Of course it is only an approximation. But in most cases, it's a quite good one. I compared the values from this app with an optical measurement (photoresistor + audacity) and till the 1/250 the results are quite accurate, the deviation is not more than 1/3 f-stops. Of course this also depends on the camera you are using.
This app isn't designed to measure the shutter speeds at microsecond-level. You won't be able to tell if the 1/250 is more likely a 1/220, but this isn't really necessary as you don't see the difference in your final image anyway. What you can detect, are larger deviations, wich really affect your image. For example if your 1/125 is rather a 1/80.
I had the same idea with the light-sensor, but unfortunately you can't read this sensor without the use of private APIs (forbidden commands). What you could do, is attaching an phototransistor to the mic-input of the iPhone, but I haven't tested this yet.
So the people who wrote the lighmeter app could get access to the private APIs? 1/3 of a stop is too large an error for the measurement instrument while it's kind of acceptable if the shutter is within 1/3 stop. Using a photocell is the same as people have been doing uisng sound recorder app on the pc. I personally use an opposed sensing fiber optic, modulated red light sensor and make the measurement with an oscilloscope. The sensor has a 50micro second response time so it's reasonably accurate up to 1/2000.
All those lightmeter-apps use the built in camera for the measurement. The light-sensor is on the front of the iPhone, right next to the speaker.
Of course you can measure a camera without my app using your computer and a microphone/photocell. But with the app you can do it a lot easier and faster, it displays the time already in parts of a second, the deviation in thirds of an f-stop and you can save all this data in a neat table, allowing you to see all values at a look without calculating or writing down anything. And the greatest advantage: You can put all that in your pocket and check a camera in less than 5 minutes, wheather you are at home, in a camera store or on a flee market
“Photography is a complex and fluid medium, and its many factors are not applied in simple sequence. Rather, the process may be likened to the art of the juggler in keeping many balls in the air at one time!”
Ansel Adams, from the introduction to The Negative - The New Ansel Adams Photography Series / Book 2
I downloaded it before work and quickly gave it a try (winged it!) with my Contax G2 without reading your website. Not extremely intutive on the fly but I think after reading your website I can figure it out. Will try again tonight.
"Well, my photos are actually much better than they look..."
May I suggest to specify in the web site that the tool is useful for all speed that can be used with a normal electronic flash. That means up to 1/60 with some cameras, or 1/90, 1/125, 1/250 with some other cameras. 1/30 with some Zenit cameras IIRC. It can give no indication for speeds that cannot be used with a normal electronic flash.
Leaf shutters (i.e. central shutters) can be used with flash on all their speed range (normally never faster than 1/500).
Couldn't you use the default camera instead of the light sensor in the front? Say have it record in movie mode, at a very low resolution only recording the presence of light and not an image, and measure the shutter speed using a flash light or other bright light source. Then pretty much use the same analysis, the first peak, is when it open, and when it again records no light is the end of the shutter duration.
I tried it last night with an old Wollensak shutter on a miniature Busch Pressman and it works very well - can I suggest that in your comparative speeds table, you include speeds for older shutters like 1/10th, 1/25th, 1/50th etc. as well as the modern ones? I am far more likely to be suspicious of older shutter speeds and want to test/keep a note of them...
The frame rate of the camera is much too low to make such timing measurements. Audio is the best approach for decent timing on a phone unless you want external hardware like a photosensor on the microphone... in which case this app is just as applicable. Since the app is flexible enough to allow you to select endpoints, you could make a more-accurate timer by using a photosensor to modulate a 15kHz tone and feed it into the phone's mic-input.