H), not diagonals. Five of the nine rows:
| Image system | Reference | Aspect ratio | Optimal viewing angle | Optimal distance |
|---|---|---|---|---|
| 720 x 576 | ITU-R BT.601 | 4:3 | 13 degrees | 6 H |
| 1280 x 720 | ITU-R BT.1543 and BT.1874 | 16:9 | 21 degrees | 4.8 H |
| 1920 x 1080 | ITU-R BT.709 | 16:9 | 31 degrees | 3.2 H |
| 3840 x 2160 (4K UHD) | ITU-R BT.2020 | 16:9 | 58 degrees | 1.6 H |
| 7680 x 4320 (8K UHD) | ITU-R BT.2020 | 16:9 | 96 degrees | 0.8 H |
Every row falls out of one division. A pixel row occupies H divided by the active line count, and one arcminute subtends 1/3437.75 of the viewing distance, so the design viewing distance in picture heights is 3437.75 divided by the line count. For 1080 lines that gives 3.183, printed as 3.2 H. For 2160 lines, 1.592, printed as 1.6 H. For 4320 lines, 0.796, printed as 0.8 H. Applied to the older rows it lands within about 0.2 H of every published figure. The driver is the line count, not the horizontal pixel count, so doubling the lines halves the distance.
The standard adds a qualifier most summaries drop. When evaluation involves resolution the lower value should be used, but otherwise "any viewing distance in the range (for 3 840 x 2 160 format: 1.6 to 3.2 picture heights; for 7 680 x 4 320 format: 0.8 to 3.2 picture heights) may be used." So 1.6H is the near edge of a band, not a cliff. A 4K TV at 3.2H is not being used incorrectly; it is showing a picture a 1080p panel could have matched.
Running the numbers on a real screen
TVs are sold by diagonal, so convert first. For a 16:9 panel, diagonal = H x 2.0397, so H = diagonal / 2.0397. Take a 65-inch set: 65 / 2.0397 = 31.87 inches of picture height, then multiply by the ITU factors.
| Screen (16:9 diagonal) | Picture height | 1080p at 3.2H | 4K UHD at 1.6H | 8K UHD at 0.8H |
|---|---|---|---|---|
| 55 in | 26.96 in | 86.3 in / 7.2 ft / 2.19 m | 43.1 in / 3.6 ft / 1.10 m | 21.6 in / 1.8 ft / 0.55 m |
| 65 in | 31.87 in | 102.0 in / 8.5 ft / 2.59 m | 51.0 in / 4.25 ft / 1.30 m | 25.5 in / 2.1 ft / 0.65 m |
| 85 in | 41.67 in | 133.3 in / 11.1 ft / 3.39 m | 66.7 in / 5.6 ft / 1.70 m | 33.3 in / 2.8 ft / 0.85 m |
The ITU does not publish these per-size distances; the derivation is one multiplication.
Read the middle column first: on a 65-inch screen you must be inside about 4.25 feet before 4K adds detail your eye can separate. The 8K column is harsher: 25.5 inches on a 65-inch set, roughly arm's length, and still only 2.8 feet on an 85-inch set.
Where the 40-degree and 30-degree rules came from
THX recommends that the best seat-to-screen distance is one where the view angle approximates 40 degrees, presented at CES 2006, and for consumer use recommends dividing the diagonal by 0.84 for a 1080p display, about 1.2 times the diagonal. Note what THX said it was: the theoretical maximum horizontal view angle based on average human vision, a ceiling on screen width rather than the distance at which pixels stop resolving. thx.com returns HTTP 403 to automated requests, so this is attested through secondary accounts, not confirmed against THX's own wording. The arithmetic holds: dividing by 0.84 puts you at 2.43 picture heights, closer than the ITU's 3.2H, and yields about 40 degrees. The two answer different questions and should never be averaged.
The other is "SMPTE 30," the claim that a display should occupy a 30-degree field of view. The reference sources that repeat it also record that "there seems to be no direct recommendation from SMPTE on the issue." The reason 30 degrees keeps reappearing is that it is very close to the ITU's own 31 degrees at 3.2H for 1080p: the quoted 1.6264 times the diagonal works out to 3.317H and, run through the geometry, to 30.00 degrees. Several people ran the same 1-arcminute arithmetic against a 1080p pixel grid and landed in the same place.
The distance you should sit is not the distance you will sit
BT.500-15 documents a separate concept, the preferred viewing distance, "based upon viewers' preferences which have been determined empirically." Its Figure 1-1 plots eleven datasets labelled from Ohtani 1970 to Kusakabe 2012, preferred distance in picture heights against screen size in inches. Most series start between 6H and 8.5H on screens under 20 inches and fall to roughly 2.5H to 4H by 120 to 160 inches, so the preferred distance shrinks as screens grow. The spread across studies runs from about 1H to nearly 12H, a warning against treating any single preferred-distance number as settled.
The design viewing distance describes where pixel detail becomes resolvable. The preferred viewing distance describes where humans actually put the couch. The gap between them is most of the reason 4K marketing works.
The honest position on 8K
The 96-degree figure in Table 1-1 is not a living-room number. NHK found that "the sensation of presence tends to level off at a visual angle of around 80 - 100 arc-degrees," treated that as the maximum angle Super Hi-Vision should provide, and then specified 40 to 50 cycles per degree rather than the 30 that plain 20/20 acuity implies. The prototype they described was 7680 x 4320, 16:9, 60 frames per second, 10 bits per pixel, Rec. 709 colorimetry. 8K was engineered to fill most of your vision, by screen size or by seating distance.
The market has followed the geometry. Checked on 2026-07-27, Samsung's dedicated 8K TV category page returned "Sorry, no results were found. Choose other filter options to find a better match for you." with no models listed, and LG's US TV page carries the line "LG's 8K TVs deliver over 33 million pixels (7680 x 4320) for an unmatched cinematic experience, with clarity and smoothness beyond 4K TVs" without naming a purchasable model, size or price. That is a first-hand observation of two vendor pages on one date, not a discontinuation announcement; no such announcement was found, and nothing here establishes what Sony or any other brand sells.
The content side is thinner still. As of the end of 2024 there were no standalone Blu-ray players certified as 8K capable and no home video releases in 8K on physical media by any major studio. NHK launched BS8K in Japan on 1 December 2018, and China's CCTV-8K ran test broadcasts from 2021 before an official launch in 2022.
What to do with your own room
Measure from your eyes to the screen face, not from the wall or the back of the sofa. Then work backwards: divide that distance by 3.2 to get the picture height that 1080p would already saturate, multiply by 2.0397, and you have the diagonal below which resolution genuinely stops being the limiting factor. Buy the biggest screen the room allows, buy 4K because that is what is manufactured at every size and price, and do not pay a premium for pixel count you cannot resolve. For a picture that visibly improves at 9 feet, spend on contrast, peak brightness and panel type, none of which the arithmetic above governs.
Two traps remain. A calculator's output is not a single correct answer, so ask which goal it optimises: the acuity answer and the maximum-angle answer differ by about a third. And these distances assume 20/20 acuity. Sharper eyesight pushes them back somewhat, by no amount quoted here, and by nothing close to what would turn 25 inches into a sofa distance.
Sources
- https://www.itu.int/dms_pubrec/itu-r/rec/bt/R-REC-BT.500-15-202305-I!!PDF-E.pdf
- https://www.itu.int/rec/R-REC-BT.500/en
- https://tech.ebu.ch/docs/techreview/trev_2008-Q2_nhk-ultra-hd.pdf
- https://en.wikipedia.org/wiki/Visual_acuity
- https://en.wikipedia.org/wiki/Optimum_HDTV_viewing_distance
- https://en.wikipedia.org/wiki/8K_resolution
- https://www.samsung.com/us/televisions-home-theater/tvs/8k-tvs/
- https://www.lg.com/us/tvs