The four o'clock test

There is a moment in every northern winter — roughly four in the afternoon, late November — when the sun is low, the sky is flat, and there is perhaps forty minutes of usable daylight left. It is the least flattering light a room will ever receive.

We design to that moment. Everything else takes care of itself.

Why noon is a bad brief

Daylight models are usually run at the equinox, at midday, under a clear sky. That combination happens for a handful of hours a year. It produces a number that satisfies a planning authority and tells you almost nothing about whether the building is pleasant to be inside.

A room that works at noon in June works because the sun is doing all the labour. The window could be in the wrong place, the ceiling could be the wrong colour, and you would not notice.

What the test actually involves

Three things, none of which require software.

One: stand in the room and read something. If you have moved toward the window without deciding to, the ambient level is too low and the distribution is too steep. A room should not have a reading corner by accident.

Two: look at the ceiling. In low sun, the ceiling is the light source, because most of what reaches the back of a room has bounced off it. If the ceiling is grey, or busy, or too low at the window end, the room dies from the back forward. We keep ceiling reflectance above 0.8 and we keep the first metre inboard of the window free of anything.

Three: look at the deepest corner. Measure the contrast between the brightest surface and that corner. Above roughly 20:1 the eye keeps re-adapting and the room feels tiring. Below about 3:1 it feels flat and institutional. Somewhere near 10:1 a room feels calm and legible.

What it changes

Once you design for the worst hour, the fixes are almost always in the same three places: the head height of the window rather than its area; the reflectance of the ceiling and the wall opposite the glass; and whether there is a second, smaller opening on a different orientation.

That last one does most of the work. A single large window gives you a bright zone and a dark zone. Two modest openings on different walls give you a room. It costs less glass, not more.

The part nobody likes

The test is unkind to the things that photograph well — the dark feature wall, the deep plan, the double-height space with glazing on only one side. We have lost arguments about all three, and been proved right about all three roughly two winters later.