Your Phone's Night Mode Isn't a Photograph. It's a Reconstruction.

Night mode on your phone isn't a photograph. It's a guess. A really good one — but still a guess.

Most people don't realize what's actually happening when they tap the shutter in low light. The camera takes anywhere from 8 to 30 frames over a couple of seconds. Some are short exposures to freeze motion. Some are longer to soak up light. Then an algorithm merges them, denoises aggressively, and fills in detail that was never cleanly captured by any single frame.

The thing nobody talks about is how much of what you see at the end is synthesized rather than recorded.

This works spectacularly well most of the time. It's one of the reasons the gap between smartphone and dedicated camera has closed so dramatically in low light. Images that would have been unusable a decade ago now look clean enough to print.

But it breaks in interesting ways. Moving subjects get smeared or doubled. Fine textures — fabric, skin pores, bark — turn into a slightly painted, watercolor version of themselves. Dark skin can lose tonal gradation because the denoiser is tuned on averages across a broader population of training scenes.

Here's what I keep coming back to. We've stopped calling these "photographs" in the old sense of the word. A night mode image is closer to a composite — a reconstruction based on statistical priors about what a dimly-lit scene probably looks like, rather than a direct recording of what was in front of the lens.

That's not a criticism. It's just a different medium. And most consumers, reasonably, prefer the composite. It looks nicer. It's more shareable. It makes the dinner look good.

The question worth asking is whether the image is accurate, or whether it's plausible. Most people want plausible. A few of us still care about accurate.

What do you actually want from a camera at night — the truth of the moment, or the best-looking version of it? I'd love to hear where you land on this. Drop a comment or reply to the email version.