Avoid using get(), it is slow. Use the pixels[] array instead.

1.     for (int y = 0; y < video.height; y++) {
2.       for (int x = 0; x < video.width; x++) {
3.         color pixel = video.get(x, y);
4.         values[x][y] = int(brightness(pixel));
5.       }
6.     }
7.     for (int y = 0; y < video.height; y++) {
8.       for (int x = 0; x < video.width; x++) {
9.         stroke(values[x][y]);
10.         point(x+(width-video.width)/2, y+(height-video.height)/2, -values[x][y]);
11.       }
12.     }
13.   }

there's no reason this needs to be two different loops - i mean two separate blocks of nested loops.

also, all that maths in the point call is unnecessary and can be replaced with one translate() at the start of draw()

and that renders values[x][y] pretty pointless too (at the moment, perhaps you have plans) - you'd only need one int's worth of space for the current pixel.

i think the main problem is just that getting 153,000 values and displaying 153,000 dots takes a while (and it's slower in opengl because of the get() - video cards are asymmetrical in that it's fast to send data to them, slow to get data back from them)

how about this? better? also added colour