Inconsistencies with `pixelDensity()` across P2D and FX2D renderers in MacOS

Created by: haschdl

I was playing around with loadPixels() and updatePixels() and I found a strange behaviour, in which using FX2D with MacOS creates a blur effect after redraws. Another person using Windows did not experience that, which led to me to check for pixelDensity() - adding pixeldDensity(2) to the sketch running FX2D "fixes" the problem.

It would be great to know what is happening "behind the scenes", as this difference in behaviour affects other sketches. In particular, is there a reason why leaving out pixelDensity() or adding pixelDensity(1) creates the blur effect?

As for the bug report, I believe P2D and FX2D should produce the same output for such trivial examples. Perhaps the bug is along the lines "FX2D renderer does not detect the correct pixel density in MacOS" or something like that.

The following code works as expected in P2D, but produces an unexpected blur effect when using FX2D.

import processing.javafx.*;

void setup() {
  size(500, 500);
  background(0);
}

void draw() {
  noFill();
  stroke(255, 200);
  circle(random(width), random(height), random(100, 200));
  loadPixels();
  updatePixels();
}

Using P2D:

Using FX2D:

Description

Using the code above produces different output using F2XD.

Expected Behavior

Output should be same regardless of renderer.

Current Behavior

There is a blur effect using FX2D.

Steps to Reproduce

Run a sketch using the code below, alternating between P2D and FX2D renderers.

Your Environment

• Processing version: 4.2
• Operating System and OS version: MacOS 13.1 Ventura, Macbook pro 13-inch
• Other information: Other users reported this behavior cannot be reproduced in Windows

Possible Causes / Solutions

This seems to be related to pixelDensity(), or to be precise, the default value for pixelDensity() using FX2D seems to be incorrect.

Comienza ejecutando el sketch proporcionado con P2D y FX2D en macOS, comparando los resultados con la densidad de píxeles predeterminada y con pixelDensity(1) o pixelDensity(2). Sigue cómo cada renderer gestiona pixelDensity(), loadPixels() y updatePixels(); se considera terminado cuando los sketches equivalentes producen la misma salida sin el desenfoque de FX2D.