Resolution vs. Aliasing: Implications for Motion Capture

There is often a lot of discussion about whether a particular device truly has a given resolution, such as 1080P, 4K or something else, along with how this impacts image quality in general. In this article we'd like to clear up some of that confusion by investigating the trade-offs of any discrete imaging system.


Resolution is clearly a simple and important way of describing the potential for detail in an image. Everything else being equal, higher resolution imagery holds up better when enlarged, and improves creative flexibility by permitting more aggressive cropping and higher-quality visual effects, amongst other benefits. However, resolution is just one aspect of image quality.

One can always have too much of a good thing, and resolution is no different in this regard. If the resolving power is pushed too near the pixel resolution, then unnatural digital artifacts will begin to appear. These happen regardless of sensor type or image-processing algorithm, and apply equally to scanned film:

"Aliasing" may appear as jagged edges, interference patterns ("moiré") and other false detail. Although sometimes tolerable in still images, such artifacts are highly distracting and almost always unacceptable with video. Moiré, in particular, can also confuse video compression algorithms and reduce bandwidth efficiency:

Video with Strong Aliasing
Less Real Detail, 30MB Size
(shown at 200%)
Video with Minimal Aliasing
More Real Detail, 10MB Size
(shown at 200%)

Each uses identical compression settings. Note how aliasing has a flickering quality in motion video. Although the severity of aliasing varies depending on camera, the bandwidth and detail-depriving properties persist.


Unfortunately, the only way to reduce aliasing in post-production is to decrease resolution by blurring or downsizing the image. This rarely solves the problem though, since aliasing patterns are often much coarser than the resolution limit:

Attempted Blurring to Reduce Aliasing
(examples of failed removal circled in red)

Effective removal needs to address the underlying cause of aliasing: when detail finer than the resolution of the sensor tricks it into capturing false information. The optimal solution is therefore to address aliasing during capture, using something called an optical low-pass filter (OLPF). The OLPF effectively smoothes out any detail finer than the pixels so that this is less likely to trick the camera:

Without an OLPF, adjacent pixels see either two dark and one light stripe, or two light stripes and one dark. This creates a new false pattern with larger but lower-contrast stripes. It also happens even without gaps between photosites (a "100% fill factor").

However, no OLPF is perfect; some detail larger than the pixel size also gets softened. This slightly reduces the camera's maximum resolution, presuming that lens sharpness and technique aren't more influential, but much less so than if aliasing had been reduced in post-production. In exchange, the OLPF causes detail to be captured in a way that fades naturally and gradually near the resolution limit similar to how our eyes perceive real-life objects. For these reasons, most high-end still cameras available today use an OLPF. If the "crunchy" aliased look is still desired, one can emulate this using aggressive, low radius sharpening:

The OLPF goals aren't always realized though, depending on the camera. For example, cameras that capture images using pixel skipping are much more susceptible to aliasing because the OLPF is intended for adjacent pixels. This is a common technique with smartphones and DSLR cameras that have a reduced-resolution video feature, in part because these have lower memory and processing capabilities.


Although an aliased image might at first appear deceivingly sharp, much of this is false detail, and a similar look could have been achieved using aggressive sharpening. Once present though, aliasing can be very difficult and destructive to remove, and will end up doing more harm than good in the long run especially with motion capture. All standard digital imagery therefore requires an OLPF. When used properly, this also ends up offsetting any theoretical resolution advantage of non-Bayer camera sensors.

What does this all mean in practice? A camera (or film scan) with 1000 horizontal pixels will never have this high of an artifact-free resolution. However, this shouldn't be taken to mean that the device isn't a "true" 1K camera. When done right, such a camera is capable of producing images which fully utilize and appear natural on a 1K display or projector. To truly improve detail though, it takes both more pixels and minimal aliasing.