Every long exposure app on iPhone does the same thing under the hood, and almost none of them explain it. The phone is not holding a shutter open for ten seconds. It's recording a stream of ordinary frames and merging them into one image. That process is frame stacking, or photo stacking, and understanding it changes how you shoot.

What Frame Stacking Actually Is

A traditional long exposure works by physically leaving the shutter open. Light lands on the sensor continuously for the whole duration, and anything that moves during that window smears across the frame. iPhones can't do that in any meaningful way, and the Camera app doesn't even let you request it.

Stacking sidesteps the problem. The app captures frames continuously, each exposed normally, then combines them pixel by pixel into one output image. Lento does this at 30fps with Metal compute shaders on the GPU, fast enough that you see the composite building live on screen instead of waiting for a render.

Stacking isn't one effect, though. What you get depends entirely on the rule used to combine the pixels, and there are three rules worth knowing.

Mean Blending: The Smooth Look

Mean blending averages every frame together: for each pixel position, it takes the value from all captured frames and computes the average.

Think about a waterfall. In frame one a particular pixel catches a bright droplet, in frame two the gap behind it, in frame three foam. Average a few hundred of those and you get the milky, uniform tone that reads as flowing water. The rocks beside the waterfall show the same value in every frame, so averaging leaves them as sharp as they started. Static things stay crisp, moving things dissolve into their own average.

That dissolve is genuine motion blur on iPhone, recorded at capture time rather than painted on afterward by an editing app. It is the same mechanism whether the moving thing is water, cloud, or a crowd on a sidewalk.

Averaging has a second benefit that's easy to miss: it reduces noise. Sensor noise is random frame to frame, so it averages toward nothing while the scene reinforces itself. Stack enough frames and a dim scene comes out cleaner than any single frame in the stack.

Lighten Blending: Keeping the Brightest Pixel

Lighten blending uses a different rule: for each pixel, keep the brightest value seen across all frames. Nothing gets averaged away.

This is what you want when the subject is a moving light against a dark background. A car drives through the frame with headlights on, and in each frame its lights occupy a slightly different position. Because lighten blending keeps the brightest pixel at every position, all of those positions survive into the final image as a continuous ribbon tracing the car's path.

Averaging would ruin this, since a light occupying a pixel for a few frames out of hundreds barely moves the average and the trail fades to near invisibility. Use lighten for car light trails, light painting, fireworks, and star trails.

Median Blending: Removing What Moves

Median blending takes the middle value at each pixel position rather than the average or the maximum. Small statistical difference, dramatically different result.

Picture a plaza with tourists walking through. Any patch of pavement is empty in most frames and covered by a person in a handful of them. The median for that pixel is the pavement, because that's what the majority of frames show. The people don't make the cut. The output is the scene with the transient objects gone: not blurred, not painted over, just absent. This is the mode behind removing people from photos without touching an editor.

The requirement is that people keep moving. Someone who sits on a bench for the whole capture appears in most frames and survives the median. It's a statistical tool, not magic.

Why Stacking Replaces a Tripod

With a single multi-second exposure, camera shake is fatal. The sensor collects light the entire time, so every tremor smears the whole frame, including the parts meant to stay sharp. That's why tripods exist.

With stacking, each frame is exposed at a duration short enough that hand shake doesn't blur it, and the frames are combined afterward. Small movements between frames don't destroy the image the way movement during a single exposure would, which is why handheld long exposure is realistic on iPhone at all. We cover the technique in our guide to long exposure without a tripod.

A tripod still helps on very long captures. But the gap between "better with a tripod" and "impossible without one" is the gap between shooting and not shooting.

Why Stacking Replaces an ND Filter

On a traditional camera, daytime long exposure has a brightness problem. Leave the shutter open for several seconds in daylight and the sensor floods, producing a blown out white frame. The fix is a neutral density filter, dark glass that cuts the light.

Stacking never accumulates light on the sensor for more than one frame's worth, so the problem doesn't arise. Each frame is exposed correctly for the conditions, and the combining happens mathematically afterward on values that are already correct. A daytime waterfall shot needs no filter. Our piece on ND filter alternatives on iPhone covers the tradeoffs.

Where Stacking Has Limits

Stacking needs frames and frames need time, so a long capture is still a long capture. And the output is only as good as the input: stacking cleans up random noise well, but it can't recover detail that was never recorded. Shooting RAW helps, and Lento exports 16-bit DNG at 3264x2448 with P3 wide color and full EXIF.

Frequently Asked Questions

Is photo stacking the same as HDR?

No. HDR combines frames shot at different exposure levels to extend dynamic range where highlights and shadows are too far apart for one exposure. Frame stacking combines frames shot at the same exposure level to accumulate motion or reduce noise. HDR is about brightness range, stacking is about duration.

How many frames do I need to stack?

It depends on the effect. Smoothing water with mean blending generally needs a few seconds of capture, on the order of a hundred frames at 30fps. Median blending needs enough frames that any given person occupies a spot in only a minority of them, so longer captures in busier scenes. Light trails are less about frame count than about covering the full path of the moving light.

Which iPhones can do real time frame stacking?

Lento runs on iPhone XR and newer with iOS 18, since GPU stacking relies on Metal 2 and the A12 chip. All three blend modes work at 30fps with a live preview. The free version includes Smooth and Light Trail with a 5 second cap; Pro adds Ghost Remove, unlimited duration, and RAW DNG export at $9.99 per year or $29.99 for lifetime access.