Loading page…
The craft behind the games. A working record of original model studies, animation questions, and practical engineering experiments across Infinite Odyssey and Oracle’s Return.
Every useful check has a defined scope—and an honest boundary around what it can tell us.
Identify one visible behavior or a specific question.
Describe the asset, state, or bounded check being examined.
Separate a component result from a whole-game claim.
State the next test needed before drawing a wider conclusion.
Original artwork and small, inspectable questions connect visual development to the systems that will carry it.
Engineering notes for Oracle’s Return: richer light and materials, with a smaller-device path kept in view.
Methods under evaluation · not a performance benchmark
An exponential moving average can soften noisy frame-time samples. Separate recovery thresholds and time limits can reduce quality chatter; smoothing alone does not prove stability. Adaptive quality should remain optional and respect the selected preset.
NIST (n.d.)Cull conservatively against the receivers that may be visible, using one verified light direction. Refresh moving bounds and restore authored renderer settings. A fixed 48 m extrusion is not a safe rule for tall objects under a low sun.
Unity Technologies (n.d.-b) · Zhang et al. (2007)SSAO can strengthen contact in stone corners, at a sampling cost. ACES changes hue and saturation, so it must be compared with Neutral under matching exposure. Rough moss, damp stone and restrained magical cores need separate material responses; depth-of-field remains optional.
Unity Technologies (n.d.-a) · Unity Technologies (n.d.-c)APA-style primary sources support the methods, not project-specific performance claims. Documentation checked October 20, 2026. This engineering notebook is not a peer-reviewed publication.
