Thursday, April 24, 2008

Race Photography

In my comment on Profgrrrrl's blog about making photographs of fast-moving vehicles, I mentioned my experiences at Indianapolis "last year". Actually, it was 2006 or it would have been in my blog from last summer! These were taken during open practice, where parking and admission together were $5, during a pit stop in the middle of a 600 mile drive. Made me late for dinner, but worth it.

Here is one of them (all three photos were cropped to 2400x1200 out of the full frame, reduced to 600x300, and "saved for web" to save space so not nearly as nice as the originals).

Details and two other examples (one spectacularly lucky) below the fold.

The key detail is that Tony Kanaan is going 219 mph when this was taken, which is 321 ft/s. I had set the camera to shutter priority at 1/200, meaning a blur distance of 1.6 feet. I was tracking the car, so the background is blurred. You will also notice that the tires and wheels are blurred, because I am only freezing the linear motion of the car.

The first pictures I took were awful (frozen car), but then I remembered that I could calculate the right shutter speed in my head. Long ago my dad taught me the conversion rule that 60 mph = 88 ft/s. Useful for driving safety, among other things. I didn't do long division, just estimated that 220 is between 180 (3x60) and 240 (4x60) miles per hour so the speed was between 3x90 (270) and 4x90 (360) feet per second. That meant 1.5 feet at 1/200 and 3 feet at 1/100.

OK, so can I get the car blurred and the background in focus? Easier said than done when using a telephoto lens. I deleted lots of photos with half a car or no car before I got a handle on my reaction time. (The camera has no delay other than the mirror flipping up, but human reaction time of 0.15 s means the car travels almost 50 feet in the time it takes to react. Shoot when you see it in the viewfinder and it is gone.) The trick I came up with was to focus the camera the way I normally do (one eye closed) but then open both eyes. My right eye kept the camera centered on the Indy logo and the track. I pulled the trigger when my left eye got a glimpse of the car appear to the right of the camera, and I got lots that look like this one:

I think that is a "3" on the car, so this is probably Helio Castroneves. I'm pretty sure this was made at 1/100, since there are others that don't have quite this nice smooth blur (and where you can read the numbers).

Those pictures were all taken from the Tower Terrace near the end of the front straight, looking down over the fence so there is a clear view of the track. This last one was taken on the backstretch looking through the fencing from one of the viewing mounds.

This is Danica Patrick in her "T" car. This is back at 1/200, but you can see how the fence in front of me is blurred from both depth of field and motional blurring. Only the horizontal parts of the fence show up.

Closing remark about animated films:
Motional blurring is a big deal in getting realism in digital (or hand) animation. Freeze any frame from a movie and you will see what I mean. If you go back to the beginning of digital work (Tron), you will notice that some of its lack of realism comes from how they handled motion back in 1982.


dr. dave said...

I actually use photos like this when teaching the Uncertainty Principle to my liberal arts students. Yes it's in imperfect analogy for all sorts of reasons (and I tell them that) but it gets the point across that position and velocity are complimentary quantities. In one picture you can tell exactly where the car is, and in the other, if you know the shutter speed, you can tell how fast it's going. But you can't take a picture that tells you both exactly.

Yeah, it's a stretch.

Doctor Pion said...

Thanks for that idea! Feel free to use these pictures in your class.

The analogy is not that much of a stretch, actually, since neither picture really tells you both x and v. The car may be sharp in the first picture, but we don't know where it is (the track is blurred). The track may be sharp in the second picture, but we don't know where the car is (at the center?).

Both remain true as shutter speed increases, although we might not notice the problem. I should think about using this in a discussion of why we always measure average velocity in the lab.

BTW, do you know that a more literal translation is Heisenberg's Unfocused Principle?