Smallest unit of a digital image — each pixel carries color and brightness data. Resolution measured by pixel count per side.
On set and in editing, we work with them daily without much thought. But as soon as it comes to quality, storage size, or compositing, you face the problem: every digital image is composed of individual, tiny color dots. These dots are called pixels. Each pixel stores color and brightness information. The higher the pixel density (measured in width × height), the sharper and more detailed your image appears – at the same time, the amount of data grows exponentially.
In practice, this means: a 4K recording (4096 × 2160 pixels) quickly produces 500 GB per hour when shooting with a 10-bit color space and ProRes codec. An 8K project – we're talking about 8192 × 4320 pixels – quickly overwhelms your storage and your editing suite. Therefore, you have to decide early on: do I really need this resolution, or is 2K (2048 × 1080) sufficient for DCP distribution? In VFX compositing, pixel density becomes the direct working size. You render your 3D layers at the same resolution as your original footage, otherwise you'll have matte problems and edge artifacts. A single pixel can carry various information – not just RGB color, but also an alpha channel (transparency), a depth channel, or motion vectors for optical flow.
The practical trick: with high-resolution projects, you tend to shoot in oversampling. A 6K RAW recording (RED, Alexa LF) gives you headroom for reframing or stabilization in post. You'll cut your project down to 4K later, but you'll still have reserves. Conversely: pixelation artifacts occur if your footage is already too compressed during shooting (too aggressive codec setting) or if you scale and resample too often. Every interpolation costs sharpness.
In the context of VFX, pixel logic is also semantically expanded: pixel pushing in compositing means painstakingly correcting each individual pixel manually – time-consuming, but sometimes unavoidable for fine masks or rotoscoping. The term also describes the precision of motion tracking or stabilization at a sub-pixel level. The industry now works with 32-bit and float pixels, not just 8-bit or 16-bit, to have maximum flexibility in color correction and grading.