Jump to content
The Education Forum

Jorma Jormakka

  • Content Count

  • Joined

  • Last visited

About Jorma Jormakka

  • Rank
  1. I looked at ther motorcycle noise on the Dictabelt. It is as follows: The timing is that the bell is 56.2s to 56.8s. from 0 to 37.5s there is laud motorcycle noise, basically uninterrupted 37.5s-45.5s motorcycle noise, but more silent, the cycle moves slower but does not stop 45.5s-48.7s the motorcycle is silent, no talking 47.8s-49.3s talking: hold it up right 49.3s-50.5s talking: there, we´re gonna do it 50.5s-51,7s talking: I, I got it 51.7s-52.5s talking unclear: let`s see of you got it 52.5s putative shot1 53.2s putative additional shot (shot4) 55.6s putative shot2, loud beep 55.8s laud beep the motorcycle is silent from 45.5s to 52.5s 57.0s-58.8s the motorcycle speeds up 58.8s putative shot3 (and putative shot5) 59.0-59.5 the motorcycle has stopped 59.5-1:04.00 continuous motorcycle noise, driving faster 1:04.00-1:04.06 beep 1:04.06-1:25.00 continuous motorcycle noise, continues from that longer 1:25.00-1:27.00 motorcycle noise, faint siren in the background, JFK taken to hospital (there is a clearer siren much later, at 2:45 ) So, from this data. The head shot had to be 58.8s since after that the motorcycle does not stop, while the motorcycle drives very slowly from 45.5s to 59.0s and stops totally after 58.8s. This indicates that the motorcycle is in the motorcade and reacts to the shots. About the Barger et al times from 1979. There is absolutely nothing hinting to gunshots in their times. 40.9s and 42.9s there is only motorcycle noise. 48.7s and 48.2s is just at the time of "hold it up right", or if you want "hold everything". The whistling is in two places 1:35.2-1:38.8 and 2:25.5-2:27.3 both times are after the stuck motorcycle has apparently started working ok. The tape has stopped before there times, they are not good for timing and the whistler need not be the one driving the motorbike. The lowest frequency peak from the motorcycle on the Dictabelt is around 380 Hz, corresponding to 22.800 rpm. I do not know how to connect this information to the motorbike type, but to me the noise from the three wheeler is not the same. This is about all I can get from the motorcycle noise. It supports the timing of shots which I proposed. There is nothing that supports the timing of shots as in Barger 1979.
  2. Can you give a youtube video where there is a motorcycle like the one McClain used? The sound on the Dictabelt is passed through the radio channel and is attenuated both on low and on high frequences, so it does not sound the same as throught the air, but if the engine is of the same type, maybe the plot will look similar either in the time domain or in the frequency domain. 3000 rotations per minute should be 50 rotations per second, that is 50 Hz. The channel does not carry so low frequences well. Normally such channels have been designed for male voice 600 Hz to 3400 Hz, the upper bound being even lower in old army digital equipment, but this was for sure analog radio as there is no cutoff from sampling, only the sampling limit which audacity makes 22.05 kHz. There is indeed a whistler on the Dictabelt but around that time there are many people talking, so the channel is already in use by many police officers. It is not clear if the whistler is the motorcyclist. I think the best would be to have an audio from the motorcycle on McClain`s type. Notice, that I do not really care if McClain was the police officer with the motorcycle, as I do not rely on echo correlation and doubt if it can be reliably made on a so noisy data. About this, can echo correlation be made, I take the putative shot at 55.6s in Kennedy_article2. Using tests shots from the 1979 report, I do not find a match to TSBD for this shot, still according to the Zapruder film and the wounds, this shot came from the back, it is Zapruder frame 254. There is a strong beep. Can this beep cover the echo peaks? Firstly, as the gunshot has wide spectrum, the echo peaks also have wide spectrum. It would be very difficult to cancel a wide spectrum peak with another peak to the opposite direction. One should cancel nicely all relevant frequences. It is highly unlikely. But instead of cancelling a peak, how about covering it with other noise? That is certainly possible. Any other wide spectrum noise will cover up a reasonably weak echo peak. Therefore I still accept the possibility that 55.6s putative shot is a shot from TSBD, though it fails the echo test.
  3. Yes, cycles per second should be smaller that cycles per minute, it should be 60 times cycles per minute, so how do I get rpm from those plots where the power spectrum is in Hz?
  4. I do not know about the motor rotations per minute (rpm). We have to solve this problem first. So, the plots of the Dictabelt have a maximum at 1100 Hz=cycles/second, divided by 60 that is 18.3 cycles/minute. If that is not 18.3 rpm, then I am lost, but help me, I do not know anything about bikes. There is fig.7 in the second paper I wrote, Kennedt_article2. It shows thje motorcycle noise at 58.8s, when I think the head shot occurred.The motorcycle stops, but starts again. At 56.2 starts a bell, what you see there is the bell, the motorcycle is very silent. So, indeed if you take the shot times that the Barger et al study from 1979 found, the motorcycle was running at a constant speed, but that is 9.6 s before the times I propose. In the time I propose, the motyorcycle has stopped. The times Barger et al proposed are exactly at the time when there is the talk we`re gonna do it, I, I got it,... it is not a correct place. I have synchronized the Dictabelt to the Zapruder film in Kennedy_article2. I say, Frame 313, the head shot, is at the time 58.8s in a Dictabelt recording where the bell starts at 56.2s. That is an exact timing. So, from that we can say that the recording of the Dictabelt started 58.8s before the Zapruder film frame 313. So far I have not looked at when the motorcycle noise started, but I will do the same as Thomas did, only it will take some time, my program only reads *.au files, that is, small parts of the audacity data. But answer about the rmp, we should get that issue straight. You see from the power spectrums that the maximum is at 1100 Hz. How does that relate to the rmp? If at all. Must be at all, if you listen you hear 1100 Hz noise.
  5. Thanks, I will try at some point to match the motorcycle movements. In Barger`s report from 1979 the last shot on the head is 49.2s, I place it to 58.8s, that is 9.6s difference. It is possible that the motorcycle movements fit quite well to my shot times also. Look at fig.7 in my article 2. The motorcycle nearly stops at the head shot 58.8s, and it is going really slow when the shots come. Clearly, the motorcycle seems to be in the motorcade as its speed seems to be responding to the purported shot times. It always stops after a shot. The Zapruder film fits to my shot times. I does not fit the times of Barger`s report, as I time their times by the bell. It may be incorrect to assume that there are sounds from only one radio and drawing conclusions based on the exact location of the single radio. If there are sounds from several radios, this way will lead to contradictions. I just say that if it is a rifle shot, there is more power on high frequences than in motor noise or voice, and so there still is more power on high frequences than in motor noise and voice after passing the low-band filter effect imposed by the communication network, and looking for thick tails is the natural way to locate gunshots, if they exist on the tape.
  6. I contacted Donald Thomas when I first started this study, because I looked at this after having watched his video in the web. Now, he was not very willing to share any information and he was not interested on my views either, or that´s how I saw it. But I got into some kind of a dispute with Steve Barber, I guess you understand he did not like my work that much, I did send my pdfs to him but have not got any answer yet. Yes, the match with the correct speed and everything looks good, but their timing of the shots is difficult to fit to anything on the Dictabelt and in the Zapruder film. Their time between the two first shots is less than 2.3 s, the third shot comes from front, cannot match these and my math does not confirm that their match is significant. In general, in echo correlation analysis one usually knows when the shot came and the task is to say where it came from. What they did was to find a match for a pattern. That is not what their competence was about, I think, if they did echo correlation analysis for the army. But I agree that my argument should be made stronger. I´ll think about it. Cannot improve it now. Thanks to you.
  7. To Chris: I will measure the noise from the three wheeler from the video, but the noise in the Dictabelt can be seen in fig 4 in article two, you say 26,00 rpm, rounds per minute, that is then 60*26=1560 1/s, or 1.56 KHz. Plot in fig 4 is motorcycle noise and there is a line at 3kHz. Clearly 1,56 kHz should be about half between 0 and 3 kHz. I measured the Harvey. The plots are in http://www.pienisalaliittotutkimus.com/2017/12/18/who-killed-j-f-kennedy-and-why/ there is a link to Answer_to_Chris. To compare these two motorcycles, one first have to understand that the one in Dictabelt is recorded through a communication network (police analog radio) which attenuates, but does not completely cut, high and very low frequences. Thus, do not look at the power spectrum expecting that they should look the same, they do not. What one must look at is the rpm. The maximum frequence in the Dictabelt is 1100 Hz, which is 18 rpm (divide by 60), while in the Harvey we have about 1560 as you say, 26 rpm. I think these cycles are not the same type. To Eddy: I put a video available. There is the word boundary problem between Hold1 and Hold2.
  8. Answer to Eddy, Thanks for your answer. It is very helpful. I try to address some of the issues now, some will require more time, checking and such. The report Barger J. E., “Analysis of recorded sounds relating to the assassination of President John F. Kennedy,” report no 3947, to HSCA, 1979, has information of the trial shots. They have measured the echo peaks ion different positions of the shooter and the target and they find matches for the peaks in the traffic in the Dictabelt. The traffic in the Dictabelt is very noisy because of the motorcycle noise and voice traffic. The question is, does the fact that the test shots match a large number of echo peaks, mean that such a match is significant? In the report it is calculated that such a match is highly significant, in my opinion it is not significant. To explain, if they had used only three echo peaks, they certainly could find a place from the Dictabelt where there is a perfect match to the echo peaks, simply because there are so many noise peaks in the Dictabelt. I made a small estimation assuming they used 6 echo peaks, and it is about 50-50 to find a match to ANY given pattern of 6 echo peaks. They have measured 11 echo peaks for the test shots, but some of these peaks are weaker. I do not think Barger et al are justified in claiming that the matches are significant. It is very difficult to estimate what is the probability of the event of finding matches. The matches they found are in places where I do not see what there should be in gun shot noise: the thick tail. I will explain why it must be there. You pick up signal from the air with a microphone and send it through a network. This communication channel typically attenuates the high and very low frequencies, but if you have two signals and one has more strength on high frequencies than the other, you get to the receiving end more strength on high frequencies than from the other. As a gunshot has very high power levels on high frequencies, much more than voice or motorcycle motor noise, you must get more power to the receiving end. You get the thicker tail. The motorcycle may have a banging exhaust pipe, so it can give similar noise, but if the noise is from a gunshot, there must be a thick tail, thicker than from voice. These places that Barger et al identify do not pass this requirement, so they must be wrong. The firm was very competent. I doubt the reason is competence, more like unwillingness to go for the conspiracy claim. So, I did not, by reading Barger, notice any other reason for claiming that the noises look like gunshots than that the echo peaks match. There was no argument that the signal form is like in the test shots. Indeed could not be, as the test shots absolutely surely had thick tails in the frequency domain. Barger shows no plots in the frequency domain. They characterize a gunshot in the general level. About the microphone position. It was essential to the analysis of Barger et al. It is not essential to my analysis. My analysis identifies gunshots by their looks in frequency and time domains. Their analysis identified the gunshots by matching echo peaks. That is why in their analysis, the echo peaks must be exactly correct, one has to know the position of the radio, which picked up the gunshots. Whether this was the same radio that transmitted the noise, I say, does not need to be the same. It is not necessary that McLain`s radio was stuck and it was his motorcycle noise on the Dictabelt. It is still quite possible it was McLain´s radio which recorded the gunshots and the echo correlation analysis assumption is correct. (But their result was not, in my opinion). If one tries to prove the case by echo correlation analysis, then it is essential to know the position of the radio. In my opinion such an effort will not succeed because the data is too noisy. You cannot identify the shots in another way and use the echos to locate the shooter, which is the normal way to do echo correlation analysis. They had to locate the shots from noisy data by matching echos. That is risky, not standard, not good. About the crosstalks. My arguments against Hold1 being Hold2 are like this: Both Hold1 and Hold2 are parts of a longer talk. Listen to the Dictabelt and try to hear if Hold2: Hold everything secure until homicide and other investigators get there Hold1: Hold it up right there, we´re gonna do it, I, I got it,……, you got it. You ask for something else than listening and I will come to it immediately, but compare the endings until homicide… and we`re gonna do it… They are different. It should be visible in the signal. Look at fig 1 in Kennedy_article1. From 589.6 ms the signals Hold1 and Hold2 do not look similar. I have not shown the power spectrum, but from 598.6 ms the power spectrums are different. The beginning is the phrase Hold everything secure or Hold them up right there. That is similar in spectrum and similar in correlation, the continuation is not. Should it be? I think it should be. The talk continues in both signals. Second proof. The signals Hold1 and Hold2, the parts Hold everything secure, do not produce motorcycle noise when Hold2 is subtracted from Hold1 moved as closely to the correct starting point and weighted in an optimal way. Notice that this is male speech, which never reaches frequency 22 kHz, it is radio speech and more like 3500 Hz maximum, so as the sampling rate is 44100 samples/s, sampling error can be ignored for speech. Moving Hold2 in time it must at some point match Hold1 if they are the same signal. Weighting Hold2 with some weight it must cancel totally the signal from Hold1. I have the software in the web page, you can try. I did try and they do not cancel and leave motorcycle noise. So, these are not the same signal. I am pretty sure about it. Third proof. You can check this from the Dictabelt and from samples of Audiograph. I took my Audiograph (Hold2) from a video of Thomas. Anyway, the signal Hold2 is too regular. What Hold1 should be: addition of motorcycle noise and Hold2. If Hold2 is regular at these places, and Hold1 is not, then the reason should be the rather humming motorcycle. No, the motorcycle noise is rather regular, human voice is not regular. I am not an acoustic expert, but I find regular human voice signal odd. Especially so if, when I listen to the signal in parts, it starts feeling even more odd. I think Hold everything secure is manufactured from the signal Hold it up right there in the Dictabelt, or manufactured to look and sound like it. Fourth proof: I did not include this to the paper, but there is a word break in Hold1 and no world break in Hold2. As Hold1 is the one with noise and the word break is where it should be by listening, the break is not caused by added noise to Hold1. Thus, it should be in Hold2, but is not. How sure I am, so sure that I post a paper claiming it, knowing fully well that if I am right, there is talk on the Dictabelt just before the shots saying highly strange things. See where the shots are: The talk on Dictabelt “hold it up right there, we re gonna do it, I I got it … you got it. Ends just before 55.6s the second shot, then comes the bell 56,2s then comes the third shot 58.8s and in my opinion the fifth 58.9s. Two shots are at the end of the phrase 52.6 and 53.2s. To Chris Bristow: Thanks for you comment. I have not measured the motorcycle noise, but that is easy to do. The frequencies are just from the power spectrum. I can look at this, it will be a while.
  9. I wrote two short pdfs on the Dictabelt, in http://www.pienisalaliittotutkimus.com/category/posts-in-english/kennedy-assassination/ I would appreciate comments, noticed misunderstandings, errors etc. It is not really my field, so there may be anything. Thanks, Jorma
  10. My name is Jorma Jormakka. For a cv see http://www.pienisalaliittotutkimus.com/wp-content/uploads/2017/12/short_cv.pdf
  11. I am Jorma Jormakka. Nowadays I am independent researcher, meaning I can study different open problems on whatever fields. Up to 2010 I was a professor, in military technology and in communication networks, since summer 2010 I tried to write novels. I wrote two, but did not get them published, so gave it up.
  • Create New...