Short answer
Because they reflect sunlight or bright lamps specularly — like a mirror — rather than diffusely, sending a concentrated beam towards people's eyes. Colour and light reflectance value describe how much light a surface returns; finish and geometry decide where it goes. The worst cases combine glossy or glazed surfaces with curved or angled façades, as at 20 Fenchurch Street in London in 2013, where a concave glass face focused sunlight onto streets hot enough to damage a parked car.
A surface's light reflectance value tells you what fraction of visible light it returns, averaged over all directions. It says nothing about the direction. A matt white render and a polished white stone can have similar LRVs; the render scatters light in all directions, so it looks evenly bright from anywhere, while the polished stone sends much of its reflection in one direction, producing a bright highlight for some viewers and a darker appearance for others. Glass, polished metal and gloss paint are strongly specular. Glare problems from buildings are therefore mostly about finish and geometry, with colour controlling only how much energy there is to redirect.
A flat mirror reflects the sun's image; a concave one concentrates it. At 20 Fenchurch Street, the concave south face acted as a curved mirror for a period of up to about two hours a day when the sun was in the right position during summer 2013, focusing a beam onto the streets to the south. Media reported spot readings above 90 °C, and parts of a parked car's bodywork were damaged; the developer paid for the repair and, with the City of London's approval, installed temporary screening while longer-term fixes were evaluated. The lasting remedy was external shading on the façade, which intercepts the sun before it reaches the glass. The case shows that glare assessment belongs at design stage, using the sun's path and the façade's shape, not after occupation.
Solar glare is predictable. The sun's position for every hour of the year is known in advance, so reflections from a proposed façade can be modelled before anything is built.
Glass is the special case because its reflection is almost entirely mirror-like and because how much it reflects depends on angle. The City of London's planning advice on solar glare puts standard low-emissivity double glazing at around 13% specular reflectance with the sun square on, which it notes can already be enough to dazzle, and solar-control glass, coated to keep heat out, typically at 15 to 40%. Those are the catalogue figures. At a glancing angle far more is returned: for clear double glazing the note gives 15% at 45 degrees, 22% at 60 and 49% at 75. The table sets those beside a calculation for a single uncoated pane, which climbs in the same way, so the rise belongs to the physics of a glass surface and no choice of clear glass removes it. A tower that looks transparent face-on becomes a mirror when seen, or lit, along its length.
A glazed façade has little colour of its own from outside in daylight. What is seen is mostly the reflected sky and neighbours, over whatever tint the glass and its coating add; a façade facing away from the sun may reflect nothing but sky.
Geometry decides who receives the reflection. A vertical glass wall dazzles mainly when the sun is low. Glass that leans back from the vertical can turn high summer sun down along the street, which the City's note treats as the more worrying case: the high sun is brighter, less likely to be blocked by other buildings, and drivers do not expect it. Glass that leans forward is unlikely to cause trouble. The note separates disability glare, which hides other things from view, from discomfort glare, and asks for assessment at planning stage from the places where it matters — junctions, signals, crossings and railway approaches — treating reflections more than about 25 degrees from a driver's line of sight as usually negligible. For concave reflective shapes a companion note sets ceilings on concentrated sunlight, listed in the table: ordinary summer sunshine in London never exceeds 1 kW/m², and no surface should receive ten times that.
Inside buildings the concern is usually discomfort glare from bright luminaires, windows and reflections in screens and shiny surfaces. The CIE's Unified Glare Rating, set out in CIE 117-1995, estimates discomfort from the luminance of light sources in view, their size and position, and the background luminance. Surface colour enters through the background: dark walls and ceilings make the same luminaire more glaring than light ones, because the contrast is higher. That is one reason lighting designers ask for reasonably light room surfaces, and why the same fitting can be comfortable in one room and harsh in another.
Very high-reflectance outdoor surfaces — white paving, bright render, reflective roofs — can make streets and courtyards uncomfortably bright in strong sun and reflect heat onto neighbouring walls and people, even if they keep their own surface cooler. The trade-off between cooling a surface and glare or heat on those around it is real and site-specific. Designers manage it with matt finishes, texture that breaks up reflections, vegetation and shading, and by modelling where reflections will fall. None of those tools changes a colour's hue; they change how light leaves the surface.
| Angle between sun and the perpendicular to the glass | Clear double glazing (published) | Single uncoated pane (calculated) | What it means |
|---|---|---|---|
| 0° (sun square on to the glass) | Not given for clear glass | 8% | The figure manufacturers quote |
| 45° | 15% | 10% | Already above the catalogue figure |
| 60° | 22% | 17% | Already above the catalogue figure |
| 75° | 49% | 41% | Glancing sun: about half the beam is thrown back |
| Glazing | Specular visible reflectance | Note |
|---|---|---|
| Standard low-emissivity double glazing | About 13% | Described as around 13%, which the note says can be enough to cause glare. |
| Solar-control glass | 15–40% | Coated to reject heat; typically more reflective than plain low-emissivity units. |
| Irradiance | Kind of figure | Meaning |
|---|---|---|
| 0.5 kW/m² | Ordinary sunshine | Common on a sunny summer day in London. |
| 1 kW/m² | Ordinary sunshine | Never exceeded by unconcentrated sunshine in London. |
| 1.5 kW/m² | Recommended ceiling | Street-level areas where people are present: reflected irradiance above this, and preferably above 1 kW/m², to be minimised. |
| 2.5 kW/m² | Recommended ceiling | Places where people are likely to be, including windows of occupied rooms: not to be exceeded for more than 30 seconds, the level at which skin can burn in that time. |
| 10 kW/m² | Recommended ceiling | No area, even at roof level, to receive this much: above it timber, plastics, fabric and paper can catch fire. |
Why: Glass leaning back from the vertical reflects high sun downwards along the ground, at an angle where it reflects strongly.
Fix: Keep reflective glazing vertical or leaning forward where it faces roads, or shade it externally; check it in the reflection study.
Why: Specular reflection of low sun from angled glazing towards a road.
Fix: Model solar reflections at design stage; adjust glazing angle, use lower-reflectance glass or add shading.
Why: High-reflectance paving and walls return much of the sunlight into the space.
Fix: Use mid-reflectance, textured or planted surfaces, and shade where people sit.
Why: Darker walls raise the contrast between luminaires and their background.
Fix: Keep ceilings and upper walls light, or choose luminaires with lower luminance in view.
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.
FactModerate evidence
During construction of 20 Fenchurch Street in 2013, its concave south façade focused reflected sunlight onto streets to the south for about two hours a day, with a media-reported spot reading of 91.3 °C and damage to a parked car.
Caveat: Temperatures are spot readings reported by media rather than controlled measurements.
Source: London skyscraper 'melts' car (and follow-up reporting, 2–3 September 2013); 20 Fenchurch Street — solar glare problem (Wikipedia)
FactStrong evidence
CIE 117-1995 defines the Unified Glare Rating, a practical system for evaluating discomfort glare in interior lighting.
FactStrong evidence
Light reflectance value measures the total visible light a surface reflects in all directions, so it does not distinguish a matt surface from a glossy one of similar reflectance.
FactStrong evidence
The City of London's 2017 advice on solar glare gives the specular reflectance of standard low-emissivity double glazing as around 13% at normal incidence and of solar-control glass as typically 15–40%, and gives clear double glazing as reflecting 15% with the sun at 45° to the glazing, 22% at 60° and 49% at 75°.
Caveat: Representative figures in planning advice, not measurements of a particular product; coatings change them.
StandardStrong evidence· City of London
The City of London's 2017 advice on solar convergence recommends that no area receive reflected solar irradiance of 10 kW/m² or more, that places where people are likely to be present not receive more than 2.5 kW/m² for more than 30 seconds, and that street-level areas above 1.5 kW/m² be minimised; it states that only a concave reflective face can concentrate the sun.
Caveat: Planning advice from one authority, July 2017 edition; not a national standard and not a compliance assessment.
FactStrong evidence
External shading was used to prevent solar convergence at 20 Fenchurch Street, and the City of London's advice names external louvres, fins or blinds as a remedy once a building is complete.
Colourwise analysisModerate evidence
A single uncoated pane of ordinary glass reflects about 8% of unpolarised light arriving square on, about 17% at 60° and about 41% at 75°, so the steep rise in published double-glazing figures follows from the optics of a glass surface.
Based on: Colourwise calculation with the Fresnel equations for a refractive index of 1.52, averaging the two polarisations and summing the reflections between the pane's two surfaces; no absorption or coating. Code and tests are in the architecture-light dataset.
Caveat: An idealised single pane; real units have two or three panes and coatings, and reflect more.
Source: HyperPhysics (optics, atmospheric optics and vision pages); Solar Glare: guidelines and best practice for assessing solar glare in the City of London (Planning Advice Note, July 2017)
Colourwise interpretationModerate evidence
Façade glare risk depends more on surface finish and geometry than on colour, which mainly sets how much light is available to be redirected.
Based on: Follows from the difference between specular and diffuse reflection and from the Fenchurch Street case, where geometry concentrated reflected sunlight.
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.