The HSCA Acoustical Evidence: Proof of a Second Gunman in the JFK Assassination

[Michael T. Griffith]

Here is the newly revised and expanded version of my 2013 article on the HSCA acoustical evidence:

The HSCA’s Acoustical Evidence: Proof of a Second Gunman in the JFK Assassination
https://miketgriffith.com/files/hscaacous.pdf

Anyone who has read the 2013 version will see that the new version is very different and much more detailed. Here are the section headings in the new version:

Introduction
The NRC Panel
The Motorcycle with the Open Microphone
No “Audible” Shots on the Dictabelt Tape
How Could the Grassy Knoll Shot Have Missed?
Crowd Noise and the Carillon Bell
The Decker “Hold Everything” Crosstalk
Larry Sabato’s Sonalysts Study
The Grassy Knoll Shot and the Zapruder Film
Five Gunshots on the Dictabelt Tape?
A Summary of the Acoustical Evidence
Bibliography

[Joffrey van de Wiel]

Thanks for sharing your article Michael  Its subtitle could be "The acoustical evidence for dummies" as it starts with the basics, it's a great introduction into this complicated matter.

Two quick questions:

1) Would shots fired using silencers show up on the Dictabelt as N-waves?

2) The grassy knoll shot. Dr. David Mantik writes that a second headshot was fired not from the grassy knoll, but from the storm drain on the north overpass. This according to his interpretation of the medical evidence. It can't be both (I assume), so which is it?

[Michael T. Griffith]

Thanks for sharing your article Michael  Its subtitle could be "The acoustical evidence for dummies" as it starts with the basics, it's a great introduction into this complicated matter.

Two quick questions:

1) Would shots fired using silencers show up on the Dictabelt as N-waves?

No. In fact, they would not show up at all, because of their vastly reduced amplitude and  because of the engine noise.

2) The grassy knoll shot. Dr. David Mantik writes that a second headshot was fired not from the grassy knoll, but from the storm drain on the north overpass. This according to his interpretation of the medical evidence. It can't be both (I assume), so which is it?

This is one of the many reasons that the Dealey Plaza test firing should have included shots fired from more than just two locations and should have included many more microphones. Many researchers have noted that one of the impulse patterns might have come from the Dal-Tex Building, where Mafia man Eugene Brading was arrested minutes after the shooting, and if test shots had been fired from there, correlations to one of those testshots might have been found on the police tape.

When we look at the acoustical evidence, we need to constantly keep in mind that the test firing was limited to two firing positions. 

[Michael T. Griffith]

Went straight to the Decker cross talk section and saw that you continue talking through your ass. The tapes are in sync whenever there's cross talk which you still don't get.

Whaaat???!!! LOL! This howler is further proof that you have no business talking about the acoustical evidence. No, the tapes are not in sync whenever this is crosstalk. The crosstalk transmissions don't even synchronize with each other, for crying out loud. How can you not know this? Don't you ever get tired of embarrassing yourself with gaffe after gaffe?

The offset BS from your fake Jesus, Dr. Thomas, can't save you.

How in the world would you know, Mr. "I Can't Understand Herb Blenner"? No credible student of the acoustical evidence denies that there are offsets between the crosstalk on the two channels. The debate is over the cause of the offsets.

I guess it did not occur to you that the offsets disprove your claim that the channels synchronize whenever there is crosstalk.

[Joe Elliott]

The BBN started with the noblest of intentions and ended as a farce.

It was thought, that analyzing the Dictabelt tape, they could deduce:

1.   The position of the shooter.
2.   The position of where the shot was aimed at, that is, where it stuck.
3.   The position of the motorcycle with the stuck microphone.

Going into this, I sure Dr. Barger had hopes of getting excellent data. And if this excellent data showed the motorcycle was elsewhere, or was at Dealey Plaza but recorded three shots from the TSBD, that would be a disappointment. But, if he discovered multiple firing locations, it would be the greatest scientific crime discovery of the century. I am certain his hopes were up.

So, they ran some firing tests in Dealey Plaza on August 20, 1978, and recorded the shots on 36 microphones arranged along Houston and Elm Street.

Below is a simplified version of BBN’s Exhibit F-367, which shows the correlations that they discovered.

Test	Beginning Time of	Zap.	Rifle	Target
ID	First impulse on	Frame	Location	Location
	Tape Segments (sec)	(Thomas)
A	136.20
B	137.70	176	TSBD	Location-z 155
D	137.70	176	TSBD	Location-z 313
E	137.70	176	KNOLL	Location- Tague
G	139.27	205	TSBD	Location-z 313
I	139.27	205	TSBD	Location-z 313
J	139.27	205	KNOLL	Location-z 313
K	140.32	224	TSBD	Location-z 313
L	145.15	313	KNOLL	Location-z 313
M	145.15	313	TSBD	Location-z 224
N	145.15	313	TSBD	Location-z 313
O	145.61	321	TSBD	Location-z 313
P	145.61	321	TSBD	Location- Tague
Q	145.61	321	TSBD	Location-z 224
R	146.30

The “Zap. Frames” are Dr. Thomas’s estimate, not the original BBN estimate. They don’t differ by enough to make much of a difference so to avoid confusion, I just went with Dr. Thomas’s time estimate.

On this table, I don’t differentiate between the rifle in the TSBD being 2 feet behind the plane of the window, or sticking out further. It didn’t seem to make a lot of difference in the firing tests. So, I just consider these shots to be from the TSBD.

In the August 20, 1978 firing tests, they fired at four locations. Target 1, 2, 3 and 4.
Target 1 is about where the limousine was at the location shown in Zapruder frame z155.
Target 2 is about where the limousine was at the location shown in Zapruder frame z224.
Target 3 is about where the limousine was at the location shown in Zapruder frame z313.
Target 4 is about Mr. Tague was standing, about 240 feet beyond the z313 location.
So instead of listing the “Target Location” as “3”, I give it as “Location-z 313”.



So, how did they do?

1.   The position of the shooter.

Terrible.
Shot 1 at z176 - They got correlations found for both the TSBD and the Grassy Knoll. There should have only been a correlation for one location. One of these results must be bad. The degree of correlation was not high enough,
Shot 2 at z205 – Again, bad, correlations found for both the TSBD and the Grassy Knoll.
Shot 3 at z224 – One correlation for a shot from the TSBD. Getting just one correlation is actually good. It’s too bad they didn’t get this for all 5 shots.
Shot 4 at z304 – Again, bad, correlations found for both the TSBD and the Grassy Knoll.
Shot 5 at z313 – Correlations only for a shot from the TSBD. This is a good result.



2.   The position of where the shot was aimed at, that is, where it stuck.
Shot 1 at z176 – the shot struck at:
Location-z 155 – miss by 19 feet.
Location-z 313 – miss by 132 feet.
Location- Tague – miss by 272 feet.

Shot 2 at z205 – the shot struck at:
Location-z 313 – miss by 92 feet.

Shot 3 at z224 – the shot struck at:
Location-z 313 – miss by 75 feet.

Shot 4 at z304 – the shot struck at:
Location-z 224 – miss by 70 feet.
Location-z 313 – a hit on the limousine

Shot 5 at z313 – the shot struck at:
Location-z 224 – miss by 75 feet.
Location-z 313 – a hit on the limousine
Location- Tague – miss by 240 feet.



One the whole, terrible. The data contradicts itself. Clearly false correlations are more common than true.

Dr. Barger doesn’t seem the least bit phase. When the results contradict themselves, we should just assume we got some “false alarms”. We should just ignore the ones that don’t make sense and consider the ones that do reliable. These “false alarms” are a serious problem. They are more properly known as “false positives”. So many false positives in the data, where false positives are more common then true positives, bring all the true positives into doubt.

And if we follow Dr. Barger, throw out all the “false alarms”, what proposed shots have good support:

Shot 1 – at z152 – good support (a miss by 19 feet is good enough), if the two false positives are ignored.
Shot 2 – at z205 – no support, all correlations give unbelievable “Target location” that is off by 92 feet.
Shot 3 – at z224 – no support, the correlation gives unbelievable “Target location” that is off by 75 feet.
Shot 4 – at z304 – good support, if the two false positives are ignored.
Shot 5 – at z313 – good support, if the two false positives are ignored.

Two of the five “found” shots, strictly speaking, should be thrown out, because misses by 70 feet or more is just too unbelievable.


3.   The position of the motorcycle with the stuck microphone.

Here we have much better correlations. These correlations are what Mr. Griffith really likes to emphasis.

The data is consistent with a motorcycle traveling at around 11 mph, trailing behind the limousine by 120 to 160 feet. Which should plausible. But there is a problem with the film evidence. The Hughes film shows the only possible officer, Officer McLain, along with his partner, Officer Baker, trailing behind the limousine by 300 feet. By the time the film shuts off, Officer McLain has 1.5 seconds to cover 170 feet, impossible. Also, he should appear in the Altgens photograph.

But let’s ignore all that. We still, seemingly, have a remarkable correlation with the data and a plausible set of positions of the motorcycle. What are we to make of this data as a whole? Random looking correlations for the position of the shooter and the area of the street the bullets struck. Good correlation for the position of the motorcycle. This is truly hot ice and wondrous strange snow.

But there is a simple explanation. For a through check of all possible correlations:
432 strip charts from the 1978 firing tests, 36 microphones times 12 test shots
Need to be each compared with the 6 possible shots from the 1963 Dictabelt recording.
So, 2,592 comparisons need to be made between pairs of strip charts for the 1978 tests and the 1963 recording.

This sounds difficult to do in 15 days, to get done in time to present the data to the HSCA. They have to:

Decide which of the 1978 78 test shots are to be compared, they selected 12 of them.
Print out 12 strip charts for them.
Print out 432 strip charts for the 1978 data.
Make all 2,592 comparisons, which involve counting waves within a certain number of milliseconds and doing some calculations to get a “correlation coefficient”
Organize all the data and print them out in a neat form to present to the HSCA.

I would suggest that maybe they did not perform all 2,592 combinations of comparisons. From one shot, it seemed the motorcycle was 150 feet behind. With the time limit, it made sense to only search for where the motorcycle might be. Why search the data from a stretch of street that you probably wouldn’t find a motorcycle. And if a correlation was found about where expected, 150 feet behind, under the time pressure, it’s time to move on to the next shot.

If something like this happened, all the correlations found would be consistent with a motorcycle averaging 11 mph, 120 to 160 feet behind the limousine. Even if a through processing of the entire data set may have found many contradictions on the location of the motorcycle, just as was found with the location of the shooter and the location of the target.

Until proven otherwise, I hold that this is the most likely scenario that explains this “hot ice and wondrous strange snow” results.

Dr. Barger’s fundamental error, is that he allowed his hopes to affect his judgement. These “false alarms”, should have set off a real alarm in his head. It should have been apparent to him that he was measuring “noise”, garbage data that contradicted itself which was what data is that contains too many false positives. This work will be his legacy, it will overshadow everything else he was done in his career. He has gone from being a respected scientist to having an insect expert as his main ally.

[Joe Elliott]

This is one of the many reasons that the Dealey Plaza test firing should have included shots fired from more than just two locations and should have included many more microphones. Many researchers have noted that one of the impulse patterns might have come from the Dal-Tex Building, where Mafia man Eugene Brading was arrested minutes after the shooting, and if test shots had been fired from there, correlations to one of those testshots might have been found on the police tape.

When we look at the acoustical evidence, we need to constantly keep in mind that the test firing was limited to two firing positions.

What do you do if these expanded firing tests find that the shot at 139.27 (z205) came from the TSBD, the Grassy Knoll and the Dal-Tex building?

Which of the three “confirmed” firing locations should be considered accurate for that shot?

[Joe Elliott]

2) The grassy knoll shot. Dr. David Mantik writes that a second headshot was fired not from the grassy knoll, but from the storm drain on the north overpass. This according to his interpretation of the medical evidence. It can't be both (I assume), so which is it?

This is one of the many reasons that the Dealey Plaza test firing should have included shots fired from more than just two locations and should have included many more microphones. Many researchers have noted that one of the impulse patterns might have come from the Dal-Tex Building, where Mafia man Eugene Brading was arrested minutes after the shooting, and if test shots had been fired from there, correlations to one of those testshots might have been found on the police tape.

When we look at the acoustical evidence, we need to constantly keep in mind that the test firing was limited to two firing positions.

Yes. The BBN confirmed that the shot at z304 (their estimate of time) came from the grassy knoll. And Acoustics analysts Mark Weiss and Ernest Aschkenasy confirmed that there was a 95% chance that the shot from the grassy knoll. But with a shooting tests from the sewer, this may have confirmed that the z304 shot did indeed come from the TSBD, and the grassy knoll, and from the sewer.

I don’t think they needed more shooting locations. I think they needed less, to reduce the number of these pesky false positives.

[Michael T. Griffith]

The BBN started with the noblest of intentions and ended as a farce.

It was thought, that analyzing the Dictabelt tape, they could deduce:

1.   The position of the shooter.
2.   The position of where the shot was aimed at, that is, where it stuck.
3.   The position of the motorcycle with the stuck microphone.

Going into this, I sure Dr. Barger had hopes of getting excellent data. And if this excellent data showed the motorcycle was elsewhere, or was at Dealey Plaza but recorded three shots from the TSBD, that would be a disappointment. But, if he discovered multiple firing locations, it would be the greatest scientific crime discovery of the century. I am certain his hopes were up.

So, they ran some firing tests in Dealey Plaza on August 27, 1978, and recorded the shots on 36 microphones arranged along Houston and Elm Street.

Below is a simplified version of BBN’s Exhibit F-367, which shows the correlations that they discovered.

Test	Beginning Time of	Zap.	Rifle	Target
ID	First impulse on	Frame	Location	Location
	Tape Segments (sec)	(Thomas)
A	136.20
B	137.70	176	TSBD	Location-z 155
D	137.70	176	TSBD	Location-z 313
E	137.70	176	KNOLL	Location- Tague
G	139.27	205	TSBD	Location-z 313
I	139.27	205	TSBD	Location-z 313
J	139.27	205	KNOLL	Location-z 313
K	140.32	224	TSBD	Location-z 313
L	145.15	313	KNOLL	Location-z 313
M	145.15	313	TSBD	Location-z 224
N	145.15	313	TSBD	Location-z 313
O	145.61	321	TSBD	Location-z 313
P	145.61	321	TSBD	Location- Tague
Q	145.61	321	TSBD	Location-z 224
R	146.30

The “Zap. Frames” are Dr. Thomas’s estimate, not the original BBN estimate. They don’t differ by enough to make much of a difference so to avoid confusion, I just went with Dr. Thomas’s time estimate.

On this table, I don’t differentiate between the rifle in the TSBD being 2 feet behind the plane of the window, or sticking out further. It didn’t seem to make a lot of difference in the firing tests. So, I just consider these shots to be from the TSBD.

In the August 27, 1978 firing tests, they fired at four locations. Target 1, 2, 3 and 4.
Target 1 is about where the limousine was at the location shown in Zapruder frame z155.
Target 2 is about where the limousine was at the location shown in Zapruder frame z224.
Target 3 is about where the limousine was at the location shown in Zapruder frame z313.
Target 4 is about Mr. Tague was standing, about 240 feet beyond the z313 location.
So instead of listing the “Target Location” as “3”, I give it as “Location-z 313”.

So, how did they do?

1.   The position of the shooter.

Terrible.
Shot 1 at z176 - They got correlations found for both the TSBD and the Grassy Knoll. There should have only been a correlation for one location. One of these results must be bad. The degree of correlation was not high enough,
Shot 2 at z205 – Again, bad, correlations found for both the TSBD and the Grassy Knoll.
Shot 3 at z224 – One correlation for a shot from the TSBD. Getting just one correlation is actually good. It’s too bad they didn’t get this for all 5 shots.
Shot 4 at z304 – Again, bad, correlations found for both the TSBD and the Grassy Knoll.
Shot 5 at z313 – Correlations only for a shot from the TSBD. This is a good result.

2.   The position of where the shot was aimed at, that is, where it stuck.
Shot 1 at z176 – the shot struck at:
Location-z 155 – miss by 19 feet.
Location-z 313 – miss by 132 feet.
Location- Tague – miss by 272 feet.

Shot 2 at z205 – the shot struck at:
Location-z 313 – miss by 92 feet.

Shot 3 at z224 – the shot struck at:
Location-z 313 – miss by 75 feet.

Shot 4 at z304 – the shot struck at:
Location-z 224 – miss by 70 feet.
Location-z 313 – a hit on the limousine

Shot 5 at z313 – the shot struck at:
Location-z 224 – miss by 75 feet.
Location-z 313 – a hit on the limousine
Location- Tague – miss by 240 feet.

One the whole, terrible. The data contradicts itself. Clearly false correlations are more common than true.

Dr. Barger doesn’t seem the least bit phase. When the results contradict themselves, we should just assume we got some “false alarms”. We should just ignore the ones that don’t make sense and consider the ones that do reliable. These “false alarms” are a serious problem. They are more properly known as “false positives”. So many false positives in the data, where false positives are more common then true positives, bring all the true positives into doubt.

And if we follow Dr. Barger, throw out all the “false alarms”, what proposed shots have good support:

Shot 1 – at z152 – good support (a miss by 19 feet is good enough), if the two false positives are ignored.
Shot 2 – at z205 – no support, all correlations give unbelievable “Target location” that is off by 92 feet.
Shot 3 – at z224 – no support, the correlation gives unbelievable “Target location” that is off by 75 feet.
Shot 4 – at z304 – good support, if the two false positives are ignored.
Shot 5 – at z313 – good support, if the two false positives are ignored.

Two of the five “found” shots, strictly speaking, should be thrown out, because misses by 70 feet or more is just too unbelievable.

3.   The position of the motorcycle with the stuck microphone.

Here we have much better correlations. These correlations are what Mr. Griffith really likes to emphasis.

The data is consistent with a motorcycle traveling at around 11 mph, trailing behind the limousine by 120 to 160 feet. Which should plausible. But there is a problem with the film evidence. The Hughes film shows the only possible officer, Officer McLain, along with his partner, Officer Baker, trailing behind the limousine by 300 feet. By the time the film shuts off, Officer McLain has 1.5 seconds to cover 170 feet, impossible. Also, he should appear in the Altgens photograph.

But let’s ignore all that. We still, seemingly, have a remarkable correlation with the data and a plausible set of positions of the motorcycle. What are we to make of this data as a whole? Random looking correlations for the position of the shooter and the area of the street the bullets struck. Good correlation for the position of the motorcycle. This is truly hot ice and wondrous strange snow.

But there is a simple explanation. For a through check of all possible correlations:
432 strip charts from the 1978 firing tests, 36 microphones times 12 test shots
Need to be each compared with the 6 possible shots from the 1963 Dictabelt recording.
So, 2,592 comparisons need to be made between pairs of strip charts for the 1978 tests and the 1963 recording.

This sounds difficult to do in 15 days, to get done in time to present the data to the HSCA. They have to:

Decide which of the 1978 78 test shots are to be compared, they selected 12 of them.
Print out 12 strip charts for them.
Print out 432 strip charts for the 1978 data.
Make all 2,592 comparisons, which involve counting waves within a certain number of milliseconds and doing some calculations to get a “correlation coefficient”
Organize all the data and print them out in a neat form to present to the HSCA.

I would suggest that maybe they did not perform all 2,592 combinations of comparisons. From one shot, it seemed the motorcycle was 150 feet behind. With the time limit, it made sense to only search for where the motorcycle might be. Why search the data from a stretch of street that you probably wouldn’t find a motorcycle. And if a correlation was found about where expected, 150 feet behind, under the time pressure, it’s time to move on to the next shot.

If something like this happened, all the correlations found would be consistent with a motorcycle averaging 11 mph, 120 to 160 feet behind the limousine. Even if a through processing of the entire data set may have found many contradictions on the location of the motorcycle, just as was found with the location of the shooter and the location of the target.

Until proven otherwise, I hold that this is the most likely scenario that explains this “hot ice and wondrous strange snow” results.

Dr. Barger’s fundamental error, is that he allowed his hopes to affect his judgement. These “false alarms”, should have set off a real alarm in his head. It should have been apparent to him that he was measuring “noise”, garbage data that contradicted itself which was what data is that contains too many false positives. This work will be his legacy, it will overshadow everything else he was done in his career.

Just about every single statement in your reply is either false, irrelevant, or based on a misreading/misrepresentation of the HSCA materials due to your ignorance on the subject. You've already admitted that you refuse to read Dr. Thomas's book, and it's clear that you still have not read the BBN report (if you have, one wonders how on earth you could say the things you say). How about Dr. Chambers' chapter on the acoustical evidence?

Let's just take your last paragraph, the one about false alarms. BBN and WA developed very reliable tests to distinguish between the false alarms and the gunshot impulse patterns to a degree of certainty of well over 95%. Dr. Weiss explained to one congressman who asked him if the third and fourth gunshot impulses could be "acoustical mirages" of one shot that this was impossible because the third shot--the grassy knoll shot--contained specific echo-pattern characteristics of a test-firing shot from the grassy knoll, because each echo has "its own peculiar distortion, transmission characteristics":

Mr. WEISS. No, sir; because in order for that to be true, you would have to, in effect, have had the sound of the muzzle blast transported by some means to the
location of the grassy knoll area, and there emitted as if it had originated from that point. Since every echo that was predicted corresponded to an echo arising from a sound rising from that location [in the test firing], what you would have required is that echoes otherwise generated from a shot fired, say, from the depository window, would each have had its own peculiar distortion, transmission characteristics such that by some marvelous process it occurred at the microphone, intact, and at the correct position. (5 HSCA 608)

People who understand the acoustical evidence will realize that you are still just posting paragraphs of diversionary smoke and are still avoiding the powerful, intricate correlations between the dictabelt shots and the test-firing shots. If all those gunshot impulses are merely false alarms, then you need to explain the correlations, unless you're merely going to argue that they are all amazing, staggering coincidences.

You can't explain the correlations because there is no Web article that deals with those correlations, and so you have no source from which to copy and paste to address them. That's why you keep avoiding them.

How about if you just explain one of the correlations in the acoustical evidence: the windshield-distortion correlations? The HSCA scientists realized that soundwaves passing through a motorcycle windshield would experience distortion. So they ran tests to determine the characteristics of windshield distortion. And, lo and behold, they found that the dictabelt shots that match test shots from the Book Depository all contained windshield distortions but that the dictabelt shot that match test shots from the grassy knoll contain no such distortions. My, my, my!

Try to fathom the odds that three out of four dictabelt gunshot impulses would just happen to contain windshield distortions in the first place, and then try to fathom the odds that the one gunshot impulse that should not contain those distortions would not contain them.

He has gone from being a respected scientist to having an insect expert as his main ally.

This petty, juvenile, and erroneous comment shows why you are not credible, why you are not to be taken seriously. First of all, all of the BBN acoustical scientists stood behind the acoustical evidence, not just Dr. Barger. So did Weiss and Aschkenasy. Second, your continued harping on the fact that Dr. Thomas is an entomologist ignores the fact that he is also an expert in statistics, and that Dr. Chambers, a world-renowned physicist, and Dr. Scheim, a graduate of MIT in mathematics, also support the acoustical evidence. You don't mind citing a urologist on ballistics and forensic issues, but you object when an entomologist is cited on the acoustical evidence, even though his acoustical research has been praised by Dr. Barger and other scientists.