ANALYSIS OF SOURCE FOUR SIGNATURES/ECHOES
@cmartenson, I took a look at the Source 4 audio, where your spectragram showed visible differences in frequency and echoes between the first group of 3 shots and the second group of 5 shots. There is a lot of talking/screaming and shuffling going on, which I have tried to suppress so that we can see the shot signatures. Most of this background noise is luckily below 200 Hz and above 800 Hz, and there is significant energy for the gun signature between 250 Hz and 550 Hz (there is energy elsewhere, but most of the energy is here).
I applied a mild high-pass filter with a 250 Hz corner frequency. I didn’t want to make this a strong filter because the noise below this frequency isn’t severe and I didn’t want to perturb the shot waveforms.
There is a LOT more noise at high frequencies, and by placing the corner frequency at 800 Hz, there shouldn’t be significant distortions in the range of interest. I applied a fairly strong 24 dB/decade low-pass filter to cut this noise.
Here is one example for the region around a shot, before and after filtering. There is still some noise, but it isn’t significantly masking the shot and echo signatures that I wanted to look at.
I cut out and aligned all eight shots based on the large, first negative pressure waves (mainly because it is hard to precisely place the preceding pressure peaks on shots 1-3). I aligned them top to bottom in the figure below, which I have annotated and marked with colored boxes.
SHOTS 1-3: These shots sound muffled for several reasons (the darker red overlay boxes are the “meat” of the primary response). They each attenuate to a fairly low level within 10 ms. But interestingly, each shot is followed by an almost immediate echo (what I would consider multipath off of a fairly close surface) within 15 to 18 ms. This is only a 20-foot increase in path distance to the microphone, which means something fairly close to the gun barrel (10-15 ft?) reflected a large portion of the initial gun sound, and this is likely not an “echo” per se of the energy in the first pulse, just a different directional part of the gun blast that found its way out of, shall we say, an enclosure???
Continuing, we can clearly see Echo 2, which you noted in your video. This echo occurs at around 80 to 90 ms after the initial blast, so the sound bounced and traveled an additional 100 feet or so to reach the mic.
The other thing I find interesting is there is a faint Echo 3 sitting at around 155-160 ms from the initial blast. This is very close to the same timing of the strong echo in the last five shots, which given that not as much sound energy is propagating out in the first three shots, could be the same echo surface (but with shots 1-3 taken slightly closer to the reflector). The other squiggles you see in these waveforms are by and large residual human noise at higher frequencies.
SHOTS 4-8: These shots sound “brighter” when listened to in the recording. In general, there is a larger, longer primary response (over 20 ms), followed by a reverb (clearest in shots 5 and 6) for 100 ms. I would guess this to be consistent with a gun barrel discharged close to a large metallic plane, with the mic in a good position to capture ringing of the surface. I could believe that Crooks’ gun next to a metal roof was the source of these five shots.
I could not see evidence of a prominent multipath reflection right after the shots, and I could not see evidence of a reflection agreeing with the strong Echo 2 from the first shots.
There is a prominent echo visible in each of the last five shots (for shot 7 we can only see the start of the echo as shot 8 then masks it) that is centered about 165 ms after the initial blast, so a reflection path that is about 190 ft longer than the direct path to the mic.
Please make your own conclusions and explanations. Here are mine:
- Shots 1-3 appear taken from the same gun A and same position A.
- Shots 4-8 appear taken from the same gun B and same position B.
- These two positions A and B are different in significant ways, which also says there are two guns A and B given two seconds between the shot groups.
- Position A appears to be an enclosure that provides a strong multipath signature and significant attenuation of the signal.
- Position B appears to be an open position that does not attenuate the initial blast and excites a nearby open-air surface that rings for a significant duration, as would metal.
- A strong Echo 2 appears in the first three shots that does not appear in the second five shots. One explanation could be a strong bounce off of the internal end of a large enclosure (we can just call it “Building 6” for grins) from which the three shots were taken.
- A strong Echo 1 appears in the last five shots. It could potentially match with a weaker Echo 3 seen in the first three shots, but with a slightly shorter path distance for the first three shots. This could be because the mic repositioned slightly between the groups, the sources aren’t quite the same distance, or some combination of the two.
- I may laugh out loud if you say this is because the first three bullets were different or Crooks changed his stance slightly.
- I may also laugh out loud if you say that because a microphone is in a car all of these signatures from a different direction would look like Octamom had them inside that car.
to attempt it (even Jason Bourne).
