Video Analysis With the iPhone 5s

I don’t have an iPhone 5s, but my brother does. In case you aren’t aware, the iPhone 5s has a slow motion video mode. In this mode, it records at 120 frames per second. The above video is an example of this new feature. Of course, the phone doesn’t play the video at 120 fps. […]

I don't have an iPhone 5s, but my brother does. In case you aren't aware, the iPhone 5s has a slow motion video mode. In this mode, it records at 120 frames per second. The above video is an example of this new feature. Of course, the phone doesn't play the video at 120 fps. Instead it changes the video so that it plays back in slow motion.

How well does this work with video analysis? That is the question. Let's find out. The plan is to use Tracker Video Analysis using a slow motion video as well as a normal video. I will just see what happens.

Here's what I did. I used my iPhone 5 and the iPhone 5s at the same time to record the same motion. I dropped some golf balls so that I could measure the motion. Oh, this was a mostly impromptu experiment so I just placed a 42 inch level on the ground for scale. I didn't realize that I would be all the way in the frame, otherwise I would have changed my hat. Oops, here is the normal video from the iPhone 5 for comparison.

With the iPhone 5's 30 frame per second video, I get the following for the y-position of that yellow-green ball after it bounces.

Data Tool

On the bounce, I get an acceleration of -8.79 m/s2. Yes, that isn't what it is supposed to be but that's ok. I suspect this is because of a scaling problem. If the orange level on the ground isn't perfectly perpendicular to the camera, it won't give the correct scale. Really, I am comparing two cameras so it doesn't matter too much.

Now to look at the same ball at the same time with the iPhone 5s. Here is the plot of the vertical motion in that case.

Data Tool

You will notice that this plot has many more data points. However, the acceleration from the fit is still just -8.812 m/2. This is still too low. I suspect some type of calibration problem.

Are more data points better? Well, yes. However, you do lose something. Here is a frame from both videos around the same time.

Fall 13 Sketches key

The iPhone 5 video in 30 fps has better resolution.

So, which should you use? I guess in this case both cameras give reasonably similar results. I should problem try some examples where the iPhone 5 isn't fast enough to be useful and see if the 5s can do the job.

If you need a refresher on video analysis, here is a detailed example showing the acceleration of a grasshopper.