U-value and g-value, and why they pull in opposite directions
Every sealed unit carries two performance figures, they describe two different things, and on overhead glazing they fight each other. Understanding the mechanism, rather than the numbers, is what lets you decide which way to trade.
Two questions about the same pane
The U-value answers: how fast does heat escape through this unit? The g-value answers: how much of the sun’s energy passes through it?
One is about heat you have paid for leaving. The other is about heat arriving whether you want it or not. They are measured differently, they are improved by different things, and the things that improve them get in each other’s way.
What the U-value is
Heat transfer per square metre per degree of temperature difference, in watts. Lower is better, and the figure covers all three ways heat moves through a window.
- Conduction through the glass and the frame.
- Convection within the cavity, as gas circulates between the warm inner pane and the cold outer one.
- Radiation across the cavity, from the warm pane to the cold one, as infrared.
Radiation is much the largest of the three in a plain double glazed unit, which is why the single most effective thing you can do to a window is stop it.
What a low emissivity coating does
Emissivity is how readily a surface radiates heat. Ordinary glass is a good radiator, which is unhelpful.
A low emissivity coating is a microscopically thin metallic layer on one of the cavity faces of the glass. It is transparent to visible light and it reflects long wave infrared, so the warm inner pane radiates far less of your heat across the cavity to the cold outer one.
It is invisible to look through, it costs very little, and it transforms the U-value. Every current unit has one, and it is the reason a modern rooflight loses a fraction of the heat a 1980s one did.
What the g-value is
The proportion of solar energy arriving at the glass that ends up in the room, expressed from 0 to 1. Sometimes called solar factor or, in the older literature, shading coefficient on a slightly different scale.
A clear double glazed unit with a low emissivity coating sits somewhere around 0.6 to 0.7, meaning most of the solar energy hitting it comes through. On a north facing rooflight in February that is free heat. On a south facing one in July it is the problem.

Where the conflict comes from
| Low emissivity coating | Solar control coating | |
|---|---|---|
| Blocks | Long wave infrared, from inside | Short wave solar energy, from outside |
| Improves | U-value | g-value |
| Sits on | A cavity face of the inner pane | A face of the outer pane |
| Costs you | Almost nothing | Light transmission |
| Wanted | Always | Where solar gain is a problem |
The two coatings are not in conflict with each other. The conflict is between blocking solar energy and letting daylight through, because they arrive together in the same beam.
The real trade is light, not heat
This is the point that matters and it is usually stated wrongly.
You can have a very low U-value and a very low g-value at the same time. What you cannot have is a very low g-value and a very high light transmission, because the coatings that reject solar energy reject part of the visible spectrum along with it.
So the choice is not between warm and cool. It is between how much solar heat you reject and how bright the room is, and pushing the g-value down far enough makes the room read grey.
Selectivity, and the figure worth asking for
Because the trade is light against solar rejection, the useful measure of a solar control glass is how much light it lets through for a given g-value.
Glass makers express that as a ratio of light transmission to g-value, and a higher ratio means a more selective coating: more daylight for the same solar rejection. It is the number that separates a good solar control unit from a merely dark one.
You will not see it on most quotations, and it is worth asking for where a room needs both brightness and control.
Why overhead glazing makes it worse
A vertical window facing south takes the sun at a low angle in winter and a glancing angle at midsummer, because the sun is high overhead in June.
A rooflight does the opposite. In midsummer the sun is close to normal to the glass for hours either side of noon, which is the angle at which the most energy passes through. The same unit that behaves well in a wall behaves badly in a roof.
Which is why solar control matters more on a rooflight than on a window of the same area and the same orientation.
Angle of incidence
The g-value published for a unit is measured at normal incidence, meaning the energy arriving perpendicular to the glass.
At shallower angles more energy is reflected and less passes through, so real performance varies through the day and through the year. On a vertical window that works in your favour in summer; on a flat rooflight it works against you, because summer noon is close to normal.
On a pitched unit like a lantern, the pitch decides how much of that reflection you get, which is why a steeper slope behaves more seasonally.

The cavity, and what fills it
Cavity width and gas fill both affect the U-value, and both have limits.
Too narrow and the panes conduct through the gas too readily. Too wide and the gas convects, circulating heat from the warm pane to the cold one, which is why a very wide cavity performs worse than an optimal one rather than better.
Argon is denser and less conductive than air, and it convects less readily, so it improves the figure for no visible change. It is standard on any current unit, and how long it stays in there has its own page.
The edge, and the frame
A published centre-pane figure describes the middle of the glass, where the coating and the gas are doing their work undisturbed.
The edge of the unit is a different environment, because the spacer bridges the two panes, and the frame is different again. Both perform worse than the middle, which is why the figure for a whole unit is always worse than the figure for its glass.
Which figure you are being quoted matters a great deal, and it has its own page.
What the regulations ask for
Building Regulations set a limiting standard for the U-value of a replacement rooflight, and any current product from a reputable supplier clears it comfortably.
The g-value is not controlled in the same way for a replacement. Where it is controlled is in the overheating assessment for new dwellings under Part O, and through the glazed area rules for extensions, both of which are covered on the overheating page.
So in practice the U-value is a compliance question with an easy answer, and the g-value is a comfort question with a real decision in it.
Which way to trade, by orientation
The short version, before the detailed pages.
North facing: take the light. A low g-value here rejects daylight you wanted and solar gain that was never a problem, and it is the wrong answer for reasons the north light page sets out.
South and west facing: control the gain. West is frequently the worse of the two, because the peak arrives in the late afternoon onto a room that has been warming all day, at exactly the hour a kitchen is busiest.
East facing: usually straightforward, because the peak is in the morning when the room is coolest.
The room decides more than the compass
Two rooms with identical rooflights behave differently, and the difference is what happens underneath.
A kitchen generates its own heat from cooking and appliances. A room with a hard floor, hard worktops and no soft furnishing stores and re-radiates heat differently from a carpeted one. A single storey extension with one set of doors has no cross ventilation, so nothing removes what does get in.
Glass specification deals with energy arriving. It does nothing about energy generated, and nothing about removing what is already there.

Ventilation is the other half
Solar control reduces the heat coming in. It does not remove heat already in the room, and a kitchen makes plenty of its own.
An opening at the highest point of a room removes warm air far more effectively than an opening at floor level, because that is where warm air collects. Paired with an opening low down on the opposite side, it produces a genuine flow path.
Specifying good glass and no ventilation is solving half the problem twice.
Reading a specification sheet
Manufacturers publish the same handful of figures and they are worth being able to read at a glance.
The U-value, with its basis. The g-value. Light transmission, usually a percentage, which is the figure that tells you how bright the room will be. Sometimes a selectivity ratio. And the glass build-up itself, written as the thicknesses and the cavity, which tells you what is actually in the unit.
A build-up written out will tell you whether the inner pane is laminated, which it should be on any overhead glazing, and whether the two panes differ in thickness, which improves the acoustic performance. Both are worth checking rather than assuming.
What the numbers do not tell you
Three things matter to how a room feels and appear in no figure on the sheet.
The colour of the light. Some solar control coatings are close to neutral and some give a faint tint that reads as slightly cool or slightly green from inside. On a large overhead area that is noticeable, and it varies between products with similar numbers.
The view. A heavily coated unit changes how the sky looks through it, which on a rooflight is most of what you see. And the reflection from outside: a strongly coated outer pane reads as more mirrored from the garden, which matters on a unit that is visible.
Where those matter, a sample is worth more than a datasheet, and manufacturers will supply one.
Winter, and the thing solar control costs you
A low g-value applies all year, and in December it is rejecting solar gain you would have been glad of.
On a south facing rooflight that free winter heat is a real contribution, and a high performance solar control unit gives some of it up. On a well insulated house with a modest heating demand the loss is small; on a poorly insulated one it is more noticeable.
It is a genuine trade rather than a free improvement, which is why the middle specification suits most rooms rather than the strongest one.
What we take on the glazing
We take the orientation and the real shading from the site rather than from a plan, ask what the room is used for and when, and write the glass specification into the quote so it is fixed before anything is ordered. We say which figure we are quoting, whole unit or centre pane, and we do not push a g-value lower than the room needs, because the cost of that is the daylight you wanted in the first place.
The U-value is a compliance question with an easy answer. The g-value is a decision about how the room feels, and it is made once.
Thinking about this job? We survey across Worthing, West Sussex and the South Downs and give you a fixed price before any work starts, with a 10-year workmanship guarantee.