Understanding the HSCA's Acoustical Evidence of Multiple Gunmen

[Michael T. Griffith]

Confusion and misinformation abound regarding the HSCA's acoustical evidence. The HSCA's six acoustical experts determined that the Dallas police dictabelt contains at least four impulses caused by gunfire in Dealey Plaza, and that one of the shots came from the grassy knoll. Most researchers know this much, but many don't know the events that led up to this historic finding.

On May 12, 1978, the HSCA asked the scientific firm BBN to analyze the Dallas police dictabelt and determine if it contained gunshots. The Committee chose BBN because the firm specialized in acoustical analysis and performed such work as locating submarines by analyzing underwater sound impulses. BBN also pioneered the technique of using sound recordings to determine the timing and direction of gunfire from a tape that was recorded during the 1970 shootings at Kent State University.

Dr. James E. Barger, the lead scientist at BBN, doubted that the tape contained gunfire based on a preliminary examination of the Channel One recording thought by researcher Gary Mack to contain gunshots. Barger determined that the recording, to the human ear, contained no audible sounds of gunfire, contrary to Mack’s claim that seven shots could be heard. Barger told the HSCA that he was “not hopeful about the prospects of recovering anything from the tape.”

Indeed, Barger agreed to clean up the recording and run a series of tests on it in order to prove there were no gunshots on the tape. He reasoned that if preliminary tests showed that no gunshots were on the tape, there would be no need to conduct additional acoustical tests in Dallas, saving the HSCA time and money.

This is important to keep in mind: Far from going into their analysis with a bias toward finding gunfire on the tape, the BBN acoustical scientists believed their preliminary tests would prove the dictabelt contained no gunfire. They thought they were doing the tests in order to spare the HSCA from wasting time and money on acoustical testing in Dealey Plaza.

Two months later, on July 13, 1978, Dr. Barger called the HSCA and dropped a bombshell: he reported that the preliminary tests found evidence of three to five gunshots on the dictabelt.

Barger told the Committee that a more sophisticated test was required to determine if any of the impulse sounds were actually gunfire. He suggested that BBN be allowed to conduct an acoustical reconstruction in Dealey Plaza during which test shots would be fired from locations suggested by eyewitness accounts and compared with the impulse sounds on the police dictabelt. If any of these acoustical fingerprints matched the impulse sounds on the recording, BBN would be able to determine the timing of the shots, the location of the gunman, and the target for each shot fired.

The major points of the acoustical evidence are as follows:

* At least four sets of gunshot impulse patterns with echo patterns unique to Dealey Plaza occur on the dictabelt recording. In other words, their acoustical fingerprints match those of some of the test shots fired in Dealey Plaza.

* Those echo patterns occur in the correct topographic order, which is an amazing correlation all by itself.

* Remarkable locational-movement correlations were found between the dictabelt gunshots and the test-firing gunshots. The BBN scientists determined that the probability that chance caused these correlations was “less than 1%.” Even the NRC/NAS/Ramsey panel admitted that their own calculations showed there was a 93% probability that the correlations were not the result of chance--a fact that is always ignored by critics who cite the panel's findings.

* The dictabelt contains N-waves from supersonic rifle fire, and those N-waves occur only among the identified gunshot impulse patterns, and only in the two impulse patterns that were recorded when the motorcycle’s microphone was in position to record them--yet another impressive correlation.

* Remarkably, the dictabelt not only contains N-waves but it also contains muzzle blasts and muzzle-blast echoes, and those N-waves, muzzle blasts, and muzzle-blast echoes occur in the correct order and interval, proving beyond any rational doubt that the dictabelt contains gunshot impulses.

* Windshield distortions occur in the dictabelt's gunshot impulse patterns when they should occur and do not occur when they should not, another stunning correlation. The HSCA arranged for a separate test to confirm the science and effects of windshield distortion.

* Mark Weiss and Ernest Aschkenasy, acoustical scientists from Queens College who specialized in processing acoustical signals for military applications, determined that gunshot impulse 144.9 (initially timed at 145.1) came from the grassy knoll. Weiss and Aschkenasy calculated there was no more than a 5.3% probability (P=0.053) that the 144.9 impulse pattern was not caused by gunfire, and they argued that the probability was likely lower than that. This is why they reported there was a 95% probability or higher that this shot came from the grassy knoll.

In another oblique admission, the NRC/NAS/Ramsey panel reported that their analysis found that the probability that the 144.9 impulse pattern was not gunfire from the knoll was 22.3% (P=0.223), which means their analysis found that the probability that the impulse pattern was caused by gunfire from the knoll was 77.7%, another fact that is always ignored by critics who cite the panel's findings.

Moreover, former USDA research scientist Dr. Donald Thomas has proved that the NRC/NAS/Ramsey panel committed crucial errors in reaching their P=0.223 calculation, and that the probability that the 144.9 impulse pattern was caused by grassy knoll gunfire is virtually 100 percent (https://www.maryferrell.org/pages/Essay_-_Acoustics_Overview_and_History_-_part_2.html).

Incidentally, a major reason that Weiss and Aschkenasy were selected to do a refined analysis of the BBN findings was that they were recommended by the Acoustical Society of America.

However, some conspiracy theorists reject, minimize, or ignore the HSCA acoustical evidence only because they believe it does not support their shooting scenario, even though it refutes the lone-gunman theory. They think the acoustical evidence requires them to believe that only four shots were fired, that three of those shots came from the TSBD's sixth-floor window, and that only one shot came from the grassy knoll (and that it missed).

However, the acoustical evidence does not require us to accept these positions. It is crucial to understand that the four impulse patterns that the HSCA acoustical experts identified as shots on the police tape were identified as gunfire because they matched the impulse patterns of four shots fired during the field test in Dealey Plaza, and that during that test, shots were fired from only two locations: the sixth-floor window and the grassy knoll.

During the HSCA test firing, no shots were fired from any window in the Dal-Tex Building or the County Records Building, or from any other alternative point in the plaza, but only from the sixth-floor window and from a spot behind the picket fence on the grassy knoll.

In addition, a fifth impulse pattern on the dictabelt was arguably wrongly rejected as gunfire. Due to pressure from the HSCA, the BBN experts identified the impulse pattern at 140.3 as a false alarm, even though it passed the echo-delay matching test, and even though 8 of its 10 impulses matched the impulses of one of the test shots.

The BBN experts said they rejected the 140.3 impulse pattern as a shot because it occurs just 1.07 seconds after the second shot from the sixth-floor window, and because it "only" had one microphone match and "only" had a coefficient score of 0.6. A coefficient score of 0.6 means that 8 of the 10 impulses in the impulse pattern matched the impulses of a pattern of one of the field-test shots. The rejection of the 140.3 impulse pattern as gunfire seems to have been based on something other than science.

Furthermore, if there was a third gunman and he used a silencer or fired from a point five or six feet behind a window, even if the HSCA field test had fired shots from his location or from a nearby location, either his shots would have created no impulse patterns on the tape or they would have created patterns that were too weak to be matched to test-shot patterns.

For more information on the HSCA acoustical evidence, see the following:

"The HSCA's Acoustical Evidence: Proof of a Second Gunman"
https://drive.google.com/file/d/1KvdvH8gTqFgMn-2vTI5ppg_egWxRKg9U/view

"Overview and History of the Acoustical Evidence in the Kennedy Assassination Case" (a three-part article)
https://www.maryferrell.org/pages/Essay_-_Acoustics_Overview_and_History.html
https://www.maryferrell.org/pages/Essay_-_Acoustics_Overview_and_History_-_part_2.html
https://www.maryferrell.org/pages/Essay_-_Acoustics_Overview_and_History_-_part_3.html

"The Bike with the Mike"
https://www.maryferrell.org/showDoc.html?docId=99545

"Debugging Bugliosi"
https://www.maryferrell.org/showDoc.html?docId=58061

[Steve Barber]

 There are no gunshots on channel 1.

 There is no "carillon bell" on channel 1 .

 There is no N.T. Fisher "crosstalk" on channel 1.

 You don't understand anything regarding the acoustics, and neither did Don Thomas:

 https://jfkfiles.blogspot.com/2007/07/of-crosstalk-and-bells-rebuttal-to-don.html

 Written in late 1999:  https://www.jfk-assassination.net/barber.htm

[Michael T. Griffith]

There are no gunshots on channel 1.

 There is no "carillon bell" on channel 1 .

 There is no N.T. Fisher "crosstalk" on channel 1.

 You don't understand anything regarding the acoustics, and neither did Don Thomas:

 https://jfkfiles.blogspot.com/2007/07/of-crosstalk-and-bells-rebuttal-to-don.html

 Written in late 1999:  https://www.jfk-assassination.net/barber.htm

No, you're the one who doesn't understand anything about the acoustics. Every one of your time-worn arguments is answered in my article on the acoustical evidence. Dr. Thomas also answers your arguments in his article. Apparently, you didn't bother to read any of the cited articles before replying.

Your 1999 article was obsolete and erroneous in 1999--it is even more so now. Your article says absolutely nothing--not one word--about the windshield distortion correlations, the time-movement correlations, the presence of N-waves and the resulting echoes that occur in the correct order and timing, the Ramsey panel's admission that there is a 77% probability that the impulse identified as a grassy knoll shot is indeed a gunshot from the knoll, the Ramsey panel's admission that there is a 93% probability that the time-movement correlations are not the result of chance, etc., etc.

Your article says nothing about any of this evidence, yet you pretend that it explains the acoustical evidence. I'm guessing you've never even read the BBN and W&A reports on the acoustical evidence, nor even the HSCA's detailed section on the acoustical evidence in its final report.

Here's a link to the section on the acoustical evidence in the HSCA's final report:

https://www.archives.gov/research/jfk/select-committee-report/part-1b.html

BTW, I take it you're unaware that Alvarez, on whom you rely so heavily, was caught falsifying his test data, right?

[Michael T. Griffith]

Dr. G. Paul Chambers, a physicist who has worked with NASA, with the Naval Surface Warfare Center, and with the Naval Research Laboratory, is unpersuaded by the NRC/NAS/Ramsey panel's attacks on the acoustical evidence:

The National Research Council admitted that its calculations of the relevant
probabilities may have been “unduly conservative.” In fact, the report uses the
word “conservative” with regard to the statistical analysis a number of times. It
then goes on to state: “Except for the rather conservative analysis above, the
data do tend to cast doubt on the hypothesis of random impulse locations.” In
other words, the sounds on the tape do not appear to be caused by standard
random noises, like cars backfiring in Dealey Plaza. This seems to me a lot like
an admission that the Bolt, Beranek, & Newman impulse sequences were indeed
nonrandom data, and therefore attributable to something significant. As gunshots
were the only reasonable source that matched the acoustic signatures, it would
appear that the NRC doubted its own conclusions.

With regard to the crosstalk synchronization, the NRC report states, in an
appendix discussing this analysis, that:

“The analyses in Appendices B-1 and B-2 [of the crosstalk] may be subject to
some criticism. A certain amount of subjectivity derives from the fact that the first
observer was looking at the sound spectrograms from both channels while he
marked points on Channel 1. . . . However, this experiment was supplemented by
several variations that derived similar results. Some of these were more careful
to avoid the subjectivity and to reduce considerably the dependence aspects of
the experiment presented here. These are not reported in detail, because they
were carried out using xerographic copies of photographs using several scales,
and relatively crude measuring instruments (graph paper in place of rulers).”

This is an admission that the first technique used to confirm the crosstalk
between the channels and to determine that the impulses of interest were not
concurrent with the assassination was highly subjective and therefore subject to
doubt. Moreover, the method they used to confirm the first technique was so low-
tech that they didn’t even want to describe it.

I could speculate that they employed the zoom function on the photocopier;
however, precisely how they did this analysis was deliberately left vague and so
there is no hope of reproducing their work. In science, if other scientists are
to take conclusions seriously, it is essential to describe the methodology of what
was done to reach them in detail. Otherwise, the work doesn’t have validity,
because no one will be able to reproduce the results.

Dr. Barger himself responded to the issue of the crosstalk synchronization in a
letter to G. Robert Blakey in 1983, after the release of the NRC report, concluding
that the NRC synchronization was doubtful and that the original BBN conclusions
remained valid. (Chambers, Head Shot, pp. 106-107)

Dr. Chambers discusses some the impressive aspects of the acoustical evidence:

Weiss and Aschkenasy reviewed Barger’s analysis and conclusions. They found
that Barger’s analysis was valid and his conclusions supported by the evidence
on the tape. They concurred with his recommendation to conduct live-fire tests in
Dealey Plaza to determine the origin and direction of the gunshots, and they
approved his plan for acoustical reconstruction. . . .

In Dealey Plaza, the sounds of gunshots would produce similar echoes. When
recorded and captured on a specialized electronic device like an oscilloscope
that converts sound patterns into pictures, these echoes appear as “acoustical
waveforms” and appear as unique signatures of sound-producing events.

In the case of a rifle shot in Dealey Plaza, the acoustical signatures would differ
based on the origin, direction, and velocity of the shot, as well as the location
of the recording microphone. The echo patterns would depend on the timing of
sound reflections off building or other structures and obstructions in the plaza. . . .
A recording was made of the sounds received at each microphone during each
test shot, making a total of 432 recordings of impulse sequences. . . . Each
recorded impulse sequence was then compared with each of the six impulse
patterns on the channel 1 Dictabelt recording to see the degree to which
significant points in each impulse pattern matched. . . .

The time of the arrival of the impulses, or echoes, in each sequence of impulses
was the characteristic being compared, not the shape, amplitude, or any other
characteristic of the impulses or sequences. . . .

When the BBN team performed their analysis of the acoustical waveforms, they
found something extraordinary. When they compared the impulse sequences
from the acoustical reconstruction to the sequences on the original Dictabelt
recording, they found a number of significant matches. When the locations of the
microphones that recorded matches in the reconstruction were plotted on a
graph of time versus distance, it was found that the location of the microphones
that recorded matches were clustered around a line on the graph that was
consistent with the known speed of the motorcade. . . .

Of the thirty-six microphones placed along the motorcade route, the one that
recorded the sequence of impulses that matched the third impulse on the 1963
dispatch tape [the dictabelt tape] was farther along the route than the one that
recorded the impulses that matched the second impulse on the dispatch tape.
The locations of the microphones were consistent with the distance a motorcycle
traveling at about 11 mph would cover in the elapsed time between impulses on
the dispatch tape. . . . Applying a statistical formula, Barger estimated that since
the microphones clustered around a line representing the speed of the
motorcade, there was a 99 percent probability that the Dallas police dispatch
tape did, in fact, contain impulses transmitted by a microphone in the motorcade
in Dealey Plaza during the assassination. . . .

Weiss and Aschkenasy, specialists in sonar applications . . . examined Dealey
Plaza carefully to determine which structures were most likely to have caused the
echoes recorded by the microphone in the acoustical reconstruction that had
exhibited a match to the shot from the grassy knoll. They verified and refined
their identifications of echo-producing structures by examining the results of the
1978 reconstruction [the test firing in Dealey Plaza]. This approach allowed them
to look for matches in the data with a 1 ms [millisecond] correlation. . . . Matches
at this level of temporal precision substantially reduced the possibility that a
match could occur as a result of random noise.

In Dealey Plaza, echoes from gunshot test patterns arrive in two discrete
[different] clusters, differing in time by about 190 ms. Echoes originating from
structures along Elm Street arrive within 85 ms, while echoes from structures
farther back on Houston Street arrive in the last 95 ms of a typical 370-ms-
duration test pattern. I

n addition, a “muzzle blast” is usually prominent at the beginning of a gunshot
acoustical pattern, while an N-wave (a shock wave traveling faster than the
speed of sound due to the rifle bullet exceeding the sound barrier) arrives prior
to the muzzle blast. The waveform identified as the grassy knoll shot is shown
in figure 13. The presence of an N-wave in this waveform was consistent with
the acoustical signature of a supersonic rifle bullet. (Head Shot, pp. 96-102)