Sunset or Sunrise for Photography? The Measurable Differences

Most of this argument is preference, and honestly so. Two parts of it are not, and both are computable.

The sunrise-versus-sunset argument is one photography's most durable, and most of it is preference wearing the clothes of an argument. Cleaner air, fewer people, harder to get out of bed — all true, all subjective, all weighted differently by different people.

Two parts of it are not subjective. The two golden hours are not the same length on the same day, and the sun does not rise and set in mirror-image positions. Both are computable, so here they are.

Morning and evening golden hour on the same day, with the bearings the sun rises and sets on.
CityLatitudeMorning goldenEvening goldenSunrise bearingSunset bearing
Longyearbyen78.2°N9h 07m8° N344° NNW
Tromsø69.6°N2h 16m2h 14m55° NE304° NW
Reykjavík64.1°N101 min100 min63° ENE297° WNW
Stockholm59.3°N84 min83 min67° ENE293° WNW
London51.5°N67 min67 min71° ENE288° WNW
New York City40.7°N54 min54 min75° ENE285° WNW
Cairo30.1°N47 min47 min77° ENE283° WNW
Mumbai19.1°N43 min43 min78° ENE282° WNW
Bogotá4.6°N41 min41 min79° E281° W
Nairobi1.3°S41 min41 min79° E281° W
Rio de Janeiro22.9°S44 min44 min79° ENE281° W
Sydney33.9°S50 min50 min78° ENE282° WNW
Wellington41.3°S55 min55 min77° ENE283° WNW
Ushuaia54.8°S74 min74 min73° ENE287° WNW

The lengths are not identical

Morning and evening golden hour are the same ten-degree band of sun altitude, crossed in opposite directions. You would expect them to take the same time, and they nearly do — but not exactly.

The sun's declination changes over the course of a day. By evening it is at a slightly different value than it was at dawn, which tilts the geometry a little and changes how long the crossing takes. Near the equinoxes, when declination is changing fastest, the difference is largest.

At mid-latitudes it amounts to a few minutes. It is not usually enough to decide anything, but it is real, and it means any tool that reports one figure for "golden hour" is giving you an average of two different numbers.

The bearings are genuinely different

This is the difference that matters for planning, and it is much larger.

If the sun sets at 300 degrees it rose at roughly 60 — reflected about the north-south line, not repeated. At high latitude those two points can be most of the sky apart. In London in June the sun rises in the northeast and sets in the northwest; in December it rises in the southeast and sets in the southwest.

For a fixed viewpoint that is decisive. A west-facing coastline is a sunset location and never a sunrise one, whatever the light quality argument says. A valley that opens to the east has the opposite constraint. The sunset direction article has the full picture of how far both points travel across the year.

Choose the time of day your subject faces, then argue about air quality afterwards.

There is one exception worth knowing. At the equinoxes the sun rises due east and sets due west everywhere on Earth, so a subject aligned east-west works at both ends of the day. Those are the two dates a year when the choice is genuinely free.

The parts that are real but not measurable here

Air is usually cleaner in the morning. Overnight cooling settles particulates, the boundary layer is shallow and stable, and a day of traffic and industry has not happened yet. This is a genuine effect and the strongest argument for sunrise. How much it matters varies enormously — in a clean rural location, hardly at all; in a large city in summer, a great deal.

Cloud usually builds through the day. Convective cloud forms as the surface heats, so mornings are more often clear and evenings more often have something in the sky. Whether that is an advantage depends on what you want: a clean gradient or something for the light to land on.

Mist and fog are morning phenomena. Radiation fog forms overnight and burns off after sunrise. If you want a layer of mist in a valley, that is a dawn subject and there is no evening equivalent.

People. Almost every well-known viewpoint is empty at dawn and crowded at sunset. For anything involving a foreground you want to control, this is often the deciding factor.

The practical asymmetry

The two are not equally forgiving, and this is rarely stated.

At sunset you arrive in daylight. You can see your composition, adjust, move to a better spot, and watch the light develop towards its peak. Your eyes adapt gradually along with the scene. If you are twenty minutes late, you still get most of it.

At sunrise you arrive in the dark. You cannot see what you are framing until it is already happening, and the light is at its most interesting within minutes of your being able to see at all. If you are twenty minutes late, you have missed it.

That is why sunset is the better place to learn and sunrise the better place to be once you know the location. Scout in the evening, shoot in the morning, is advice that survives contact with reality.

The order of events reverses

Worth stating because it changes the plan.

Evening: golden hour, then sunset, then blue hour, then dark. The best light comes at the end, and you can keep working as it improves.

Morning: dark, then blue hour, then sunrise, then golden hour. The most delicate light comes first, before the sun is up, when you have least ability to see what you are doing.

So the evening sequence lets you build up to the peak, and the morning sequence demands you be ready before it starts. Same events, opposite pressure.

The answer

There is not one, and anyone who gives you one is describing their own habits.

What is true: pick the end of the day your subject faces, because that constraint is geometric and the rest are preferences. Check the bearing for the actual date rather than assuming east and west. And if the air quality argument matters where you are — a hazy city in summer — it is a real effect and worth the alarm.

Every city page here shows both golden hours, both blue hours, and both bearings for today, so you can settle the geometry before getting into the taste.

Times for your own city Every figure in this article is computed for 733 cities, each with its own page.