Zoom, baby, zoom*

For a few months now, I have been spending (wasting?) some time with a gadget called Gigapan, a robot that can take hundreds of shots of the same scene with a simple point-and-shoot camera. The pictures are taken in a well-defined rectangular grid pattern so that there is the right amount of overlap between all neighbors. Later the photos can be stitched into a gigantic photograph on a computer and shared with the world through the Gigapan.org website and, even better, through Google Earth. [If you are a tiny bit familiar with geoblogs, you must have seen some of the gigapans that Ron Schott has put together; he is one of the earliest and most enthusiastic adopters of the technology and has assembled an impressive set of panoramas on the gigapan site.]

I have to confess that I had to actually buy this thing and start playing with it to realize how different gigapixel panoramas are from the usual few-megapixel digital photographs. The idea is simple: a ten megapixel camera takes photos that contain ten million pixels; if you put together a 10x10 grid of such photographs into one image, you end up with a gigapixel panorama. Because some overlap is needed between the photographs, more than 100 pictures are necessary to exceed the gigapixel limit. But the point is that the more pixels there are in a photograph, the more information it contains and the more sense it makes to zoom in and see the details - details that are usually non-existent in a conventional digital picture. The other side of the coin is that it is only worth taking gigapans of scenes with plenty of small-scale and variable detail (although I am getting to the point that I see a potential gigapan everywhere).

I do not think that gigapixel images will replace conventional (that is, megapixel) photography. There is only a limited number of things that the human eye can see at one time; and often the value of a good photograph comes not from the pixels it captures, but from the ones it consciously ignores. Beauty and the message an image can hold are scale-dependent; and zooming in to see the irrelevant detail could be a distraction.

That being said, I am all for taking home as many pixels as possible from outcrops and landscapes in general. The gigapan system is simple and works surprisingly well, and it *is* exciting to explore big outcrop panels from the scale of entire depositional systems to the laminae of single ripples or even grains.

No photos or panoramas posted/embedded this time; but here is a link to my giga-experiments.

* title is courtesy of Kilgore661