[robgough]

That explanation never made sense to me.

If windows has rotated everything, then chrome is now the right way round again and should appear to render properly.

The only reason it does that I can think of, is that chrome loaded when the screen was half way through a refresh - so we saw the now rendered bottom appear before it went back round and refreshed the top of the screen.

[Arve]

The point here is that the rotation is done by the OS, in software, so the screen is still painting from the physical left-to-right and top-to-bottom. This is something you'll never notice with the naked eye, but their camera uses a staggering 2700fps. This means that a full repaint of the screen - assuming 60Hz refresh happens in 45 frames. Those 45 frames, played back at 30 fps means that the full repaint takes a second and a half, leaving you with the ability to see the individual lines being (re)painted.

[boredguy8]

Windows doesn't change the hardware of the monitor. It just changes the orientation of what's on the screen.

[sapphirecat]

The display always paints from its own top to bottom. Since it has been turned upside down, when you watch really quickly, you can see it painting bottom-to-top from your point of view. This would be true whether the display has Windows rotation or not, and regardless of how apps are writing into the framebuffer.

Windows rotation makes the graphics driver output the framebuffer in reverse, so that the bottom lines of the image appear on the bottom lines of the display from your point of view. This makes the displayed image upright again.

To avoid flicker, Chrome may render offscreen and then copy the result top-to-bottom into the framebuffer. Hence Google's statement that Chrome renders top-to-bottom, and that the screen is displaying bottom-to-top due to rotation.

OK, I think that's enough pedantry for one comment ;)