Using a microphone to test your shutter speed

I shoot a lot with my vintage cameras. Many date from the 1950s and have probably never been serviced. I’ve no idea how accurate the shutter speeds on them are. For the mostpart, they are probably acceptably accurate, since they seem to produce reasonably well-exposed negatives.

Except for my newest lens, a Horseman 65mm f/5.6 view camera lens. The negatives seemed to be consistently overexposed.

I wondered how I might test the shutter speeds. I don’t have a high speed video camera, but I do have a set of nice microphones and audio equipment. I set up a condenser microphone to record digitally, and I decided to sample at 192kHz to get as many data points as possible. For people with more normal sound hardware, any microphone should be OK, and the standard 44.1kHz will be fine.

Then it was a simple matter of speaking to say which shutter speed I was testing, and recording the sounds of the shutters at all speeds. By ear, you can’t really distinguish the separate clicks until you get to around 1/15. This sound clip is my test of the Horseman 65mm, which has a leaf shutter in the lens.

The tricky bit is picking through the recordings in a graphical sound editor. These three screenshots show relatively straightforward readouts from the sound file. Others were extremely hard or impossible to make out.

Assuming these readings are accurate, it would explain why the negatives I exposed with a 1″ shutter speed are overexposed!

However, I ran into a few snags with this experiment:

• Cameras make other noises than just the sound of the shutter. The sounds of actuation levers, clockwork and springs can mask the click of a shutter quite easily, especially at fast shutter speeds.
• Focal plane shutters in particular make more than two clicks (because they have two curtains, and both curtains start and stop). It’s hard to know which two clicks to measure the time distance between.
• There doesn’t seem to be a systematic error. The same shutter can be too fast and too slow at different speeds. This makes it really hard to use the data to “calibrate” itself and decide which of the peaks are the right ones.

Maybe this is a potentially useful technique, but I reckon the only way to be sure about your shutter speed is to video it with a high speed video camera, and inspect the footage.