Aspect ratio correction

Aspect ratio correction is implemented through a combination of geometric transforms, sampling algorithms and conditional rules based on device metadata. At a basic level, the correction algorithm derives the target aspect ratio from the player’s native resolution or an EDID report, and the source aspect ratio from the asset’s encoded width and height. A decision tree then selects an operation: uniform scaling preserves aspect ratio by scaling both axes equally, which may introduce letterboxing or pillarboxing where the screen is larger in one dimension; non-uniform scaling fits the asset to bounds but distorts geometry; and centred cropping trims the overflow region to fill the screen while losing peripheral content. More advanced pipelines use content-aware techniques that evaluate safe areas and salient regions—text and logos are kept inside a margin, while flexible background imagery can be cropped. For video, chroma subsampling and pixel aspect ratio complicate outcomes: a 4:3 video with a non-square pixel aspect ratio can appear stretched if pixel aspect metadata is ignored, so professional signage pipelines preserve or normalise pixel aspect metadata during transcode. On the player side, interpolation methods matter; bilinear scaling is fast but softens text, whereas bicubic or lanczos produce crisper edges at the cost of CPU or GPU usage. Hardware acceleration on modern media players and SoC-driven devices offloads scaling to dedicated blocks, reducing playback stutter and preserving frame timing. When Fugo.ai serves HTML5 layouts, CSS object-fit and transform properties allow predictable browser-based correction; when serving native video, the platform negotiates with the player to use either pre-rendered variants or on-device scaling according to device profiles and network bandwidth.