Two-Shooter Audio Analysis
In my last post, I dismissed the Cruiser dashcam audio because it recorded the primary response of all eight shot reports essentially identically, even to the point of having echoes from the preceding shots disappeared and “written over.” To further back that up, I will show responses from other, more reliable audio sources so you can see how different the eight shot waveforms truly are. These sources were not totally stationary, BUT, they do appear relatively stationary during the last five shots, and the L/R stereo microphones seems to have extremely similar responses in each case, which says there was probably no “thumb” recorded and these are good recordings. Further, there are large acoustic signature changes between shot group 1 (1-3) and shot group 2 (4-8), in all the recordings that didn’t spend a month in police custody.
For the tl;dr crowd, the resonant spectral peaks of reports 1-3 differ substantially from the resonant spectral peaks of shots 4-8, as well as the spectral distribution generally. By my measurements of first pressure wave arriving from the boom (which I believe to be very accurate), the first shooter was somewhere generally along the shot line of the patsy, but a different gun and a different location.
For those interested in details, first, let’s look at the timing results, and then we can see how we got there. I calculated differential arrival times from the podium arrival times of the booms for source 2 (got a gun), source 3 (TMZ), and source 6 (DStew). I time-aligned these audio sources at shot 4. Why? I had to choose somewhere, and this was a clean choice, and we believe it was likely from Shooter 2 (Patsy) so we know its location. The purple dots in the plot are the podium reference. If all 8 shots came from the same location, and the sources never moved, all we would see are the purple dots as all the sources would have the same time delays and their dots would be occluded by purple dots.
This is almost exactly true for shots 4-8. All sources agree almost perfectly to less than a foot (with two one-foot variations) for these five shots.
For shots 1-3, NONE of the sources agree on the report spacings with the podium. Variations of up to almost 20 feet exist in arrival times (at the speed of sound) from what would be expected from the podium report spacings, if the patsy took all the shots. Note this doesn’t say where the shots were taken from, it just notes the discrepancies.
This COULD be due to the patsy taking the first three shots also, if sources 2 and 6 moved away from the shooter from the time of shot 1 to the time of shot 4, then stopped moving, and source 3 was slightly attracted towards the shooter. But this 18-19 foot delta pattern doesn’t seem that way to me. DStew was moving mostly sideways and not really changing his distance away from Crooks that much. Source 2 couldn’t really move away directly from Crooks, a fence was in the way. Sure, part of this could be them moving some, but not 20 feet. I think the times are reliable, but the microphones had feet until shot 4, so that could be part of the answer.
So now, let’s look at the source 3 TMZ audio to see if we can convince ourselves we have a single firearm. I have cut out and aligned all eight shots in the figure below. I have also noted to the right of each the peak “ring” frequency of the first 15 ms or so of each shot. The first three shots ring at 438-440 Hz, with quite a lot of energy at frequencies lower than that. The last five shots ring at a peak of 485-519 Hz, and have much less energy at lower frequencies. Also look at the red and green vertical lines and note how they hit the first two pressure peaks of the reports for the first three shots, but the second peak is well inside the green line for the last five shots (meaning a higher frequency, probably a shorter gun barrel and/or a different close resonant surface). Lastly, I will note that shot 3 is hard to pick out visually here as there is a high-pitched scream competing with it in the waveform—spectrally they are quite separable.
Anyway, I say this signature means: 1) two different guns (different resonances); 2) two different distances (different spacings from the podium); 3) demonstrates that the dash cam audio with no audio variability is BS.
Now let’s look at the shots aligned for source 2 by the fence. These are hard to align because a) there are loud supersonic cracks/reflections stepping on the beginning portions of booms, and b) they are all different to some extent. The high-frequency squiggles are the cracks—I am not trying to align those. I am trying to align the lower-frequency pressure peaks between the neighboring shots, and I do get perfect agreement with the podium for shots 4-8.
The yellow line passes through the first “boom” pressure peak for each shot (a few of these are hardly visible). The green line passes through the second “boom” pressure peak for the last five shots—basically invisible in shots 4 and 5. The red line passes through the second “boom” pressure peak in shots 1-3—you can see how broad this peak is and how nothing matches it in the last five shots. The blue, purple, and brown lines pass through other pressure peaks used to confirm alignment for shots 4-8.
These alignments are at the msec level—if shifted enough for shots 1-3 to agree with the podium for the one-shooter theory, that is about half the visible graph and nothing aligns with anything. Something is substantially different here between the two shot groups, and it isn’t just the phone moving a few feet.
I won’t show the DStew plots here. The supersonic crack/reflections are LOUD on all of these and stomp all over the boom with a 9 kHz peak that sounds like breaking glass. Underneath that, however, the first three shots have a prominent low-frequency peak at 546, 531, and 528 Hz, respectively. For the last five shots, the low-frequency peak occurs at 376, 351, 367, 360, and 354 Hz. NOT the same gun (unless recorded in a cop car and released three weeks later). BTW, shots 9 and 10 both have extremely low frequencies (highest below 200 Hz).
Also, at Trump’s mike, shots 1-3 peak at 494, 477, and 475 Hz. Shots 4-8 peak at 544, 528, 532, 528, and 532 Hz, with extremely strong second and third harmonics.
To summarize, at four dispersed locations, the signatures of shots 1-3 and 4-8 show two different guns, a stationary firing position for shots 4-8, and a different TBD firing position for shots 1-3.
