Calculating FPS by sound

[Pendra]

I got the idea yesterday that it must be possible to measure the FPS by using the sound of the gun and the sound of the hit. I set up a little experiment to check the possibilities:

Got a mic and installed a demo version of Sound Forge (but I think any other Wave recording software will do)

Put a wodden plate (it has a sharp sound when the BB hits) to the wall

Put down a mark at exactly 1 m far from the wodden plate

Put on the safety googles as the BBs ricochet wildly

Started the recoding

Put the end of the barrel to the mark and fired off some rounds

 

Here are my findings

1st I used my M16 (EG1000, Alu piston head, 400 FPS, but not going to tell you how I achieved that FPS)

The recoding showed 2 major groups (with 2 major peaks) very close to each other, followed by a smaller group then the group of the hit.

I'm sure one of the first 2 peaks is the BB leaving the barrel. I think the 1st peak was the piston hitting the gearbox, and I think the 2nd peak was the BB leaving the muzzle. I measured the time between the 2nd group and the hit group by selecting the region between the peak amplitudo of the 2nd group and the first high amplitudo in the hit group. I used the following calculation 1m/0.0085s=117 m/s that is around 385 FPS. The other shots gave me the same results.

 

Then I used my L85 (EG700, Alu piston head, 330 FPS)

The recording was not that clear, since it lacks the semi auto. I could barely see 2 groups next to each other, but no really high peaks. Also the 3rd small group was missing. Now I took the time between the middle of the 2nd group and the first high peak of the hit and it gave me 0.0105. I calculated 1m/0.0105s=95 m/s and it is around 312 FPS.

 

The problem with this is that you have to clearly identify the time when the BB leaves the barrel. A small error and the calculation will be off. I think it may bo good idea using 2m instead of 1m, beacuse in this case a small error would be still fine. I tried it and it had about the same results.

It seems all right, but that are only 2 guns. If anyone could make some more experiment that would be cool. If this thing really works, then it could be a cheap (and less accurate) alternative for the Chrono.

 

EDITED some messy senteces.

[Tommygunn]

Have you tested the guns with a proper chrono as an accurate comparison?

[Pendra]

Sure I tested them about 4 months ago. So I actually knew what to look for. That is why some more experiments are needed. So someone has to do a test using a gun with unknown FPS. Calculate the FPS with the method above (time between the 2nd sound group and the hit group) then use a Chrono to check if the results are valid or not.

[McMadkat]

Don't forget the larger the distance with between the gun and board the slower the BB will be going.

[Hissing_Sid]

I would have thought something like 10m and 2 microphones would be the minimum you'd need.

 

Bearing in mind the BB is doing 100m/sec then you're only recording 0.01 of a seconds worth of travel. To achieve +/- 1 fps accuracy you'd need to be accurate to within 0.003 of a second.

Relative to this, bear in mind that the speed of sound is approximately 300m/sec which mean that the time taken for sound to travel back to the microphone from 1m away would be enough to destroy your 1fps accuracy.

 

Also, last time I checked, a chrono was cheaper and easier to set up than a laptop.

 

Nice idea for an experiment or for those who already have the bits and no desire to spend £30 on a chrono but not really practical.

[Tommygunn]

Why does somebody have to go and do the test with a gun of unknown fps? Why don't you just tell us if your computer fps result was the same as your chrono result?

[Hissing_Sid]

Here are my findings

1st I used my M16 (EG1000, Alu piston head, 400 FPS)

I used the following calculation 1m/0.0085s=117 m/s that is around 385 FPS.

 

Then I used my L85 (EG700, Alu piston head, 330 FPS)

I calculated 1m/0.0105s=95 m/s and it is around 312 FPS.

<{POST_SNAPBACK}>

 

 

Why don't you just tell us if your computer fps result was the same as your chrono result?

<{POST_SNAPBACK}>

Erm...

 

He already did.

In both cases it appears the reading is around 4% low. As long as this remains nearly constant it's probably better than a bloody coke can, at least.

[Tommygunn]

Oh yeah. I'll look harder next time.

[Pendra]

Using a Laptop is a good idea to make it portable. I think I will make a solid, transportable frame and check some guns on the next skirmish. In the experiments I put the Mic half way between the plate and the barrel. In the next one I will put the mic to the barrel of the gun so the highest peak will be the BB leaving the barrel.

I'm not aiming for accuracy to 1 digit, I'd like to get a good estimate. Like the gun is around 280, 290, 300... FPS.

[Hissing_Sid]

Right...

 

Just had a go at this myself.

 

I used a laptop running Windows Sound Recorder for recording and Audacity waveform editing software to examine the results.

I measured out a 2m distance to target and used a biscuit tin as a target.

The gun I used was a stock Marui P90.

 

The end results show that the gun seems to produce 2 distinct peaks during the firing cycle.

I recorded the time-lapse between both peaks and the sound of the target hit cos I figured one or the other would be right.

The exact time the BB hits the biscuit tin is MUCH more easy to see. No problems there at all.

 

A couple of thoughts...

 

First, the noise of an AEG shooting isn't that sharp. I put the microphone half way between the gun and the target. It would have been better closer to the muzzle, especially if you're using a loud target such as a biscuit tin.

Second, I did the test using an AEG. If you use a GBB there might be more mechanical noise which might mess up the results.

I imagine that, using a system like this, experience will be what tells a person which part of the sound shows the BB leaving the barrel.

 

The results...

 

I fired 10 shots and measured them.

As you can see from the waveform, the first shot was actually a dud. The first shot usually is from a P90.

 

I then zoomed in on all the sounds individually and found that the time between the muzzle report and the strike was about 0.025 seconds. I highlighted the time between launch and hit and you can see the elapsed time at the bottom of the screen.

Incidentally, there's always an increase in noise in the middle of the flight time. I wonder if that's the BB passing the mic?

 

I repeated this with all 10 shots and the results are surprisingly consistant.

 

So, 1 / 0.025 = 40.

40 x 2 = 80 (x2 cos the distance to target was 2m)

So, we're gonna guess at a muzzle velocity of 80m/sec

 

Now, I can only promise that I DID NOT chrono the P90 until afterwards cos I didn't want that information interfering with my judgement. I did chrono it afterwards though.

 

Throw away your coke cans.

[Skorn]

the only problem is that some guns sound different (but if u are measuring the time difference, sorry that this is all wrong), like the ICS sounds "wierd" or so i have heard

[bullzebub]

"weird"? well ... the sound you look for is the "last" one done by the gun (motor starts - pulls piston back - relesases piston - piston strikes cylinderhead - bb leaves barrel (if you have wrong cylinder size this may change)) so the "weirdness" shouldnt matter ...

:-)

[Pendra]

Yeah if you put the mic to the end of barrel, then you can be sure that the loudest sound is the BB leaving the barrel. It doesn't matter what kind of weird noises the gun makes.

[[H&K]Ravizzone]

Nice, that is pretty awesome.

[systema]

Very nice Sid. So basically one "twap" (to the human ear on with .wav file) is actually two; the sound of the aeg and the sound of the impact of the tin can?

[Hissing_Sid]

Very nice Sid.  So basically one "twap" (to the human ear on with .wav file) is actually two; the sound of the aeg and the sound of the impact of the tin can?

<{POST_SNAPBACK}>

Yep.

 

That wav file has the sounds you see in the pics. If you download Audacity for yourself you can muck about with it for yourself.

[Kukailimoku]

The difference in calculations made could be due to the very small amount of time it takes the gearbox to cycle and fire another round?

 

Or I could be just talking out of my *albatross* here, I'm not too sure.

[bullzebub]

no ... since you are looking for the sound when the BB leaves the barrel ... :-)

[Thanatos]

You are not measuring the time when the BB leaves the barrel it is from the intial strike of the piston and the BB. The BB has to travel down the barrel, so your calculations would be off, and the longer the barrel you use the worse the calculations would be measuring from the time the BB leaves the barrel.

 

You just have to measure the change in time between the hit of the piston head on the bb and the hit of the bb on the target. This will give you the time of travel for the bb from point A to point B.

Vavg = D1 - D2/ T1- T2

The difference of D1 and D2 is the distance between gun and target. The difference in T's is the invertval of your two noise mentioned above.

 

The only problem with this horribly simple chorno is that you have to be able to distuinsh between gearbox noise and the hit of the piston. Though the hit should measure as a sharp peak, and the diffence between the two sounds time is minimal.

 

How did nobody think of this before, it is so simply.

[Hissing_Sid]

Well, as I just said in the other thread on roughly the same subject...

 

My recordings definately always show 2 peaks during the firing cycle. I'm guessing that one is the piston hitting the cylinder head and the other is the muzzle crack as the BB leaves the barrel.

 

Don't forget that, even at such low power, you have compressed air behind the BB and, once the BB leaves the barrel, that air will decompress and, in doing so, probably accelerate to supersonic velocity. Ergo, you will get a muzzle crack as the BB leaves the barrel. And, of course, we can confirm this theory cos, ummm, we do get a muzzle crack.

[Pendra]

Plus I think the small peak in the middle is indeed the BB passing the mic. Thou I have no idea why it shows like this. In my tests I also had this small peak and it was in the middle of the 2 peaks, and I put the mic to the middle if the test range. This needs some more tests thou, like placing the mic to different places along the range should move the middle peak accordingly.

If it works, then the calculation could be made like this: Place the mic 2m away from the target, fire the gun from like 3 m. Te recording will register the gearbox peaks, then a silent period, then the smaller middle peak (BB leaves the mic), then an other silent period, then the hit. By measuring the distance between the middle peak and the hit would give good results as well. And it would be completely clear from any gearbox noises.

Also it would be possible to check how much the BB slows down in specific distances. For example you fire the BB from 12 m and place the mic at exactly 2 m from the target. When the BB leaves the 10m mark it leaves a small peak on the recording. From that you can calculate the speed and the power loss.

In any way, I think if you put the mic at the barrel of the gun the highest peak will be when the BB leaving the air and the air decompressing.

[systema]

I'd say the best way to measure the sound on the gun's end is to put the mic right next to the muzzle. The sound to measure from and to the tin can would be from the loudest, and that should be right when the bb exits the barrel and goes over the mic.

[snorkelman]

to get a GBB profile best bet would be a first 'profile shot' of the GBB being shot at metal target at very close range (like a CM or two)?

 

eg as you can spot the sharp peak where the strike is pretty easily, you can then deduce that any noise following on after that is the action cycling, and that BB exiting the barrel point is going to be the the part of the waveform just a few ms prior to the strike (given proximity of barrel to target)

 

when you come to do the test proper at your choice of range you now know what point to consider the muzzle exit point in the waveform.

 

stevie

[Pendra]

I made a little application which, in theory, should ease the FPS calculation. I calibrated it to Sid's Wav file, but you can recalibrate it to your wav.

You can set the sample rate, the distance and the weight of the BB. You need to put the mic (or the wave/mp3 recorder) to the end of the barrel. And use something that makes a lound bang at the impact. It would also be good if there was no other sound at the recording, but if you make 10-15 semi shots, then it would be able to fine.

The most important is the Barrel leave treshold setting. That filters out background noises. The impact noise should be equal or higher than the BB leave noise. The BB leave noise should be higher than the background noise, otherwise the calculation will off. Try to reduce it from 0.9 with a stepping of 0.1. You will get no results for a while, then, at one point, you will have a result. It should be almost accurate. At that point use a stepping of 0.01. The next results should be more accurate. If you get the same results with 2 or 3 different treshold setting, then it should be a good result. If you lower it too much (to the background noise level) then the results will go off again. There are buttons to increase or decrese the treshold, so you don't have to type the values in again and again.

Faulty input datas will yield wrong, or no results at all!

Only wav files can be opened!

You can DL it from here:

www.variagame.hu/SCrono5.exe

[Pendra]

http://www.variagame.hu/SCrono5.exe

Sorry but it seems, the edit button doesn't work for some reason.