Short answer
Because the two walls are lit by different light. In the northern hemisphere a south-facing wall takes the sun's warm, very bright beam for most of the day, while a north-facing one is lit by blue skylight alone from the September equinox to the March one; south of the equator the roles swap. A Swedish study of about 3,600 façade observations found walls in shade or under cloud read blacker, less white and bluer than the same paint in sun — and found a larger, constant gap between any sample and the finished house, which always looked lighter.
A wall outdoors is lit by the sun's beam, by the sky, or by both. The beam is strong and comparatively warm; the clear sky is weaker and blue, because it is sunlight scattered by air. Measurements of daylight in Granada, taken in all weathers over two years, ran from under 4,000 K with the sun low to above 30,000 K in shade lit only by clear sky, with 5,700 K the most typical value. Brightness differs as much as colour: the façade study described below quotes about 60,000 lux from sun and sky together in clear weather against about 10,000 lux from a clear sky with the sun excluded. A wall in sun and a wall in shade are therefore under lights several thousand kelvin and a factor of several in intensity apart, at the same moment, on the same building.
Eyes adapt to the dominant light, which hides much of this. The difference shows most where a sunlit and a shaded elevation meet at a corner and can be compared directly.
This page is framed for the northern hemisphere outside the tropics, where the sun crosses the southern sky. There a south-facing wall can receive direct sun on every clear day of the year, east walls take it in the morning and west walls in the afternoon. A north-facing wall receives none between the September and March equinoxes, because the sun rises and sets south of the east–west line. In summer it rises north of east and sets north of west, so in London a north wall can catch up to 7.1 hours of sun in June — but only early and late, at a glancing angle. South of the equator everything reverses: the sun crosses the northern sky, the north wall is the sunny one and the south wall the shaded one, as the Sydney columns of the table show. Between the tropics the sun passes to either side during the year, and both walls take their turn.
Sun angle matters as well as duration. In London the noon sun stands about 15° above the horizon in December and about 62° in June, so a south wall is struck almost square-on in winter and at a steep glance in summer, when eaves and balconies shade it too.
The fullest evidence is Karin Fridell Anter's doctoral thesis at KTH in Stockholm, published in 2000. She and trained colleagues, with groups of school students, made about 3,600 observations of 125 painted timber and rendered façades around Uppsala, Sweden, usually from about 50 metres and roughly square to the wall, describing what they saw in Natural Colour System terms by 6 different methods. Each wall's 'inherent' colour was fixed by holding NCS samples against it. Sorting the observations by light gave a clear split: direct sun on one side, and on the other diffuse light, whether from an overcast sky or from blue sky on a shaded wall. In diffuse light the same paint was seen with more blackness and less whiteness, and tended to look less yellow or more blue. That is the north-wall effect, measured.
The study's strongest result concerns every elevation, not just shaded ones. Perceived colour had less blackness than the inherent colour on every façade observed, and on average slightly more chromaticness: the house looked lighter and a little more colourful than the sample held against it. Pale colours gained whiteness; deep ones gained chromaticness. Hue drifted too, yellows towards green, reds towards blue, and neutral greys picked up a faint blue. These gaps were larger than any difference between sun and shade, near and far, or summer and winter surroundings. Fridell Anter offers explanations only as speculation: outdoor light is many times stronger than the roughly 1,000 lux of the cabinet in which sample colours are defined, and samples are judged against white paper. Size alone is not shown to be the cause.
No systematic change was found between about 4 and 50 metres. Paler, cooler colours appeared only beyond about 600 metres, and on too few observations to rely on.
Orientation sets the light on a flat wall; the building's own shape then takes some of it away. Deep window reveals, projecting eaves, balconies and returns cast shadows that move through the day, and those shadowed patches are lit like a north wall whatever the elevation. Rough render and vertical boarding shade themselves at a smaller scale, so raking sun shows a wall darker and more textured than frontal sun does. A colour scheme that depends on a subtle difference between two paints can lose it altogether where one of them sits in the building's own shadow, and a strong contrast between body and trim can look harsher on the sunlit side than the designer intended.
The measured pattern supports a practical habit: choose the sample a little darker and greyer than the colour wanted on the house, then test it where it will go. A board of about a metre square, painted in the real product and finish, should be stood against each elevation in turn and looked at from the street in sun, in shade and under cloud. Where one colour must serve all four sides, judge it on the shaded elevation, where it will look bluest and darkest, and on the sunniest, where it will look lightest and warmest, and accept the pair. Very light colours in full sun can be uncomfortably bright for neighbours and passers-by; that is a matter of reflectance, taken up on the glare page.
| Elevation | London, 51.5° N: June solstice | London, 51.5° N: equinoxes | London, 51.5° N: December solstice | Sydney, 33.9° S: June solstice | Sydney, 33.9° S: equinoxes | Sydney, 33.9° S: December solstice |
|---|---|---|---|---|---|---|
| North-facing | 7.1 h | None | None | 9.7 h | 12.0 h | 6.6 h |
| East-facing | 8.2 h | 6.0 h | 3.8 h | 4.9 h | 6.0 h | 7.1 h |
| South-facing | 9.3 h | 12.0 h | 7.6 h | None | None | 7.6 h |
| West-facing | 8.2 h | 6.0 h | 3.8 h | 4.9 h | 6.0 h | 7.1 h |
| Light situation | What lights the wall | When it occurs | Reported effect |
|---|---|---|---|
| Direct sun, square on | The sun's beam plus light from the whole sky | The sun stands in front of the wall: around midday on a façade facing the equator, morning on an east one, late afternoon on a west one. | Least perceived blackness and most whiteness of the four situations; the yellowest reading of the same wall. |
| Direct sun, raking | A glancing beam plus skylight; texture throws its own small shadows | The sun is nearly in line with the wall, shortly after it comes round the corner or before it leaves. | Grouped with square-on sun in the study: both gave lighter, less blackish colours than diffuse light. |
| Cloudy, hazy or overcast sky | Diffuse light from the whole sky dome, with no beam and no hard shadows | Any orientation under cloud; all four elevations receive much the same light. | More perceived blackness and less whiteness than in sun, with a tendency towards less yellowness or more blueness. |
| Shade under a clear sky | Blue skylight only; the sun is behind the building or blocked | The elevation facing away from the equator for most of a clear day, and any elevation while the sun is behind it. | Behaved like overcast light in the study: more blackness, less whiteness, and a bluer or less yellow reading than the same wall in sun. |
| Attribute | What was found | How firm | What it means for a sample |
|---|---|---|---|
| Blackness | Perceived colour had less blackness than the inherent colour on every façade observed. | Consistent across every façade observed | The finished house reads lighter than the sample held against it. |
| Whiteness, for pale inherent colours | Inherent colours with much whiteness gave perceived colours with still more whiteness and unchanged chromaticness. | Recurring tendency | Off-whites and pale tints drift towards plain white outdoors. |
| Chromaticness, for inherent colours with little whiteness | Where the inherent colour had little whiteness, chromaticness rose most and whiteness stayed about the same. | Recurring tendency | Deep, strong samples come out more colourful on the wall, not merely lighter. |
| Hue | Perceived hue moved away from a stable point near Y50R (orange) and towards a point between blue and R70B: yellower and greenish colours anticlockwise round the NCS circle, through green; redder ones clockwise, through red. | Recurring tendency | Yellows turn slightly greenish, reds slightly bluish; blues barely move. |
| Neutral greys | Inherently neutral greys tended to gain a little chromaticness with a hue near blue. | Recurring tendency | A grey chosen as neutral can look faintly blue on the building. |
| Viewing distance, 4 m to 50 m | No pervading difference between about 4 m and 50 m, though for some colour areas the 4 m view sat closer to the inherent colour. | Uncertain: few observations | Stepping back across a street changes little compared with the sample-to-wall jump. |
| Viewing distance, beyond about 600 m | Whiteness appeared to increase for most colour areas, chromaticness to fall for blue and green houses, and pink, red and green houses to shift from yellow towards blue. | Uncertain: few observations | Across a valley or a harbour, colours pale and cool; few observations support the detail. |
| Seasonal surroundings | No systematic difference between winter brown, spring, summer and autumn green, and snow-covered surroundings. | Uncertain: few observations | Greenery or snow around a house mattered less than the method could detect. |
Why: One elevation is in sun and the other is lit only by blue sky, which reads darker and bluer.
Fix: Judge the colour on both elevations before committing; if the gap is unwelcome, a warmer, slightly lighter version on the shaded side narrows it.
Why: Neutral greys tended to pick up a bluish cast outdoors in the façade study, and skylight on a shaded wall adds to it.
Fix: Test a grey with a little warmth in it against the neutral one, on the shaded elevation.
Why: Façades were consistently perceived with less blackness, and often more chromaticness, than their samples.
Fix: Step the sample darker and greyer, and decide from a large painted board viewed from the street.
Why: Diffuse light adds blackness to both and flattens a small lightness difference; self-shading under eaves does the same.
Fix: Increase the lightness step between the two, and check it in shade.
Each statement is labelled by kind — established fact, a standard’s requirement, observed market data, a convention, or Colourwise’s own interpretation or analysis — with the strength of the evidence behind it.
FactStrong evidence· Uppsala region, Sweden
Fridell Anter's study of perceived façade colour rests on about 3,600 observations of 125 painted timber and rendered façades around Uppsala, Sweden, mostly from about 50 metres, made by trained colour researchers and school students and compared with each wall's inherent colour, determined by holding Natural Colour System samples against it.
Measured: 125 façades; trained observers and school students. Method: Six visual determination methods within the Natural Colour System
FactModerate evidence· Uppsala region, Sweden
In that study the perceived colour of a façade always had less blackness than its inherent colour and on average slightly greater chromaticness, and the variation between light situations, viewing distances and seasons was smaller than this difference.
Caveat: One study, one region and building tradition, visual assessment; demonstrated for full daylight and distances up to a few hundred metres. The author treats her explanations as speculation.
FactModerate evidence· Uppsala region, Sweden
Façades seen in diffuse light, from an overcast sky or as shaded walls under blue sky, were perceived with more blackness and less whiteness than in direct sunlight, with a tendency towards less yellowness or more blueness.
Caveat: Observations were grouped by light situation after the fact, not made as a controlled comparison of north and south elevations.
FactModerate evidence· Granada, Spain
Measured daylight in Granada ranged in correlated colour temperature from under 4,000 K to over 30,000 K, with 5,700 K the most typical value; the highest values belong to shade lit only by clear sky.
Caveat: One site and climate; other latitudes and skies differ in detail.
Source: Color and spectral analysis of daylight in southern Europe
Colourwise analysisModerate evidence
At London's latitude an unobstructed north-facing wall can receive no direct sun at the equinoxes or in December and up to about 7.1 hours on the June solstice, while a south-facing wall can receive about 7.6 hours in December and 9.3 in June; at Sydney's latitude the north and south walls exchange roles.
Based on: Colourwise calculation from standard solar geometry: for each minute of the day the sun's direction is found from latitude, declination (±23.44° at the solstices, 0° at the equinoxes) and hour angle, and counted when it is above the horizon and in front of the wall. Code and tests are in the architecture-light dataset.
Caveat: Geometry only: no refraction, no obstructions, no cloud. Real walls receive less, never more.
Source: Solar Calculation Details
Colourwise interpretationModerate evidence
Choosing an exterior sample slightly darker and greyer than the colour wanted on the building, and testing it at about a metre square on each elevation, compensates for the measured tendency of façades to look lighter and more chromatic than their samples.
Based on: Colourwise's practical reading of the consistent blackness and chromaticness differences in Fridell Anter's thesis; the thesis reports the pattern and says it could guide colour design, but does not prescribe this rule.
Reviewed 1 October 2026. Colourwise summarises its sources in its own words and does not reproduce standards text or proprietary colour data. Spotted an error? Tell us.