Home/Browse Archive/ Browse Essays/ Selected Essay/Overview and History of the Acoustical Evidence in the Kennedy Assassination Case


Overview and History of the Acoustical Evidence in the Kennedy Assassination Case

by D.B. Thomas

Part 1    Part 2    Part 3


The first suspect sound on the DPD recording matched to a test shot that was recorded on a test microphone on Houston Street near the intersection with Elm Street. The very next suspect sound on the dictabelt matched to a test shot recorded at the very next microphone, 18 ft to the north on Houston Street. The third suspect sound matched to a test shot recorded on a microphone in the intersection of Houston and Elm Street. The fourth sound matched to a test shot recorded on a microphone on Elm Street, and the fifth suspect sound matched to a test shot recorded on the next microphone to the west. Thus, rather than in a random pattern, the chronological order of the suspect sounds had matched to the topological order of the microphones that produced matches to the test shots. To a scientist this sort of orderliness is very significant because there are 125 ways to sequence five events, only one of which is 1-2-3-4-5.

If the sounds on the dictabelt are anything but Dealey Plaza echo patterns, then the matches are purely by chance, having nothing to do with the geometry of Dealey Plaza. And that being the case, any spurious match was as likely to occur at any one microphone as at any other. And if the five matches achieved were truly spurious then the sequence of matches relative to the test microphone locations would have nothing to do with the location of the motorcycle with the open microphone. They would be in nonsense, random sequence. But the results are not scattered at random. The matches are precisely in the order that would be expected of a motorcycle traveling with the motorcade.

Map showing locations of microphones at 18-foot intervals in Dealey Plaza during the 1978 acoustic tests. The figure also shows the 5 microphones which matched echo patterns.
(view larger version)

But it is not just the sequence that was ordered. The spacing on the sounds on the dictabelt matched the spacing of the test microphones in Dealey Plaza. The first three suspect sounds are each separated by a little more than a second. The first and second are 1.6 sec apart, while the second and third are 1.1 second apart. Then there is a 4.8 second gap before the fourth putative shot which is only a fraction of a second (7/10ths of a second) before the final suspect sound. The test microphones that recorded the matching echo patterns were three in a row at the intersection of Houston and Elm. The test microphones that matched the last two sounds were at two consecutive positions about 80 ft west of the intersection, skipping four microphones between.

But it wasn’t just the sequence and the spacing that matched. The distance from the first test microphone that achieved a match on Houston Street was 130 ft away from the microphone on Elm street where the last match occurred. The time lapse between the first suspect sound and the fifth suspect sound was 8.3 seconds. In order to travel 130 ft in 8.3 sec an object would require an average speed of around 12 mph. In 1964 the FBI using the Zapruder film, calculated that the President’s limousine was traveling at an average speed of 11.3 mph on Elm Street.

There has been confusion regarding the order in the data because the suspect patterns matched to more than one test pattern. It has been suggested that the order occurs only if one picks and chooses from among the several matches. This is not true. It is true that there were multiple matches of test shots to each (save one) of the suspect sound patterns. In total, there were 15 test shots that matched to the five suspect patterns (Table 2). Consider the first suspect pattern. Four test shots “matched” to this suspect pattern. All four test shots were recorded at one of two adjacent microphones (nos. 5 & 6), 18 ft apart, on Houston Street. Of course there would be no reason to expect that the suspect motorcycle would have been at precisely the same location as one of the test microphones, rather, if it was there, it would have most likely been in between two consecutive test microphone positions. This was the reason for using the ± 6 msec criteria in scoring the impulse matches. It is thus not surprising that matches would be achieved at adjacent microphones, or surprising that the degree of match would be higher at one microphone than to the other. Nearly all of the multiple matches to each suspect pattern were to adjoining microphones. If one wished to ignore the scoring coefficient evidence, never a good idea, and only look at the cluster of matches for each suspect pattern, then the order in the data still appears. The data is shown in Table 2. The cluster of matches that involved the first suspect pattern was to two consecutive microphones (5 & 6 of the 2nd array) on Houston Street near Elm Street. The second cluster occurred at the intersection of Elm and Houston with three of the four matches involving two adjoining test microphones (6 & 10 of the 2nd array) at the intersection with an outlier by itself further down Elm Street (5 of the 3rd array). The third suspect pattern did not have a cluster because there was only one match with this pattern and it was to a microphone (11 of the 2nd array) on Elm Street near the intersection at Houston. The fourth suspect pattern had three matches clustering at microphones on Elm Street (4, 7 & 8 of the 3rd array), with the fifth pattern, only a split second later, having three matches clustering near the same spot on Elm Street (5, 7 & 8 of the 3rd array). Hence, there was overlap between succeeding clusters because the shots came so close together. But the order in the sequence of clusters still arises. The only way that the matches become disordered is by deliberately picking and choosing discontinuous test positions from within overlapping clusters, ignoring the scoring coefficients, and further ignoring the other matches in the cluster. In other words, by doing exactly what the critics have falsely accused the acoustical experts of doing.

In considering all of the data, the acoustical experts concluded thusly:

"If the motorcycle were not moving through Dealey Plaza at the time of the assassination, the distance along the motorcade route would be a meaningless coordinate, and the microphone locations for the correlations that exceed the detection threshold would occur at random.” “The order in the data would not likely have occurred by chance.” “We conclude that the motorcycle was moving through Dealey Plaza and did, in fact, detect the sounds of gunfire." [BBN Report pp. 64-66.]


It was this order in the matching data, and not just the echo delay time matching, that convinced the acoustical experts that they had found the assassination gunfire on the Ch-1 recording. Importantly, one of these suspect sounds had matched to a test shot fired from the Grassy Knoll. Because of the controversial nature of these results the HSCA decided to seek a second opinion. The second laboratory on the ASA list, the computer science department at Queens College was asked to review the work of the BBN lab. These folks were the experts who write the computer programs for processing acoustical data, such as sonar, for military applications.

Their approach was based on the principles of acoustics applied to the physical nature of the recordings and the circumstances of the assassination. Either the suspect sounds were Dealey Plaza echo patterns or they were not. If they were not, then it was extremely unlikely that all five patterns would have matched in the orderly fashion that they had. But a second scenario, suggested by an FBI analysis, provided an alternative. Even if the sounds on the DPD recording are Dealey Plaza echo patterns, one or more might not have been from a gunshot. A motorcycle backfire or a large fire cracker could produce an echo pattern that would match to a test pattern if it happened in Dealey Plaza. So, even if the suspect sounds included the assassination gunfire, there could still be a false positive among them, perhaps even the grassy knoll shot.

Oscillogram of test shot fired from grassy knoll (top) and sound pattern on DPD recording (bottom). (view larger version)

Consider this fact; the fourth suspect sound had not only matched to a test shot from the grassy knoll, but had also matched to test shots from the book depository, though to a lesser degree. The Queens College scientists judged that the weakness in the analysis was due to the 18 ft spacing between the microphones, and thus the ±6 msec leeway that was applied to judge the matches. The primary factor that most influences echo delay time, is the position of the microphone relative to the buildings, with the position of the shooter being secondary. As the BBN experts explained in their testimony, their test was designed to find the motorcycle, not the assassin. The shooter location does matter, but the difference is in the details, and the details are smaller than the 6 msec gaps. To find those detailed differences the sonar experts would have to close the gap. To close the gap they would have to move the test microphones closer together. Had there been a dense array of test microphones, in a grid instead of a line, and set at 2 ft spacings instead of 18, then the window of match could be set at a demanding 1 msec precision. To actually perform such an experiment would have been logistically daunting. Some changes in the experimental design would be relatively simple, such as placing the microphones in a grid (they would need about 180 for each shot) and repeating the test shots from a line of positions along the fence on the knoll. The test could be performed when the air temperature was exactly the same as on the day of the assassination (65F) to account for the fact that the speed of sound, and thus the arrival time of the echoes, is a function of ambient temperature. But the most difficult parameter to include in the test would be the motion of the motorcycle.

The problem was solved by doing what sonar experts do. They simulated the sonar conditions with a computer. If one thinks about it, a sonar operator on a submarine faces the same problems: the water temperature varies with latitude, the vessel is traveling at variable speeds relative to the bottom and at varying angles to other ships in the water. A computer can factor in and adjust for all of these parameters. Mark Weiss and Arnold Aschkenasy wrote those computer programs for our Navy. Now, using the grassy knoll shot as a template, they wrote a program that simulated an echo pattern for each of 180 locations in a grid surrounding the position of the test microphone that achieved the best match to the suspect pattern, factoring in a trajectory westerly at 11 mph, and with an air temperature of 65 degrees.

Photograph showing the concrete pergola and beyond it the wooden fence atop the "grassy knoll," taken shortly after the gunfire in Dealey Plaza.
White House Photo taken by Cecil Stoughton.

With those adjustments implemented, they found that the suspect pattern on the DPD tape was a perfect match for a test shot simulation at a position five feet to the southwest of the test microphone position. The simulated test shot had 26 echoes. When the muzzle blast of the test shot was aligned to the first large impulse on the DPD suspect pattern, each of those 26 echoes was within 1 msec of a corresponding impulse in the suspect pattern. But this wasn’t good enough to satisfy the experts because they knew that most of the impulses on the test recording were actually the piston firings of the motorcycle motor. To separate the signal from the noise they used a noise threshold. The first large impulse in the suspect pattern was presumably the muzzle blast. The impulses in the one second period immediately before the big impulse must be mostly motorcycle motor noise and that level established a noise threshold. Therefore, for the second comparison, the experts looked at only those impulses in the suspect pattern that were larger than the noise threshold. There were 14 large impulses after the putative muzzle blast. In the test shot pattern there were only 12 large echoes (louder than the noise threshold). It was known from the lab tests that the DPD radio recording, like all radio systems, have static and that the marking rate of the static was around 8 impulses per sec. Within this one-third second long pattern they expected a couple of static impulses so the two extras were explicable. They were not disregarded however, another misunderstanding by internet critics. The two extra impulses are scored as non-matches. Of the 12 impulses in the test shot, 10 matched to the impulses in the suspect pattern to within 1 msec. It wasn’t that two impulses had failed to align to any impulse, but rather, each aligned to impulses that were smaller than the noise threshold. Interestingly, the two impulses in question were echoes that arrive from in front of the motorcycle. The sonar experts pondered this anomaly and then realized that the microphones on the police motorcycles were mounted behind the windshield. They realized that the windshield could have reduced the intensity of the sound and could explain why the two impulses were smaller than expected. Nonetheless, these two impulses were also scored as non-matches, not disregarded. Hence the final matching score, using the notation defined previously {N:n:M} was 14:12:10 with the ±1 msec bins. The odds of getting that degree of match by chance is more than one hundred thousand to one against (see my 2001 article in Science & Justice for the formal calculation).

One complaint that I have heard concerning the sonar analysis was that it should have been applied to the other matches as well. I have always suspected that one reason that Weiss and Aschkenasy were not contracted to apply the analysis to the other shots was because of the “rogue” shot, the fifth pattern swept under the rug by the HSCA counsel (more on that later). It is worth mentioning that this complaint originated with the NRC panel that was asked to review the acoustical evidence. Of course, if it were true, as the NRC claimed, that the sonar analysis was necessary in order to validate the matches, then the NRC panel should have done the analysis as part of their study. Unlike the acoustical experts who were given neither the time nor the funding to perform such an analysis, the NRC panel had both. Given their motivation to discredit the acoustical evidence, one is led to conclude that if there had been any real chance that the book depository shots would have failed the sonar analysis, the NRC panel would have done it. Certainly, the sonar analysis should be applied to the rogue shot if some lab could be funded, or convinced to do it for free. In my opinion this project would make a wonderful doctoral dissertation for a graduate student in acoustics. Such a study would not just determine the origin of the rogue shot (the second impulse pattern), but would also test the assertion that the third pattern was a “false alarm.” Moreover, such a test should include all of the patterns that matched to the book depository test shots as they provide “controls” to establish the statistical confidence level (there are some important differences between these and the grassy knoll shot which will impinge on the precision achievable in the analysis). But to claim that the test is necessary to validate the latter is no more that a pretext to hold the truth hostage to a test that is both unnecessary in that sense, and unlikely to be done. The reason is simple scientific logic. A scientifically valid hypothesis is one that is falsifiable. The sonar test cannot falsify the hypothesis that the suspect sounds are gunshots because even in the worst case scenario, the suspect sounds still match the test shots to a statistically significant degree.

Firstly, unlike the case with the grassy knoll shot, there is no basis to doubt that there were gunshots fired from the book depository. Secondly, the grassy knoll shot was a stand alone pattern whereas, the book depository shots are a grouping. A single event is far more likely to be a spurious event than a close grouping of events (when one plane crashed into the trade center towers in New York City on September 11th we thought it was an accident; when the second plane crashed just minutes later, we knew something was up). Finally, and most importantly, one should consider, in a scientific context, what the possible outcomes of such a test might be. It is already established that the suspect patterns match to the test patterns to a statistically significant degree. The sonar test is simply a more refined test of the same parameter, echo delay time. So for the sake of argument, let us suppose that the book depository patterns were subjected to the sonar test and failed to meet the ±1 msec standard, or more likely, that one or two out of the three failed. What scientifically sound conclusion could one reach? One would certainly not conclude that there was evidence for fewer than three gunshots from the book depository. Nor could one argue that there isn’t evidence for three shots from the book depository on the DPD recording, because the significant patterns are there. It would also be logical to conclude that the other shots might not have come from the Sniper’s nest window (though not defensible on non-acoustical criteria). The problem devolves to whether the initial procedure gave false positives or the more refined procedure false negatives That question is not answered by refining the same line of evidence. Refining the test cannot make the matches go away. The instantaneous speed of the motorcycle will almost certainly vary from the average speed during the shooting. Combined with acoustical shadows such as a vertical gradient in air temperature or a small gust of wind at the critical moment, could offset the adjustment made for temperature, for example. Quite reasonably, a match using a ±2 or 3 msec window should also be considered a positive result. Ultimately, the way to test the acoustical evidence is to look for concordance with the non-acoustical evidence. That is why the order in the data, and not just the echo delay matching, leads to the conclusion that the assassination gunfire is on the DPD dictabelt. And that is also why one looks to the filmed evidence to challenge the acoustical analysis, because therein lies the capacity to corroborate or falsify the acoustics hypothesis.

Aerial view of Dealey Plaza. The TSBD is at top center, with Elm street curving past it. The railroad overpass is at the left side, with the grassy knoll between it and the TSBD.
Photo courtesy City of Dallas.


The final report of the HSCA states that there were four acoustically identified gunshots on the DPD recording. But as the above section explained, and as appears in the data tables in the technical reports, the BBN experts had identified five suspect patterns and all five had matched to the test shots in Dealey Plaza. One of the patterns was swept under the rug because the Chief Counsel for the HSCA felt the acoustical evidence would be more convincing if the acoustical results had a better fit to the Warren Commission’s finding of three shots from the Book Depository. Four of the suspect patterns had matched to test shots fired from the Book Depository in BBN’s field test (Table 2).

Reference to Table 2 shows that five patterns passed the echo delay matching test. The five patterns are identified by their chronological position: 137.70, 139.27, 140.32, 145.15, and 145.61 seconds after the beginning of the motorcycle segment. One of the five, the pattern at 140.32 sec was judged to be a false alarm and discarded. This was the one glaring error in the acoustical analysis. The BBN Report states,

"The entry in Table II that occurred at 140.32 sec is a false alarm, because it occurred only 1.05 sec later than earlier correlations also obtained from the TSBD. The rifle cannot be fired that rapidly. Since there are three correlations plausibly indicating the earlier shot, the one occurring 1.05 sec later must be a false alarm."

TABLE II.- CORRELATIONS BETWEEN IMPULSE PATTERNS: SIX SEGMENTS OF THE DPD RECORD AND ECHO PATTERNS FROM 432 TEST SHOTS _________________________________________________________________________

Time of
1st Impulse
Correlation Coefficient
136.20 sec - - - all < 0.5
137.70 sec 2(5) TSBD 1 0.8
" 2(5) TSBD 3 0.7
" 2(6) TSBD 3 0.8
" 2(6) Knoll 4 0.7
139.27 sec 2(6) TSBD 3 0.8
" 2(6) TSBD 3 0.6
" 2(10) TSBD 3 0.6
" 3(5) Knoll 2 0.6
140.32 sec 2(11) TSBD 3 0.6
145.15 sec* 3(4) Knoll 3 0.8
" 3(7) TSBD 2 0.7
" 3(8) TSBD 3 0.7
145.61 sec 3(5) TSBD 3 0.8
" 3(6) TSBD 4 0.8
" 3(8) TSBD 2 0.7
146.30 sec - - - all < 0.5
*Subsequent analysis resets the onset of this impulse to 144.9 sec

The logic of this statement is that if it didn’t come from Lee Harvey Oswald’s rifle then, it was not a shot. But of course, the whole purpose of the inquiry was to test the Warren Commission’s single assassin theory against the facts, not the other way around. The fifth shot was dismissed because five shots were less palatable to the committee members than four shots. Palatability is not, however, a scientific criterion for judging the validity of evidence. Moreover, it was illogical to dismiss the pattern at 140.32 as a false positive because it was too close to the previous shot. The first two putative shots are only 1.7 sec apart, also too close together to have been fired from Lee Harvey Oswald’s rifle. If the subject sounds are the assassination gunfire, and if three of the shots are attributable to Oswald’s rifle, then the second pattern is the rogue shot, not the third. But, the second pattern, at 139.27 sec, could not be dismissed as a false alarm because it was supported by multiple correlations, including a robust correlation coefficient of 0.8. The weakest supported pattern was the pattern at 140.32, with a score of only 0.6, and only one match, and was thus selected as the “false alarm.” The fact remains that five candidate patterns passed the initial screening tests, and all five matched to a significant degree with the test shot patterns. The time intervals between these putative shots, corrected for tape speed, were: 1.7, 1.1, 4.6 and 0.7 sec.

Ultimately, one still has to ask the scientific question: how do we know the matches are not false positives? There are several steps that scientists can take to improve their confidence in their test results. One, is to have a second laboratory replicate the experiment. The HSCA had done that by bringing in the sonar experts from Queens College. Another way to assess the reliability of a test result is through statistical analysis. And here the experts had made another mistake.

The purpose of a statistical analysis is to answer the question, what are the mathematical odds that the result could have occurred by chance? When scientists conduct an experiment they try to design it in such a way that the results can be defined mathematically. This makes it easier to apply the statistical analysis. The value “p”, is the probability that the result occurred by chance. Four parameters go into the calculation. The first three are straight forward data points as discussed previously: (N) the number of echoes in the test shot, (n) the number of impulses in the suspect pattern, and (M) the number of matches between the two patterns. The fourth parameter was somewhat of an artificial number used to make the statistical analysis easier. It is (i), the number of spaces in the suspect pattern where an impulse could occur. Because the grassy knoll pattern had a duration of 360 msec, and because the standard for determining a match was a window of 2 msec (±1 msec), the total number of spaces (i) would be 360/2 = 180. But in this particular case there was a problem. An underlying assumption of statistical analysis is that the data is “normal.” One example of “non-normal” data, is data that is clustered. In this case the impulses were clustered. Although the echo pattern had a duration of 360 msec, all of the impulses occurred in the first 90 msec and the last 90 msec. This was because the first cluster was the echoes from the structures on Elm Street, whereas, the second cluster was from structures on Houston Street. In order to minimize the clustering effect, the experts simply discounted the 180 msec gap in the middle. This was a very conservative adjustment, in that it helped eliminate false positives. The mistake happened because two laboratories were involved, and each independently made the adjustment. The correct adjusted value for (i) that should have been applied to the calculation was 90. Because they double corrected, the experts used 45. The result was to underestimate the statistical significance of their result. The sonar experts reported that p was 0.05 meaning that the odds of the results happening by chance was about five percent. Actually, it was much smaller than that. The odds that the match of the suspect pattern on the DPD recording to the test shot from the grassy knoll happened by plain old bad luck, was around 1 in a hundred thousand.

To err is human. In science there are two possible errors. This is so fundamental that scientists designate them as the Type I error and the Type II error. The Type I Error is accepting as true a hypothesis which is false. The Type II error is rejecting as false a hypothesis which is actually true. The acoustical experts had taken steps to avoid the Type I error by using the conservative adjustment in the statistical analysis and by using overzealous criteria in judging false positives. Because any mistake tends to mislead, the Type I and Type II error concept has a corollary. If the hypothesis is true, then any mistake is going to tend to make it look like it’s not true, but if the hypothesis is false, then any mistake will tend to make it appear that it is not false. Each of the mistakes made by the acoustical experts hurt their case by making the hypothesis look weaker than it was. The acoustical evidence was stronger than even the HSCA’s acoustical experts realized.


If the sounds on the DPD recording were anything other than Dealey Plaza echo patterns, then they would not occur in the order and spacing that would match the trajectory of the motorcade vehicles. This order in the data is that which convinced the acoustical experts that they had found the assassination gunfire. It also suggests another way to test the validity of the acoustical evidence. The sonar analysis had placed the open microphone at a position 141 ft behind the presidential limousine at the time of the Grassy Knoll shot. Robert Groden, a film expert advising the HSCA, searched the archival footage of the motorcade for evidence of the suspect motorcycle. He was able to quickly eliminate most of the18 motorcycles in the president’s escort, leaving four in the general area required by the evidence in the mid-section of the motorcade. Newsreels show all four of these officers on Elm Street shortly after the shooting. Patrolmen Haygood and Baker are clearly in the wrong place, but officers Courson and McLain were in the right area. Of the two, McLain, in position ahead of Courson, was closest to the predicted location. In his sworn testimony before the HSCA, McLain revealed that his radio did have a problem with a sticky microphone relay. The filmed evidence is not conclusive because it only shows McLain before and after the shooting, not during the crucial shooting sequence. It can only be said that McLain, and none of the others, could have been the motorcycle patrolman with the stuck open microphone. Some researchers, among them Gary Mack of the Sixth Floor Museum, believe that there is enough archival footage to narrow down McLain’s exact position during the shooting by the process of elimination. That is, by showing where he was not during the gunfire sequence. Preliminary scrutiny suggests that a film taken by an employee of the Book Depository, Mrs. Elsie Dorman, does show the midsection of the motorcade and may be key in this analysis.

Some analysts have claimed that the motorcycle officer that appears in the Dorman film is McLain. The film shows an officer arriving at the intersection of Elm and Houston at a time which is approx. 4-5 sec after the fatal head shot. The acoustical evidence requires McLain to arrive at the intersection much earlier, at about the time of the second shot, which is about the same time as the Mayor’s car, the sixth in the motorcade. There is no direct evidence that allows one to identify the officer in this film. But there is indirect evidence that the officer in the Dorman film is Courson. Courson recalled that when he arrived at the intersection and looked down Elm Street, he saw Mrs. Kennedy on the trunk of the limousine. In the Zapruder film, Mrs. Kennedy can be seen on the trunk between frames 360-400. Because the fatal head shot impact occurred at Z-313, she was on the trunk between 3-5 sec after the head shot. Hence, if the officer in Dorman is McLain, then Courson’s version is false, because he was so far behind McLain that he cannot get to the intersection in time to see the event he said he saw. But if the officer in Dorman is Courson, then he is arriving just in time to see what he said he saw. Thus, the indirect evidence indicates that the officer in Dorman is Courson.

Most importantly, the Dorman film sequence begins with the Mayor’s car in her narrow field of view on Elm Street, then pans eastward towards the intersection. Because there is no other officer in the Dorman film, if the officer is Courson, then McLain has to have been just ahead of the Mayor’s car out of the lens view, which is exactly where he is required to be by the acoustical evidence. The Dorman film proves this much: “close but no cigar” is not an outcome. Either McLain was exactly where the acoustical evidence requires, or he was way behind and no where close.


Another obvious way to test the validity of the acoustical evidence is to compare the sequence of wounding seen in the Zapruder film with the sequence of gunfire indicated by the DPD recording. The separation between the five shots was: 1.6, 1.1, 4.8 and 0.7 sec. The next-to-last shot was the grassy knoll shot. In the Zapruder film the fatal shot that struck the President is gruesomely obvious at frame 313. The only other wounding seen in the film occurs between frames 234 to 250 during which Governor Connally is clearly showing indications of anguish before collapsing down into the arms of his wife. The Warren Commission opined that the actual impact had to have happened at some earlier point, perhaps at a time when the governor was out of Zapruder’s view. However, a frame-by-frame photogrammetric analysis by a firm called Failure Analysis Associates, under contract with the American Bar Association in 1992, discovered that the lapel of the governor’s jacket flapped outwards at frame 224. Because the governor did sustain a chest wound caused by a bullet which exited through the front of the jacket, and inasmuch as the lapel flap is just an instant before the governor exhibited the overt response to his wounds, the lapel flap provides evidence for the impact at Z-frame 224.

The separation in time between the two visible wounding events is therefore: 313-224 = 89 frames. Because Zapruder’s camera had a film speed of 18.3 frames per second, the separation between the impacts was 89/18.3 = 4.8 sec. This exactly matches the separation on the police tape between the acoustically identified Grassy knoll shot and the immediately previous acoustically identified shot from the Book Depository: 4.8 sec. This agreement provides one with even greater confidence in the acoustical evidence.

Sequence of gunfire in seconds based on acoustical evidence.


In addition to the visual detection of impact in the Zapruder film, there are also more subtle clues caused by involuntary movements of the movie camera by Zapruder in response to the gunshots. Startled by the loud gunshot blasts, Zapruder’s sudden body reflexes imparted angular acceleration to the camera body which is detectable in the film. Not surprisingly, the largest blur in the Z-film is frame 313, coincident with the impact of the fatal bullet to the president’s head. The next largest blur occurs at frame 227. Factoring in the speed of sound, a gun blast reaching Zapruder’s position, 270 ft away from the snipers nest, gives a trigger time at the equivalent of Z-222. A bullet fired from a Mannlicher-Carcano rifle (muzzle velocity 2160 fps) would have struck Governor Connally at Z-224. Hence, the blur evidence corroborates our inference that the lapel flap at Z-224 is evidence of impact. Moreover, there is close accord between the blur evidence, the acoustical evidence, and the visual evidence of wounding.

Interestingly, when the same analysis is applied to the head shot at Z-313, it demonstrates that the blur occurs too soon to have been a shot from the Book Depository. A Mannlicher-Carcano bullet shot from the sniper’s nest striking at Z-313 would have been fired at Z-310. But the sound of a shot at Z-310 cannot travel to Zapruder’s position and arrive any sooner than Z-315. Conversely, a bullet from the Grassy Knoll fired at Z-312 would have struck JFK at Z-313 and the muzzle blast would arrive at Zapruder’s position (50 ft away) at the same frame. Hence, again, there is close accord between the blur evidence, the acoustical evidence, and the filmed evidence of wounding.

© Mary Ferrell Foundation. All Rights Reserved. |Press Room |Our Policies||Site Map