psychopy.org | Reference | Downloads | Github

Antialiasing does not work properly in online experiments

URL of experiment: https://run.pavlovia.org/behrengi/attention_exp1/?__pilotToken=6512bd43d9caa6e02c990b0a82652dca&__oauthToken=b7c6c24d5e2e8aa5b30b4ca7031b65631a28f1ed93fefc9d0a19e134b3599b5f

code: Bihter Akyol / attention_exp1 · GitLab

Description of the problem:

Hello,

I am using the builder and creating stimuli using code components. I need to show some dots and an arrow. The problem is, the dots and the head of the arrow look antialiased in Psychopy but not in the online version. Aliasing gets more noticeable when stimuli are larger, and I am planing to adjust stimulus size to present them in visual angles in the actual experiment, which means I cannot simply use a fixed stimulus size and I do not know how prominent this lack of antialising issue will be for different participants.

The python code of the stimuli as follows:

heads.append(visual.ShapeStim(
        win=win, name='cue_head', size=1,
        vertices=[[-head_size[0] / 2, 0], [head_size[0] / 2, -head_size[1] / 2], [head_size[0] / 2, head_size[1] / 2]],
        ori=0, pos=-(rl + rect_size[0] / 2 + head_size[0] / 2),
        lineColorSpace='rgb', lineColor=white,
        fillColor= white, fillColorSpace='rgb',
        opacity=1,  interpolate=True))

dots[i].append(visual.Polygon(
                win=win,
                lineColorSpace='rgb', fillColorSpace='rgb',
                fillColor= white,
                lineColor= white,
                edges=32,
                pos = [-300 + (random() - 0.5) * 2 * 100, 
                       0 + (random() - 0.5) * 2 * 100],
                size=30,
                lineWidth=1,
                opacity=1, 
                interpolate=True,
                closeShape=True
                ))


This issue applies to the fixation cross when it is rotated (50 degrees, image is not included here), which I created with the predefined stimulus components in psychopy and not with a code component.

I am trying to run the experiment on a mac in Chrome. I tested with Safari too, and the problem persists. The psychopy version is v2020.2.10


I investigated this issue further by using only builder components (here). Aliasing problem applies to visual.Polygon but not to visual.TextStim. Any ideas why this may be happening?

Also, I noticed that assigning an orientation to visual.Polygon makes the stimuli rotate to different directions (clockwise vs. counterclockwise) in PsychoPy and pavlovia.

I’m having the same issue. Were you able to fix this?