Transient Optical Detection Transition Revisited Why the geometric reanalysis is stronger than the original pilot result - while remaining a model-defined detectability study

Abstract

The original Transient Optical Detection Transition (TODT) pilot study reported that a physically present falling quartzobject entered a brief model-defined subthreshold state in a natural indoor scene. In the strict observer model, frames430-442 were classified as VISIBLE, frame 443 as SUBTHRESHOLD, and frame 444 as TRANSITION; frame 443 hada measured local scene contrast of 1.86%, a median modeled effective contrast of 0.256%, and a modeled belowthreshold probability of 97.7%. The longest marginal/subthreshold run on the decoded 30 fps timeline was 33.3 ms [1].The present reanalysis does not replace that finding. Instead, it asks whether the same event can be recovered by asubstantially different, compact geometric representation that does not reuse the original probabilistic observerclassification. Using pre-event frames 431-442 to freeze the reference state and holding frame 443 out, the reanalysisplaced frame 442 on one side of a signed dimensionless state-space boundary and frame 443 well inside the modeldefined harder-to-see side. The signed state coordinate changed from -0.867 at frame 442 to +2.601 at frame 443,with an orthogonal boundary depth of +1.502 standardized units. The 442-to-443 crossing remained present in 1,920of 1,920 analysis variants, in 99.48% of 100,000 uncertainty perturbations, and throughout the full stated one-sigmabounded-error box for the baseline analysis. Thus, relative to the original TODT pilot, the reanalysis adds hold-outtesting, explicit geometric localization, method robustness, uncertainty stress testing, and a deterministic conditionalcertificate. It remains a single-object computational study and does not demonstrate physical transparency orconscious perceptual disappearance. The detailed implementation is not publicly released; the analysis is deliberatelydescribed at an abstract state-space level. Keywords: visual detectability; transient optical detection transition; state-space geometry; robustness; boundarycrossing; retinal image motion; contrast