Recording distant audio NA

[RyanShort1]

Ok, so if I wanted to record gunshots, say, 2-3 blocks away and have a real-time livestream of it, what equipment would I need? If three to four neighbors wanted to participate, could we triangulate them?

 

[flyingron]

If you had a fairly accurate, synchronized timing between the three reception sites, you could. You'd need to make sure you're measuring the initial shot and not any echos and this is all premised on the fact that the shooter was stationary. Comparing the relative times shows you how much closer it is to listener A vs B, or B vs. C, or A vs. C. Most gunshot timers are good to 10 milliseconds, so that's about 11 feet of accuracy. Assuming worst case, you'd be off 22 feet in each direction.

 

[RyanShort1]

So what kind of equipment, how to sync, stream?

 

[Initial Fix]

I’m curious why? but typing out loud, you need at least 3 microphones with known locations (similar to how gps works), each microphone synced in time and connected to a DSP that identifies the intrinsic sound spikes and correlates those to a timestamp. After that it’s just math to figure the time difference of arrival and the “location” and error ellipsoid. As a college professor used to say, I’ll leave the math to you

Also try searching on radio direction finding, or “fox hunting” as the amateur radio folks call it. The principals are similar.

 

[RyanShort1]

I’m curious why?

Nightly semi-auto fire in the middle of a big city??? Multiple 911 calls by multiple parties and nothing can be done because no one wants to talk. And yeah, I've taken a bullet to the hood of my truck already. Came almost straight down, looks like.

 

[Kenny Phillips]

Come to Cincinnati; they have such a setup installed already. It's a commercial product.

 

[asicer]

Come to Cincinnati; they have such a setup installed already. It's a commercial product.

Is it www.shotspotter.com ?

 

[Spring Ford]

I think this could perhaps be done with some PCs (or any stereo digital recorder) recording audio. Two might do it but three would be better.

You need two microphones wired up to the stereo input of the PC. Place the microphones carefully at surveyed locations that you can put on a map.

Record the audio.

Using Audacity Software which is free (I have NO experience of this software) you will I think be able to obtain the difference in arrival times of the sounds on the two channels. From the speed of sound (record the temperature too) you will be able to get a position line from the pair of microphones.

Use a second PC and microphones at a separate location and you will get two position lines and a position.

With three at the approximate corners of a sufficiently large equilateral triangle you would get a very good position in all cases.

The microphones would not need to be too far apart but I would put them a few feet apart or as much as I could easily manage. The digitised sound has a sample rate of maybe more than 50kHz. I think that this will provide a VERY accurate position line if the microphones are even a few feet apart.

Note:- No absolute times are necessary with this method other than to ensure that you have the right gunshots on the separate PCs.

I have NO idea if this will actually work but you could get the software and try it with any old microphone to see if you can actually identify the gunshots on the Audacity level graph and how precisely they can be identified. There will I am sure be many echos and other confounders.

You also need to check if you can zoom in sufficiently with audacity to measure time differences of small fractions of a millisecond - my guess is you can

The further apart the microphones the better the position line.

This may not work at all, I don't know but it is pretty cheap to try out.

If you proceed someone here may be able to do the math (maybe me for the position line if that is not your forte.

I would synchronise the time on the PCs of course to make it easy to identify the particular noises you want to look at.

Very likely someone will be along soon to shoot me down;-)

Good luck.

 

[Spring Ford]

You also need to check if you can zoom in sufficiently with audacity to measure time differences of small fractions of a millisecond - my guess is you can

https://manual.audacityteam.org/man/zooming.html

There you go go go.

"or even a tiny fraction of a second, where you can see the individual audio samples as small dots:

"

 

[RyanShort1]

Hmmm. I've got Audacity and used it for entirely different purposes before. I'm assuming I'd need to think about the kind of mic that would be relatively durable and weatherproof, but also decent enough quality.

 

[DaleB]

@Spring Ford just described a job the Army trained me for in the late 1970s, using geophones tuned for artillery or mortar fire and a 1950s or early 60s vintage computer (bonus points for anyone who remembers the FADAC).

I’d suggest putting the microphones as far apart as possible, and of course you’ll want the array oriented broadside to your suspected point of origin. A triangular or rectangular array would be more omnidirectional.

 

[Spring Ford]

@Spring Ford just described a job the Army trained me for in the late 1970s
...
I’d suggest putting the microphones as far apart as possible, and of course you’ll want the array oriented broadside to your suspected point of origin. A triangular or rectangular array would be more omnidirectional.

I just made it up. I was thinking about how to get accurate time sync across sites and all solutions looked too expensive. Suddenly noticed that it might not be necessary.

Agree with as far apart as poss. Sample rate will likely be 44 or 48kHz so need at least a few feet I would think.

Thing is a PC or other cheap recorder is Stereo and so no more channels are available without exotic (expensive) hardware.

 

[DaleB]

I just made it up. I was thinking about how to get accurate time sync across sites and all solutions looked too expensive. Suddenly noticed that it might not be necessary.

Agree with as far apart as poss. Sample rate will likely be 44 or 48kHz so need at least a few feet I would think.

Thing is a PC or other cheap recorder is Stereo and so no more channels are available without exotic (expensive) hardware.

If I were setting something like this up, I’d use Raspberry Pi’s syncing time using either NTP (free, good time sync) or low cost GPS (cheap, great time sync).

 

[neon]

You have to use a "parabolic microphone". More information about it can be found on the net.

 

[FORANE]

Nightly semi-auto fire in the middle of a big city??? Multiple 911 calls by multiple parties and nothing can be done because no one wants to talk. And yeah, I've taken a bullet to the hood of my truck already. Came almost straight down, looks like.

Maybe consider moving out of the hood?

 

[RyanShort1]

Maybe consider moving out of the hood?

Not an option right now.