elaTCSF: A Temporal Contrast Sensitivity Function for Flicker Detection and Modeling Variable Refresh Rate Flicker

TL;DR

elaTCSF extends the traditional Temporal Contrast Sensitivity Function by incorporating luminance, eccentricity, and area through a spatial probability summation framework to better predict flicker perception across variable contrast and large-field stimuli. Building on Watson's TCSF and the IDMS standard, the model combines a luminance channel $S_L(L)$, an eccentricity factor $S_{\mathrm{ecc}}(e)$, and a modified slope $S_{\omega}'(\omega,L,e)$ into a base elTCSF, then aggregates over space via $E(\omega,L,\hat{e},R)$ to yield the perceptual threshold $S_{elaTCSF}$. The model is fitted to eight datasets and validated on a VRR flicker dataset, achieving improved S-RMSE and CFF-RMSE compared with baselines and accurately predicting parafoveal CFF peaks and VRR flicker visibility. Practical applications include predicting safe VRR operating ranges, mitigating low-persistence flicker in VR headsets, and informing lighting-design flicker assessments, with a publicly available VRR flicker dataset to support ongoing validation. The work advances perceptual flicker modeling toward more realistic, condition-dependent predictions, though it remains focused on low-spatial-frequency, large-field stimuli and invites further validation across diverse contexts.

Abstract

The perception of flicker has been a prominent concern in illumination and electronic display fields for over a century. Traditional approaches often rely on Critical Flicker Frequency (CFF), primarily suited for high-contrast (full-on, full-off) flicker. To tackle varying contrast flicker, the International Committee for Display Metrology (ICDM) introduced a Temporal Contrast Sensitivity Function TCSF$_{IDMS}$ within the Information Display Measurements Standard (IDMS). Nevertheless, this standard overlooks crucial parameters: luminance, eccentricity, and area. Existing models incorporating these parameters are inadequate for flicker detection, especially at low spatial frequencies. To address these limitations, we extend the TCSF$_{IDMS}$ and combine it with a new spatial probability summation model to incorporate the effects of luminance, eccentricity, and area (elaTCSF). We train the elaTCSF on various flicker detection datasets and establish the first variable refresh rate flicker detection dataset for further verification. Additionally, we contribute to resolving a longstanding debate on whether the flicker is more visible in peripheral vision. We demonstrate how elaTCSF can be used to predict flicker due to low-persistence in VR headsets, identify flicker-free VRR operational ranges, and determine flicker sensitivity in lighting design.

Figures (15)

• Figure 1: Left: Measurement of a VRR display, which alternates between 30 Hz and 120 Hz in short time intervals. The upper plot shows drops in luminance (caused by V-blank) that vary in frequency depending on the refresh rate. The differences in the frequency of those drops cause small luminance differences and result in flicker. Right: The same signal in the frequency domain shows peaks at 4 Hz (the frequency of the refresh rate change), 30 Hz and 120 Hz (caused by V-blank).
• Figure 2: The contrast of VRR flicker at varying stimulus luminance and temporal frequencies or refresh rate switch, $F_{\mathrm{rrs}}$ [Hz]. Luminance levels above 8 cd/m$^2$ are irrelevant to us because none of our VRR experimental data reaches above that luminance.
• Figure 3: Our VRR Flicker dataset, where each point represents the average sensitivity across all participants. The error bars indicate the upper and lower bounds derived from psychometric function fitting.
• Figure 4: A photograph of the experimental setup. The experiment was taken in a dark room. The lights were added to take the photograph and were not present during the experiment.
• Figure 5: The inputs (left) and computation steps of elaTCSF.