The devicePixelContentBoxSize read-only property of the ResizeObserverEntry interface returns an array containing the size in device pixels of the observed element when the callback is run.

An array containing objects with the new size of the observed element in device pixels. The array is necessary to support elements that have multiple fragments, which occur in multi-column scenarios. Each object in the array contains two properties:

blockSize

The size of the content-box, in device pixels, of the block dimension of the observed element. For boxes with a horizontal writing-mode, this is the vertical dimension, or height; if the writing-mode is vertical, this is the horizontal dimension, or width.

inlineSize

The size of the content box, in device pixels, of the inline direction of the observed element. For boxes with a horizontal writing-mode, this is the horizontal dimension, or width; if the writing-mode is vertical, this is the vertical dimension, or height.

The following example is taken from the article Pixel-perfect rendering with devicePixelContentBox. As the callback function of a ResizeObserver is called after layout but before paint. This provides an opportunity to log the exact size in physical pixels to ensure a one-to-one mapping of canvas pixels to physical pixels.

const observer = new ResizeObserver((entries) => {
  const entry = entries.find((entry) => entry.target === canvas);
  canvas.width = entry.devicePixelContentBoxSize[0].inlineSize;
  canvas.height = entry.devicePixelContentBoxSize[0].blockSize;

  /* … render to canvas … */
});
observer.observe(canvas, { box: "device-pixel-content-box" });