Member Discussion

PPF And Frame Rate

Every PPF study I've ever read records stationary subjects. Has anyone ever tried to figure out what effect frame rate could have on PPF?

In other words, what PPF do I need to identify a human subject moving at walking speed? What if my subject is running, or on a bicycle? Does having 60FPS, like the new Samsung cameras have, make a difference?

My feeling is that it shouldn't matter much, but I'd like to know if anyone ever tried it out.

I have never tested it but I agree with you that it shouldn't matter much.

Logically, to capture a clear image of a moving subject, (walking, running, bicycling, what have you), a sufficiently short shutter speed is much more important that a high frame rate. For example, assume a subject is sprinting at 15mph (22 feet per second). If the shutter speed is 1/30s, during the time the shutter is open the subject will move a distance of 8.8". That would likely cause significant blurring of the subject. Conversely, if the shutter speed is 1/300s, then the subject only moves 0.88" while the shutter is open. Depending on the field of view, that probably is not enough to cause significant blurring. If the shutter speed is too slow relative to the speed of the subject, causing blurring, then more pixels-on-target are not going to help you in any way.

Interesting question.

Unordered thoughts follow. Assuming subject is always properly framed.

On the one hand you would think that if the shutter speed was fast enough to prevent blurring, then frame rate would be irrelevant, assuming global shutter or CCD.

But since global shutter is not the norm, you might at some speed be effectively limited by the speed the sensor can be read at.

However if the SS was too slow, and blurring DID occur then you would effectively lose pixel information; exactly how much seems hard to guess, but if the subject can move more than 2 pixels in the time that the shutter is open you might not be able id the face regardless of the resolution.

Then again, there is some sort of clarifying effect that multiple frames from slightly different perspectives has. Most of us have experienced pulling a single representative 'snapshot' out of what appears to be a decent looking stream, only to find that the still looks way more noisy/blurry/low-res on its own. But when viewed at a decent framerate it seems to get sharper.

So maybe we need a PPF/sec metric?

Ari - I've played around with some tests trying to get to a similar answer. I haven't had the time to come up with a full formula, but the main factors seem to be:


2)Shutter speed

3) Object travel speed.

Roughly speaking, I've found that a travel distance of less then 2" frame-to-frame with 70+PPF results in pretty good images. With lower PPF, you'd want less travel distance frame to frame, and I think with higher PPF you could get away with a little more travel distance.

What I was trying to do was develop a formula factoring the relation of PPF, Shutter speed, Object speed and come up with a recommendation like "If you divide X by Y and multiply by Z and the answer is between 1 and 1.2 it should be good".