Elementarraystim - applying a texture/mask to the field rather than the elements?

Coding
1

Hi everyone!

I am currently programming a so-called “dot lattice” stimulus that consist of a grid of regularly spaced dots. As the ElementArrayStim is suitable to easily display such stimuli, I started to use this object. However, in the future I would probably want to add a Gaussian window to this stimulus. This window should be applied to the whole stimulus rather than each object separately, as shown in the image below.

Am I right that it currently is not possible to apply such a “field-wide” Gaussian mask to an ElementArrayStim? What would be the best strategy to create such a stimulus? First, create an image of the stimulus, load it again, and multiply this texture with a Gaussian? I don’t feel literally stuck here, but I am curious whether there are elegant solutions to this problem.

2

You could perhaps do something like the following to get a gaussian mask (assuming dots is an ElementArrayStim):

dots.opacities = np.prod(
    [
        psychopy.filters.makeGauss(xys, sd=np.ptp(xys) / 6.0)
        for xys in dots.xys.T
    ],
    axis=0
)

This would only really work properly if the elements are not overlapping, and you would probably want to set sd to be specific to the size of the field.

3

Thanks for this suggestion! At first, I was thinking that it might be problematic that a single dot might receive the same opacity value everywhere, but from the stimulus examples I have available, this seems to be the case also. I will mark this as the final solution.

4

Yes, fair point about being constant within an element. The below is somewhat inelegant but would seem to work more completely:

import numpy as np

import psychopy.visual
import psychopy.event

import pyglet.gl

bg_col = -1.0

win = psychopy.visual.Window(
    size=[400, 400],
    units="pix",
    color=[bg_col] * 3,
    fullscr=False
)

n_dots = 500

dots = psychopy.visual.ElementArrayStim(
    win=win,
    nElements=n_dots,
    xys=np.random.uniform(-120, 120, (n_dots, 2)),
    units="pix",
    sizes=5.0,
    elementTex=None,
    elementMask="circle"
)

mask_tex = psychopy.visual.GratingStim(
    win=win,
    tex=np.ones([256, 256]) * bg_col,
    mask="gauss",
    units="pix",
    size=[256, 256]
)

dots.draw()

pyglet.gl.glBlendFunc(
    pyglet.gl.GL_ONE_MINUS_SRC_ALPHA,
    pyglet.gl.GL_SRC_ALPHA
)

mask_tex.draw()

pyglet.gl.glBlendFunc(
    pyglet.gl.GL_SRC_ALPHA,
    pyglet.gl.GL_ONE_MINUS_SRC_ALPHA
)

win.flip()

psychopy.event.waitKeys()

win.close()
5

Thanks again for this great suggestion! It seems I can only mark a single answer as the final solution, so I will leave the first one as it seems the best trade-off between what people will want to achieve and simplicity of the code.