Improving sound detection

Starting SprintTimer with sound is easy and accurate. But the start is often so far away that sound is too weak to trigger the start. In this post, I will give some tips on how to improve the sound detection. This could be used both when the sound should directly trigger the start, and when using Hand/Mic to create a correction.

Sound intensity decreases with the square of the distance. This means that if you double the distance the volume will only be 1/4. And at 100 m it will be 0.01% of the volume at 1 m. The problem, however, is usually not the weak sound as such, but rather that other sounds become relatively stronger. The risk for a false trigger therefore increases. So one has to increase the start sound in relation to all other irrelevant sounds Or in technical terms, increase the signal/noise ratio. One obvious way to do this is to bring the microphone closer to the sound source. Making a long microphone cable or using Start Sender are two ways to achieve this.

Another solution is to try to focus the sound reception to the gun and shield other sources. This can be done with a parabolic microphone. This is often used to record faint sounds, for example when filming wildlife. While a professional parabolic microphone is pretty expensive, it is fairly easy to make one yourself. More of this below.

Sound speed
Sound moves at 343.3 m/s through the air, which means that it takes about 0.3 s to travel 100 m. This is too much to neglect and is the reason why there is a ”Sound distance” parameter in the Start set up in SprintTimer. If you type in 100 in that text field the app will add 100/343.3 ≈ 0.291 s to the measured time. The number 343.3 is the speed of sound at 20° C. The speed increases with roughly 1% for every 5°. If you want, you can compensate for this by increasing the distance with the same amount. However, 1% of 100 m is 1 m, so it is usually more important to get the distance correct. And 1 % of 0.3 s is 3 ms, so the 1 m error is pretty small.

It is also important to note that the sound distance is measured from the sound source to the microphone. Any distance covered by a cable or a network should not be included. 

Making a parabolic microphone
A parabola is a curve that has the nice property that all incoming parallel rays (sound, light, etc) are reflected to one point, the focus.

This is used in everything from the headlights of your car to gigantic radio telescopes that listens to quasars billions of lightyears away. The equation for a parabola is pretty simple: y = x^2/4f were f is the point on the y-axis where the focus is located. If you want to play around with the curve you can go to this page and drag the focus point up and down.

A parabolic microphone is a parabolic reflector made of a hard material, and a normal microphone, e.g. the handsfree from your iPhone, placed at the focus point. You don’t need a perfect shape to improve the sound reception. So it is pretty easy to find ready-made objects that can be converted to a microphone. You need a bowl-shaped object 30 – 60 cm in diameter made of hard plastic or metal. For example:

  • Lampshade
  • Salad bowl
  • Casserole lid
  • Wookpan
  • Satellite dish

Some of these already have a hole in the middle (like the lampshade in the image above) so they might be restored if needed. Others like a salad bowl might require that you drill a hole in the bottom or find other means to fix the microphone. Be creative and you can probably find something at home that can be used. Some additional inspiration can be found here.

Testing
To test the theory I made parabolic reflectors out of two metallic lampshades, one fairly flat (picture above), the other more like a half-sphere. I placed a computer playing a shot 10 m away and another playing a pure noise closer by. I set the start in SprintTimer to Hand/Mic to get a sound graph and started at the sound of the shot.

I first tested with the internal microphone on the iPhone. When it was held upright in normal portrait mode the shot didn’t show up in the graph at all. When I held flat with the bottom pointed towards the sound there was a small indication of the sound.

Click to show larger

I then used a Røde lapel microphone and got the graphs above. The first curve shows just the mic pointing towards the sound. The next two are with the mic in the focus of the flat and spheric reflector respectively. The latter two clearly give a better distinction between noise and shot and the difference between them is small.

I also tested with the handsfree with and without a reflector and the result was similar to the one above. While not a dramatic effect, it might mean the difference between being able to use the sound to trigger the start or not.

Finding the focus
To find the focus point you can use the formula for the parabola given in the figure above. Measure the depth d and the radius r of your reflector and insert that into the formula. You will then get f = r^2/4d. For example, the lampshade in the image has a diameter of 40 cm and a depth of 5 cm. The focus distance then becomes f = 20^2/4*5 = 400/20 = 20 cm. If you like, you can then fine-tune the placement of the microphone by playing a constant sound, move the microphone up and down and listen where the volume reaches its max.