Testing Shutter Speed With iPhone

[ic-racer]

I saw a thread somewhere (can't find it now) where an iPhone was used to test shutter speeds.

I tried this and this is my evaluation.

On my iPhone I set it to "Slo-Mo" and made a video of various things, including a clock and various leaf and guillotine shutters.

To analyze the movie I imported it into a Digital Audio Workstation (DAW) software (Logic; Apple Computer, Cupertino California).
In this software, each movie frame is broken down in to 100 subframes for audio synch.

The DAW software clock counter (not adjustable that I can tell) was set to rollover to zero after counting up to 24.80 frames. My video of a clock gave about 180 frames per second.

I also analyzed some of the movie frames against the DAW software clock and found my movie contained a new frame about every 1/180th of a second.

That would be a Nyquist frequency of 180Hz so the best this could do for shutter speed testing was about 1/90th of a second (one-half the Nyquist frequency).

Indeed, the top speed on all my cameras measured about 1/100th of a second with this method.

Conclusion: The method was most useful for shutter speeds 1/60th and slower.

 

[Donald Qualls]

Different phones have different frame rate capabilities. My first-generation Pixel apparently can do the 240 fps referenced in the original article (I think it was posted in the Stearman Press subforum, just a couple weeks ago) -- at least, it claims 8x slo-mo and is nominally 30 fps HD in normal video mode. Also, Nyquist frequency is of less concern in this application, than counting frames in which the shutter was open vs. closed -- so you ought to be able to (approximately) time a shutter down to 1/150 or so at 180 fps and around 1/200 with 240 fps. Further, by recording multiple shutter cycles and averaging the counts, you can probably get pretty decent accuracy up to the full "one frame open" speed (in my case, about 1/250).

Sure, this still a limited method, but if you combine it with audio speed checkers (that record and detect the sound of the shutter opening and closing, for leaf shutters), you should be able to do pretty well without building or interfacing any kind of hardware detector. Honestly, for my large format operation, I don't care much about leaf shutter speeds above 1/200 or 1/250 anyway (relatively slow film, small apertures, my shutter speeds tend to run in the 1/25 range). And for my Speed Graphic, the focal plane shutter works correctly, including accurate curtain speeds, at least based on measurements made several years ago.

 

[ic-racer]

Not sure which generation phone I have, it was free from work. A shutter faster than the Nyquist frequency shows to be open the entire inter-frame period, which is 1/180th of a second. (In fact the shutter is closing during that period but no frames are obtained to show this.) This gives a calculated speed at 1/2 the Nyquist frequency. In my case about 1/100th of a second.

 

[Donald Qualls]

Not sure which generation phone I have, it was free from work. A shutter faster than the Nyquist frequency shows to be open the entire inter-frame period, which is 1/180th of a second. (In fact the shutter is closing during that period but no frames are obtained to show this.) This gives a calculated speed at 1/2 the Nyquist frequency. In my case about 1/100th of a second.

With multiple samples, however, you can determine the fraction that show open one frame, partially open in one or two frames, or not open at all (i.e. cycled completely in the inter-frame period) and from that get at least an estimate of somewhat faster speeds. The brighter the light source, the better this will work (if the light is bright enough, the exposure can be much shorter than the frame rate).

 

[ic-racer]

With my phone the video of the shutter shows the interrogated shutter to be either open or closed in all samples. Maybe faster effective shutter speeds some newer phones or different test conditions would improve the situation.

 

[ic-racer]

With multiple samples, however, you can determine the fraction that show open one frame, partially open in one or two frames, or not open at all (i.e. cycled completely in the inter-frame period) and from that get at least an estimate of somewhat faster speeds. The brighter the light source, the better this will work (if the light is bright enough, the exposure can be much shorter than the frame rate).

Do you have some images of partially open shutters that might allow calculation of estimated open or closing instant. I have read that some iDevices can operate at shutter speeds as high as 1/40,000 of a second. But mine does not.

 

[ic-racer]

Donald, I'm glad you chimed in, because this thread is a spin-off of the 16mm camera thread. I'm working on some of my 16mm cameras and testing the shutters. I was going to give up on this iPhone shutter tester thing, but maybe you can get some good images of the shutter as it is moving across.
If the aperture were infinitely small, the shutter efficiency would be 100% and the midpoint of the shutter blade would be the opening and closing point for each blade. So that would make a good timing point that could be estimated if the blades could be caught in motion.

 

[Donald Qualls]

I haven't actually done these tests with my Pixel -- though I guess I should. The iPhone is supposed to have a very good camera -- and it's the one the Pixel was built to beat -- but my Pixel is three and a half years old (planning to upgrade when the 4a comes out).

Obviously, you'd want to make such tests with the aperture wide open, to give the best ability to spot partially open shutter leaves -- and this might not work at all. I don't have a very precise idea just how fast a leaf shutter opens and closes, other than that it can't take much more than a millisecond to give a speed of 1/400 or 1/500, never mind the (mostly hype) 1/1000 on the Super Speed Graphic 1000.

I don't have time tonight -- it's after 8 and I have a 4 AM alarm -- but I'll try to remember to find an unloaded Minolta 16 or Kiev 30 or 303 (externally visible shutter) and run a few slo-mo tests tomorrow after work.