ut ut

Laboratory for Image & Video Engineering

Spatiotemporal flicker detector model of motion silencing

Motion silencing research at LIVE is being conducted in collaboration with CPS (http://www.utexas.edu/cola/cps/).


A recent study has demonstrated that objects changing in color, luminance, size, or shape appear to stop changing when they move rapidly in collective motion (Suchow & Alvarez, 2011). One hundred small dots were randomly arranged in a ring-shaped pattern around a central fixation mark (see the illusion at http://visionlab.harvard.edu/silencing/). Each dot changed continuously over time in color, luminance, size, or shape. The changes were easily noticeable when the dots were stationary, but the changes were undetectable when the dots were suddenly sent into continuous rotational motion. This motion-induced failure to detect change, known as silencing, not only suggests the tight coupling of motion and object appearance but also reveals that motion can disrupt the perception of salient changes in visual objects. The mechanisms underlying this phenomenon remain unknown, although we expect that it might be explainable using known mechanisms. Here, we aim to model the effects of the illusion using well-known spatiotemporal filter-based models.

To determine why the visual system silences changes of objects in the presence of rapid motion, we constructed similar visual stimuli to induce the conditions under which silencing occurs. Observations and systematic spatiotemporal spectral analyses of the presentation data led us to develop a simple filter-based hypothesis and a spatiotemporal flicker detector model to explain the motion silencing phenomenon. We conducted a series of human psychophysical experiments to understand whether our filter-based hypothesis could explain the motion silencing illusion.

From the results, we found that the threshold of silencing occurs when the log frequency of object replacement is roughly one quarter of the log flicker frequency (the mean slope is approximately 0.27). The dependence of silencing on object spacing may be explained as a phenomenon of temporal sampling of the stimuli by the visual system. Our proposed model successfully captures the psychophysical data over a wide range of velocities and flicker frequencies.

Example Stimuli

  • We highly recommend viewers to download video stimuli and display them on your PC at 1366 × 768 resolution (or full screen mode), 60 Hz refresh rate monitor, with viewing distance 57cm to watch the exact stimuli. (Video stimuli is 60 frame per second. Web rendering can be different from the original stimuli or choppy according to your PC setup. We recommend high performance graphics card, monitor, and solid-state drive(SSD) for smooth presentation of stimuli.)

  • Download: (ThreeRings72Dots, OneRing10Dots, OneRing12Dots, OneRing18Dots, and OneRing24Dots).

  • Instructions: Play the video while looking at the red central fixation mark of the ring. At first, the ring is motionless and it's easy to notice that the dots are changing luminance. When the ring begins to rotate, the dots suddenly appear to stop changing. However, in reality luminances of dots are changing the entire time. Compare the level of silencing between videos. When the number of dots increased, the change of luminance is less visible while the ring is rotating.

    • Three Rings 72 Dots

    • One Rings 10 Dots

    • One Rings 12 Dots

    • One Rings 18 Dots

    • One Rings 24 Dots

  • Relevant publication

  • L. K. Choi, A. C. Bovik, and L. K. Cormack, "Spatiotemporal Flicker Detector Model of Motion Silencing," Perception, vol. 43, no. 12, pp. 1286-1302, Dec. 2014. (PDF)
  • L. K. Choi, A. C. Bovik, and L. K. Cormack, "A Flicker Detector Model of the Motion Silencing Illusion," Journal of Vision, vol. 12, no 777. pp. 777, Aug. 2012. (Abstract)