I visually inspected that barn before I was chased off the property. There was no damage to it or fresh repairs.
3 Likes
hi @MacD, I have a model myself using the James Copenhaver footage to adjust the heights of the bleacher to match his eye sight. My conclusion is the first bullet came from Crooks position. The error at David Dutch position is minimal considering it came from Crooks. I even considered the drift and drop due to ballistics.
The critics to this analysis say I canât use Jamesâs footage because the time of his footage is not the time when his is been filmed by the North bleacher footage, so by luck he could have sat in the bleacher during his footage. The North bleacher footage captures James 131 seconds before and 110 seconds after James footage, in both footages he is standing up as everybody is. For me if he was sitting on the bleacher, he would be seeing the heads of the people in the front row and itâs not the best way to film Trump.
As I explained the error in all the measurements done by 3D models lead to the wrong conclusion, and an error of 1 foot at Trump position lead to an error of 6 foot at Crooks position. Google earth heights are not that accurate.
Here is line of sight of James Copenhaver camera:
line of fire on David, Trump and Crooks:
The line of fire come from 1.2 inches above the rail and 3.9 inches to the West of David Dutch center line.
2 Likes
Hey Rough Country Gypsy,
that sounds really strange. Jammed by physical interference or some sophisticated Anti-Drone equipment? Be careful, personal challenge or not. And do not announce your time table. Iâve to thank you a thousand times!
5 Likes
1 to 3 is +400 samples,
1 to 2 is about -240 samples, see below. (Here is the entire 8-shot intervall shifted.)
Iâve been working on inividual shot correlations.
The whole intervall:
This is the 1st shot (to be correlated):
Correlations of 1st shot to the whole intervall:
Of course the offset starts at zero for the first shot, because it is the first.
Channel 2 skipped since it is too noisy.
Channel 3 has an offset of +10 ms compared to channel 1, apparently throughout the whole recording. Channel 2 is in between. 500 sampels are about 10 ms. 10 ms equate to about 3.3 m distance between the outer two mics along the car axis, and this is why I smelled a rat rather early â then thinking that channel 1 was recording ambient sound in front of the car. If this is true or not I do not know for sure until today. See my post here and the also the non-discussion with greg_n.
edit: meant for @kincses-zsolt, on the thread level I notice no indication that this was a reply to him immediatly before.
edit: clarified the offset of 10 ms
2 Likes
I hope this is beneficial to clarify some data points that should be considered settled with a high confidence (90% confidence).
Note: drawing updated (v5)
My drawings and source information is different than roger-knightâs thus it can be an independent check from a different source. I believe the data sources I used are crudely approximate and thus I tried to give percentage accuracy and levels of confidence. The plan dimensions I consider the most accurate at +/-3% accuracy. The grade (soil) elevations are the most suspect, which I have given a +/- 1 foot accuracy.
2 Likes
ch2 is too noisy, maybe Iâll deal with it later.
Now the micro-correlations bother me a lot.
1&3 have approx 400 sampletime delay.
Look at the samples of the shot(1).
So it is not a smooth function of time. It is comb-ish. Thatâs why we can see micro-shift similarities.
1 Like
Neglecting the 5-10 ms phase shift for a moment: if all three mics sense sounds within the cabin, wouldnât their signals be similar? Or what do you mean by micro-shift similarities?
edit: micro-shifts of what numeric value?
I would expect the channels to correlate well with each other with the correlation peaks precisely indicating the delay between the channels, given the microphonesâ proximity. I wish we knew where the microphones are placed in the vehicle.
For each channel, I am interested in the sample numbers of all eight peaks when the first shot in that channel is correlated with the other shots in that channel. These correlations should be so accurate that the gaps between the 8 peaks are identical to the sample in each channel. If they arenât, that is also a red flag because you would at least expect them to get that right if they are simulating all eight shots from a same location. But there are other details they seemed to miss so who knows? Worth a look.
2 Likes
Wow Rough_Country_Gypsy,
Thank you so much for your efforts and really strange, as if they have been waiting for us? I canât believe it. Does this confirm that we are on the right track?
Once again, donât get your self into trouble for these photos, we will work it out in a different way!
Together we are strong!
3 Likes
Somewhere in this thread @brian60221 posted a flyer of the (?) manufacturer of such a surveillance system tailor-made to police cars. Therein they praised the functionality of a separate channel intended to listen into the potential whispering between unsuspecting suspects on the back seats (which should be the â2nd rowâ according to greg_n and I donât think that this patrol car has three seat rows)
edit: Iâm still working, also preparing cross correlation of one channelâs sample with the data sets of the other channels. Iâve solved the resolution problem and underlayed the data points with a grid of ms precision. Soon.
4 Likes
Kudos on how well you combined all 3 elements of the diagram. Very well done!
1 Like
There should only be at most 7 feet, or 6 msec of delay, between the dash and the second row seatrests, assuming the microphones are aligned perfectly longitudinally to the sound source.
Although Ford Explorers have a third row, with it up there isnât much room behind for gear, so I would think patrol cars keep them down to hold more gear.
2 Likes
Interesting, does anyone know if these Butler city police patrol cars have a tight separation between the front and the back rows similar to the old London-style cabs?
again: meant as reply to @offtheback
Hi Phase 5,
Great job!
It looks as if a person standing on the left bleachers at the crossing point of the bullet would not be in the way? Maybe it is possible for you to put one person there on the bleachers to see the clearance?
And sorry one improvement suggestion to avoid any misunderstandings. Could you use 2 different colors for the shots, 1 for Crooks and 1 for the 2nd shooter, because if some people donât study the drawing exactly, it would give the impression that the 2nd shooter misses the corner of the bleachers, which is not the case, it is Crooks that misses that shot!
I will try to present the same view as you did with my CAD system, and we can compare the values.
Once again, bravo!
2 Likes
Excellent work. I thought I might contribute my information in my CAD system as a check. First I simply straightened the photo. I did the following photo drawing interface. I projected vent and window locations from vanishing points in the photo.
This was the information gleaned from that effort:
Are the vents and windows where you expected?
Note: I updated the drawing a bit.
I finished the caéculations. Here is the first 3 shots correlated foor the whole intervall.
if all three mics sense sounds within the cabin, wouldnât their signals be similar? Or what do you mean by micro-shift similarities?
First of all, channels autocorrelate even with a few numbers of sample offset. Imagine you have 2 combs and shift them reéativelly.
1 Like
At about 00:03:07 (18:12:09 according to the clockbar) of the bodycam footage 1302_202407131800_BWC2122104-0RATF.mp4 a man with what can be understood as a MAGA cap is seen standing and probably filming in the direction of AGR Bldg. 6. Someone replied to me that this man was James Copenhaven.
edit: Apparently, there were even two guys filming (the other one standing to the right/North of the first, between the projection of the two cell phone towers)
edit: again no indication that I did reply to @vt1
Excellent work. Regarding whether James Copenhaver was standing or sitting, it doesnât matter much; both scenarios are instructive. It is unique to have a photo exactly from the target of one of the bullets. Regarding the elevation, you have used the level of his camera (which is presumably eye level), but if he was sitting, that would be approximately two feet lower than when standing, which would be about the level of his abdomen where he was hit when he was standing. So if he was sitting, you have modelled exactly the altitude path of the bullet, and if he was standing, you could subtract a couple of feet to obtain the exact path.
What your model does not include are any of the potential obstacles (visible in Copenhaverâs film). If you did this, you might be able to eliminate with certainty some possible shooter positions.