elaTCSF: A Temporal Contrast Sensitivity Function for Flicker Detection and Modeling Variable Refresh Rate Flicker
Yancheng Cai, Ali Bozorgian, Maliha Ashraf, Robert Wanat, Rafał K. Mantiuk · 2025
Read elaTCSF: A Temporal Contrast Sensitivity Function for Flicker Detection and Modeling Variable Refresh Rate Flicker by Yancheng Cai, Ali Bozorgian, Maliha Ashraf, Robert Wanat, Rafał K. Mantiuk (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-16759.txt.
Introduction
Yancheng Cai and co-authors (2025) build a better yardstick for when flickering light becomes visible. Flicker matters in lighting and in displays, and modern screens with variable refresh rates can shimmer in ways older standards miss. The classic measure, critical flicker frequency, suits only full on-off flashing. A newer standard function handles gentler contrasts but ignores three things that change sensitivity: brightness, where in the visual field the flicker sits, and how large an area flickers.
Their function, elaTCSF, extends the standard by adding exactly those three factors, plus a new model of how the visual system sums flicker signals across space. They trained it on data from varied flicker experiments. The result predicts detection across a wider range of real displays, including the low spatial frequencies where earlier models fail. Display makers can use it to certify that a screen's refresh tricks stay invisible.
Catalog record
| CREATOR | Yancheng Cai, Ali Bozorgian, Maliha Ashraf, Robert Wanat, Rafał K. Mantiuk |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 41 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-16759
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
elaTCSF: A Temporal Contrast Sensitivity Function for Flicker Detection and Modeling Variable Refresh Rate Flicker Yancheng Cai, Ali Bozorgian, Maliha Ashraf, Robert Wanat, Rafał K. Mantiuk 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. VRR variable refresh rate CFF Critical Flicker Frequency TLM temporal light modulation TCSF Temporal Contrast Sensitivity Function CSF Contrast Sensitivity Function fMRI functional magnetic resonance imaging ICDM International Committee Display Metrology IDMS Information Display Measurements Standard GPU Graphics Processing Unit V-blank Vertical Blanking Interval RMSE root-mean-square-error Conference: SIGGRAPH Asia 2024 Conference Papers; December 3–6, 2024; Tokyo, Japan SIGGRAPH Asia 2024 Conference Papers (SA Conference Papers ’24), December 3–6, 2024, Tokyo, Japan DOI: 10.1145/3680528.3687586 ISBN: 979-8-4007-1131-2/24/12 297 CCS: Computing methodologies CCS: Computing methodologies Perception Figure 1. (a) 3D Visualization of elaTCSF. elaTSCF predicts sensitivity and critical flicker frequency along three dimensions: Eccentricity, Luminance, and Area. (b) elaTCSF predicts flicker visibility in short-persistence displays, such as VR/AR headsets. The dashed lines indicate the current display capabilities (peak luminance and official refresh rate). Above the CFF threshold (colored area), flicker is unlikely to be visible. 1. Introduction Flicker, known as TLM ( TLM ), has been a significant concern in illumination engineering for over a century. It has detrimental effects on human physiology, ranging from irritation to neurological disturbances ( Miller et al., 2023 ) . Flicker has been observed in different technologies such as fluorescent lamps ( Eastman and Campbell, 1952 ) in the 1940s, cathode-ray tubes (CRTs) ( Bauer et al., 1983 ) in the 1980s, and LED ( Lehman et al., 2011 ) . The CFF ( CFF ) is the best-known flicker detection standard that has been in use for over a century ( Porter, 1902 ) , defined as the frequency at which a flickering light becomes indistinguishable from a steady, non-flickering light. However, it assumes a high-contrast flicker (light on and off), which is not always applicable. For example, flicker found in VRR ( VRR ) displays is caused by slight luminance differences at varying refresh rates, resulting in a low-contrast flicker. To address flicker of various contrast, Watson proposed a flicker detection metric ( Watson and Ahumada, 2011 ) , which relies on the TCSF ( TCSF ) equation from ( Watson et al., 1986 ) . This metric (TCSF IDMS 1 1 1 TCSF describes human visual sensitivity as a function of temporal frequency, exhibiting diverse forms. TCSF IDMS specifically denotes the standard proposed by the International Committee for Display Metrology (https://www.sid.org/Standards/ICDM). ) was later incorporated into the information display measurements standard (IDMS) v1.2 by the ICDM ( ICDM ) and was released in 2023. However, as acknowledged by Watson and Ahumada (2011) , other factors such as luminance, area, and eccentricity also affect the shape of TCSF , yet these parameters have not been integrated into the current standard, limiting its applications. Existing CSF , such as Barten’s ( 1999 ) , stelaCSF ( Mantiuk et al., 2022 ) and castleCSF ( Ashraf et al., 2024 ) predict sensitivity as a function of mentioned factors, but (a) do not offer sufficient accuracy at high temporal frequencies for modeling flicker; (b) cannot explain flicker visibility for large sources that span a significant portion of the visual field. This paper aims to extend the TCSF IDMS to incorporate the effects of eccentricity, luminance and area (hence elaTCSF), updating the standard for flicker detection. A 120-year-long debate persists in flicker detection: Is temporal sensitivity to flicker higher in the peripheral visual field (parafovea) than in the central (fovea)? The organization of the visual field is topographically represented in the striate cortex, with central regions receiving a disproportionately larger representation than peripheral areas ( Daniel and Whitteridge, 1961 ) . However, regarding temporal sensitivity, experimental measurements are conflicting and puzzling: some researchers observe a rise and fall in CFF with increasing eccentricity, peaking in the parafovea ( Porter, 1902 ; Phillips, 1933 ; Hylkema, 1942 ; Hartmann et al., 1979 ; Rovamo and Raninen, 1984 ; Tyler, 1987 ; Krajancich et al., 2021 ) , while others contend that temporal sensitivity peaks at fovea and decreases steadily with eccentricity ( Ross, 1936 ; Chapiro et al., 2023 ) . Presently, a consensus remains elusive, although the data from Hartmann et al. (1979) shows that larger stimulus areas, higher luminances, and higher contrasts are more likely to elicit peaks in the parafovea. We rely on those findings, including modern fMRI ( fMRI ) measurements of cortex activity ( Horiguchi et al., 2009 ; Himmelberg and Wade, 2019 ) , and propose a model that can explain these seemingly conflicting observations . Larger signal areas generally activate a broader range of retinal cone and rod cells, enhancing human sensitivity. Some CSF models like castleCSF ( Ashraf et al., 2024 ) have incorporated area (size) as a crucial parameter. However, they treat area and eccentricity as independent parameters. In reality, for stimuli with large areas, a range of eccentricities is involved, particularly evident in our VRR flicker dataset, which includes full-screen flicker. Relying solely on one eccentricity value is evidently insufficient. In this context, we propose a spatial probability summation model that can work with stimuli spanning any portion of the visual field. In summary, our contributions are as follows: • We introduce elaTCSF with a spatial probability summation model 2 2 2 The code for the model and the datasets used to train it can be found on the project page: https://www.cl.cam.ac.uk/research/rainbow/projects/elaTCSF/ . , which accounts for eccentricity, luminance, and area, extending the industry flicker detection standard TCSF IDMS . We also address past controversies regarding parafovea sensitivity peak. • We measure the visibility of flicker on a VRR display. The measurements are combined with publicly available flicker detection data to calibrate and test elaTCSF. The dataset will be made publicly available. • We demonstrate several applications of the model, including predicting safe refresh rate ranges for VRR displays, addressing VR headsets low-persistence flicker, and assisting in lighting design. 2. Related Work 2.1. Temporal sensitivity and Critical Flicker Frequency The neurons in the retina, thalamus, and subsequent stages of the visual pathway are sensitive to temporal variations in