Abstract
Understanding the perception of humanoid character motion can
provide insights that will enable realism, accuracy, computational
cost and data storage space to be optimally balanced. In this sketch
we describe a preliminary perceptual evaluation of human motion
time warping, a common editing method for motion capture data.
During the experiment, participants were shown pairs of walking
motion clips, both time warped and at their original speed, and asked
to identify the real animation. We found a statistically significant
difference between speeding up and slowing down, which shows
that displaying clips at higher speeds produces obvious artifacts,
whereas even significant reductions in speed were perceptually acceptable.
1 Computer Animation and Perception
Computer animation and motion perception are closely related
fields, as the result of motion synthesis is always presented to a live
observer. In this study, we focus on motion time warping, a method
used in both parametric models and state machines for matching
and transitioning between clips of the same type with different timing
or speed. Timewarping, and particularly dynamic time warping,
originates in speech recognition and was succesfully used, alongside
other signal processing techniques, on animation data (e.g.,
Bruderlin and Williams [1995]). However the perceptual implications
of such manipulations have not been extensively studied and
are hard to predict. Our evaluation approach is closely related to
the work of Reitsma and Pollard [2003], but we focus on human
locomotions rather than generic ballistic motion.
2 Experiment Design
Five motion captured clips of a walking animation served as the
stimuli for our experiment. These five animation speeds covered a
normal range of human walking, ranging from 0.8 m/s to 2.4 m/s
with 0.4 m/s increments. For each clip, we created 4 other versions
using time warping to match the speed of the other clips, leading to
a total of 25 clips (e.g., the 1.2 m/s motion was slowed down to
0.8 m/s and speeded up to 1.6, 2.0 and 2.4 m/s). We hypothesised
that time warping would be less noticeable if the time warped speed
is close to the original speed of the clip.
The experiment consisted of sequences depicting two animated
characters side-by-side (Figure 1). Both characters (Figure 1b)
were either stick figures or geometric models (the model’s realism
was found to affect perceptual sensitivity to errors in motion [Hodgins
et al. 1998]), facing forward or sideways (to test if motion error
sensitivity is affected by viewpoint), with each simultaneously displayed
pair using the same setup. One character’s animation consisted
of the original motion, randomly placed on the left or right
side of the screen, while the other was time warped to match its
speed. We also tested the real animation against itself as a control
case.
Sixty naive participants from the general public (64% male and
36% female) volunteered for this experiment. The instruction sheet
indicated that one of the motions was a real captured motion and
the other one was synthetically edited. The task was to indicate
which of the two animations was the real motion by clicking the
left or right mouse button. Each participant completed 100 trials in
randomized order (25 motion clips, 2 models, 2 viewpoints).
AbstractUnderstanding the perception of humanoid character motion canprovide insights that will enable realism, accuracy, computationalcost and data storage space to be optimally balanced. In this sketchwe describe a preliminary perceptual evaluation of human motiontime warping, a common editing method for motion capture data.During the experiment, participants were shown pairs of walkingmotion clips, both time warped and at their original speed, and askedto identify the real animation. We found a statistically significantdifference between speeding up and slowing down, which showsthat displaying clips at higher speeds produces obvious artifacts,whereas even significant reductions in speed were perceptually acceptable.1 Computer Animation and PerceptionComputer animation and motion perception are closely relatedfields, as the result of motion synthesis is always presented to a liveobserver. In this study, we focus on motion time warping, a methodused in both parametric models and state machines for matchingand transitioning between clips of the same type with different timingor speed. Timewarping, and particularly dynamic time warping,originates in speech recognition and was succesfully used, alongsideother signal processing techniques, on animation data (e.g.,Bruderlin and Williams [1995]). However the perceptual implicationsof such manipulations have not been extensively studied andare hard to predict. Our evaluation approach is closely related tothe work of Reitsma and Pollard [2003], but we focus on humanlocomotions rather than generic ballistic motion.2 Experiment DesignFive motion captured clips of a walking animation served as thestimuli for our experiment. These five animation speeds covered anormal range of human walking, ranging from 0.8 m/s to 2.4 m/swith 0.4 m/s increments. For each clip, we created 4 other versionsusing time warping to match the speed of the other clips, leading toa total of 25 clips (e.g., the 1.2 m/s motion was slowed down to0.8 m/s and speeded up to 1.6, 2.0 and 2.4 m/s). We hypothesisedthat time warping would be less noticeable if the time warped speedis close to the original speed of the clip.The experiment consisted of sequences depicting two animatedcharacters side-by-side (Figure 1). Both characters (Figure 1b)were either stick figures or geometric models (the model’s realismwas found to affect perceptual sensitivity to errors in motion [Hodginset al. 1998]), facing forward or sideways (to test if motion errorsensitivity is affected by viewpoint), with each simultaneously displayedpair using the same setup. One character’s animation consistedof the original motion, randomly placed on the left or rightside of the screen, while the other was time warped to match itsspeed. We also tested the real animation against itself as a controlcase.Sixty naive participants from the general public (64% male and36% female) volunteered for this experiment. The instruction sheetindicated that one of the motions was a real captured motion andthe other one was synthetically edited. The task was to indicatewhich of the two animations was the real motion by clicking theleft or right mouse button. Each participant completed 100 trials inrandomized order (25 motion clips, 2 models, 2 viewpoints).
การแปล กรุณารอสักครู่..
