You photograph a fast-moving vehicle and a vertical pole seems to lean. You pan your phone quickly and straight buildings appear to wobble. In video, spinning objects can look strangely stretched or slanted even though they look normal to your eyes.
These effects can come from rolling shutter, a way many camera sensors capture an image over a short span of time instead of recording every part of the frame at exactly the same instant. The effect is easiest to notice when the subject moves quickly, the camera moves quickly, or both.
Understanding rolling shutter gives you a useful mental model for several camera distortions that otherwise look like software errors or damaged hardware. It also explains why changing how you shoot can sometimes reduce them.
A frame is not necessarily one instant
It is natural to think of a photograph as a single frozen moment. That is close enough for many everyday situations, but it is not always how the camera sensor collects the image.
An image sensor is a grid of light-sensitive pixels. In a rolling-shutter design, different rows of that grid begin and end their exposure at slightly different times. The capture progresses across the sensor, commonly row by row.
The delay between neighbouring rows can be very small. But a complete frame contains many rows, so the first and last parts of the frame are not necessarily describing precisely the same moment.
For a stationary scene and a steady camera, that timing difference usually causes no obvious problem. A wall is still in the same place when the later rows are captured.
Fast motion changes the situation.
Motion turns a timing difference into a shape difference
Imagine pointing a camera at a vertical post while moving the phone sideways quickly.
Suppose the sensor records rows near one edge of the frame first. By the time it records rows farther across the sensor, the phone has moved slightly. The post therefore appears at one horizontal position in the earlier rows and another position in the later rows.
When those rows are assembled into one image, the post can look slanted even though it was vertical in the real scene.
The same principle applies when the camera is steady but the subject moves. A fast vehicle can change position while different rows are being captured, so its shape may look skewed. A rapidly rotating object can produce more complicated distortions because different parts of it are moving in different directions as the sensor progresses through the frame.
The important point is that the camera is not necessarily bending the object after capture. Different parts of the image can represent slightly different moments.
Rolling shutter is different from ordinary motion blur
Rolling-shutter distortion and motion blur can appear in the same image, but they are not the same problem.
Motion blur happens when an object or camera changes position during the exposure of a pixel. Instead of recording a sharply defined position, the camera collects light from a range of positions, producing a streak or soft edge.
Rolling-shutter distortion comes from the timing difference between different parts of the sensor. Individual parts of the subject can look reasonably sharp while the overall shape is skewed because those parts were not captured at the same time.
For example, a quickly panned photograph might show a pole that is both blurred and leaning. The blur reflects movement during exposure. The lean can reflect the row-to-row timing of the sensor.
A faster exposure can reduce motion blur, but it does not automatically eliminate rolling-shutter distortion. The sensor’s full-frame readout timing also matters.
Why video can make the effect easier to notice
Rolling shutter can affect both still photos and video when the relevant capture mode uses rolling readout. In video, however, camera movement repeats across many consecutive frames, which can make the distortion particularly noticeable.
If you move a phone rapidly from side to side, vertical lines may appear to lean in one direction and then the other. With some combinations of motion and readout timing, the scene can seem to wobble.
This is sometimes described as a “jello” effect because rigid objects appear temporarily flexible. The name describes the appearance, not a separate camera mechanism.
How strong the effect looks depends on the sensor, capture mode, readout speed, movement, stabilization, and other processing. Two phones, or even two modes on the same phone, can therefore behave differently.
Readout speed matters
A rolling shutter is not automatically a slow shutter.
The word shutter can be confusing here because two different time intervals matter. One is the exposure time for a particular part of the sensor. The other is the time separation between the capture of different rows across the frame.
A camera can use a short exposure that freezes local motion quite well while still taking some additional time to progress through the full sensor. Fast sensor readout reduces the time difference between the earliest and latest parts of the frame, which generally reduces rolling-shutter distortion for the same movement.
This is why the effect is not determined by frame rate alone. A video mode that records many frames per second still has its own sensor readout behaviour, and implementations vary between devices and modes.
Not every camera uses the same shutter method
Rolling shutter is common in CMOS image sensors, but it is not the only approach.
A global shutter is designed so that the pixels used for a frame capture the scene at the same time, or effectively share the same exposure interval before their information is read out. This avoids the row-to-row time displacement that creates rolling-shutter geometric distortion.
Global-shutter designs are valuable for situations such as industrial imaging where accurately capturing fast motion can be especially important. They also involve different sensor design requirements and trade-offs, so the presence of a CMOS sensor does not by itself tell you exactly how a particular camera captures every mode.
For everyday phone use, the practical lesson is simpler: do not assume every camera, lens, resolution, or video mode has identical readout behaviour.
Electronic stabilization can change what you see
Phones often process motion as well as capture it. Electronic video stabilization, lens stabilization, frame cropping, and other image processing can affect the final appearance of movement.
These systems may reduce visible camera shake, but stabilization and rolling shutter solve different problems. Stabilization tries to compensate for unwanted movement of the camera. Rolling-shutter distortion arises because different sensor regions are captured at different times.
Processing can sometimes reduce the visible result, but it cannot make every fast-moving scene equivalent to one captured simultaneously across the whole sensor. The outcome depends on the device and recording mode.
This is also why a strange-looking frame is not enough to diagnose a faulty stabilization system. The distortion may come from sensor timing, movement, processing, or a combination of them.
How to reduce rolling-shutter distortion
You usually cannot change the sensor design in your phone, but you can change the conditions that make the effect obvious.
The most useful step is to reduce rapid camera movement. Pan more slowly when possible, especially when the scene contains straight vertical lines such as buildings, poles, or door frames. Holding the phone more steadily also reduces the amount the viewpoint changes while a frame is being captured.
If the subject itself is moving quickly, increasing your distance can sometimes help because the subject then crosses a smaller portion of the frame during the same amount of time. This is a compositional trade-off rather than a guaranteed fix.
Different camera modes can also produce different results because they may use different sensor areas, resolutions, frame rates, processing pipelines, or readout strategies. If distortion is important for a particular shot, compare the modes your device offers rather than assuming the highest resolution or frame rate will always have the least rolling shutter.
For still photography, good light can help the camera use shorter exposures and reduce ordinary motion blur. That can make moving subjects look clearer, although it does not by itself remove the sensor’s row-to-row timing difference.
When distortion may have another cause
Not every bent line or warped object is a rolling-shutter artifact.
Wide-angle lenses can make straight lines near the edge of an image appear curved if lens distortion is not fully corrected. Perspective can make parallel lines seem to converge. Digital panorama modes intentionally combine imagery captured while you move the camera and can create their own stitching errors when subjects move. Strong video processing can introduce additional artifacts.
A useful clue is the relationship with motion. If rigid objects become noticeably skewed mainly when the camera or subject moves quickly, rolling shutter is a plausible explanation. If the same lines remain curved in a completely still scene, lens geometry or correction is more likely to be involved.
The practical takeaway
A phone camera frame may look like one frozen instant, but with rolling shutter its different rows can represent slightly different moments. Normally those differences are too small to notice. Fast movement makes them visible because the scene changes while the sensor progresses through the frame.
That is why a pole can lean during a fast pan, a moving vehicle can look skewed, or video can appear to wobble. Keeping the camera steadier, reducing rapid pans, and comparing capture modes can reduce the effect, but the exact behaviour depends on the sensor and the way each device implements its camera modes.
Once you see rolling shutter as a timing effect across the image, the strange shapes become much less mysterious.