The “blue” tiles on the top face of the left cube are NOT the same color as the “yellow” tiles on the top face of the right cube. Each tile in both sets is made of pixels of several different colors. The “blue” tiles on the left are made of several shades of slightly bluish-gray; the “yellow” tiles on the right are similarly composed of several shades of slightly yellowish-gray.
Such illusions may indeed be real, but that image does not seem to be a good example.
Edit: Interesting. The higher-resolution version of the image (in mattnewport’s post) is indeed as described, i.e. the colors of the relevant tiles are identical. Is the deviation in the lower-res image a result of compression?
It looks like it’s compression. The image in the post is JPEG, which uses lossy compression; the one in mattnewport’s post is PNG, which uses lossless compression.
I think it might be some dodgy processing or gamma correction at some point in the image’s history. Some of the ‘yellow’ tile pixels in my image have very slightly different RGB values than the ‘blue’ tile pixels. Applying a gamma curve, saturation adjustment or other image processing could have resulted in the greater differences seen in the image in the post.
The “blue” tiles on the top face of the left cube are NOT the same color as the “yellow” tiles on the top face of the right cube. Each tile in both sets is made of pixels of several different colors. The “blue” tiles on the left are made of several shades of slightly bluish-gray; the “yellow” tiles on the right are similarly composed of several shades of slightly yellowish-gray.
Such illusions may indeed be real, but that image does not seem to be a good example.
Edit: Interesting. The higher-resolution version of the image (in mattnewport’s post) is indeed as described, i.e. the colors of the relevant tiles are identical. Is the deviation in the lower-res image a result of compression?
It looks like it’s compression. The image in the post is JPEG, which uses lossy compression; the one in mattnewport’s post is PNG, which uses lossless compression.
I think it might be some dodgy processing or gamma correction at some point in the image’s history. Some of the ‘yellow’ tile pixels in my image have very slightly different RGB values than the ‘blue’ tile pixels. Applying a gamma curve, saturation adjustment or other image processing could have resulted in the greater differences seen in the image in the post.