# Getting the bearing on a sound.

I don’t know if this is the wrong place to post, or even if it is the kind of idea that audacity people are into, but it seems to me that the tool is most of the way towards what I need, I just need to work out the details.

What I want to do is to get a bearing on a sound. I am thinking that if I get a synchronised sound recording from two places then we can see the time separation and from that, get a measure of the direction that the sound is coming from…

Which we could do simply by separating the left and right microphones and comparing the traces…

I was looking for a way to synchronise the waveforms recorded from separate computers, but I’ve just realised that we probably don’t need to.
But I’ve written this now, I might as well post it: somebody might have interesting ideas from it.

You could record two microphones as a stereo track. That could be done with a computer, two microphones and a conventional sound card / audio device.

Some problems that you are likely to encounter are:

1. Timing differences between the microphones will be very small. If the two mics are 30 cm apart, then the maximum difference will be less than 1 ms.

2. Even very abrupt sounds (like a hand clap) have significant “attack” time (the sound does not start “instantly”).

3. Unless you have an anechoic chamber to perform your experiments in, the recorded sound will be “polluted” by reflections of the sound from physical objects (such as walls / floor / ceiling / furniture / …)

That was my thinking, when I eventually got there.

Some problems that you are likely to encounter are:

1. Timing differences between the microphones will be very small. If the two mics are 30 cm apart, then the > maximum > difference will be less than 1 ms.

True enough, but 30cm isn’t very far, and while a millisecond doesn’t sound like a long time, it is something like 3 million clock cycles of this computer, and this is not a particularly fast one.

1. Even very abrupt sounds (like a hand clap) have significant “attack” time (the sound does not start “instantly”).

2. Unless you have an > anechoic chamber > to perform your experiments in, the recorded sound will be “polluted” by reflections of the sound from physical objects (such as walls / floor / ceiling / furniture / …)

I don’t know if it needs to be particularly clean, just as long as there are points on the two traces that are recognisably the same…

A goniometer looks a bit like what you’re describing …

https://forum.audacityteam.org/t/free-stereo-visualizer-vst-from-izotrope/47033/1

Clock cycles are not really relevant compared to sample rate and audio bandwidth. Most affordable sound cards have a maximum sample rate less than 200 kHz, so the minimum measurable unit of time is around 5 Microseconds. What you will be looking for are small phase shifts in the recorded signals.

Yeah, but a millisecond is 200 x 5μs, you were predicting 1ms phase difference with a separation of only a foot, so with 20’ separation and a resolution of 5μs we should be well in.

Yes it does. Interesting, thank you…