This packet contains the following files:

BBNstudy.bas - The BBN echo study program written in QBASIC.
BBNstudy.exe - The BBN echo study program compiled with First BASIC.
Dplaza.gif - My computer-drawn map of Dealey Plaza.
BBNstudy.txt - This text file describing the BBN echo study program.

     The BBN echo study program is a test of method which analyzes three test
shots from the test by Bolt, Beranek and Newman on August 10, 1978 in Dealey
Plaza. Mike locations 4, 5, and 6 were chosen because they most frequently
matched DPD tape shots 1 and 2. BBN has refused to release their raw data, so
some of the varibles are my best estimates. For example, we do not know what
the temperature was for each shot, so I have estimated the temperature based
on a ratio of probable time for each shot compared to the range of temperature
over the total time for the shooting test. The location of the test rifle is
based on an analysis of photographs taken of the test shooting and plotting
reasonable trajectories and firing positions. The locations of the test mikes
is based on analysis of photographs taken of the test shooting and inferences
from HSCA testimony about the design of the test. BBN attempted to place the
microphones exactly 18 feet apart in a straight line up Houston Street. My
selection of reflective (refractive?) objects is based partly on the success
of professors Weiss and Aschkenasy in identifying echo generating objects in
Dealey Plaza and the old string method with a calculator to look for echo
generating objects. You will notice that the last impulse should always match
with the Post Office. On most test shots fired from the TSBD window #1 and
picked up by microphones at the top end of Houston Street, there are several
early impulses up to about the point of 400 or 500 milliseconds, then a gap
until the last impulse at about 900 milliseconds. It appears that the corner
of the Post Office is the only echo generating object which can produce a
strong echo in this general area which is then picked up by the test mikes.
Relying on that fact, we can attempt to change the other variables enough to
allow a perfect match between the last echo and the predicted echo pattern
from the Post Office. This is an additional confirmation of method.

     This BASIC program was converted to QBASIC for the IBM from an old
Commodore BASIC program which I wrote many years ago. Not every feature
could be translated elegantly. And there are some flaws and uncertainties
which make it impossible to achieve a perfect match between the BBN test shot
impulses and predicted echo patterns. There are several variables for which
we do not know the true values with absolute certainty. We can only use a 
reasonable estimate. I have provided default values which are assumed to be
very close to the true values for some of these variables, such as the
position of the rifle, the position of the mikes and the speed of sound.
Only direct path echoes are calculated, not compound echo paths as Weiss
and Aschkenasy were able to do. If anyone has suggestions on how to add this
type of calculation, please E-MAIL me the code. Also, you may notice that
there are some differences between the QBASIC program and the compiled BASIC.
First BASIC was the only shareware QBASIC compiler I could find, so if you
can compile it with something better, please do so and E-MAIL it to me.

     Most of the screen output and printout shoud be self-explanatory, but
there is one field which is not explained. The field just before the word
"**BLAST**" is a sum-of-squares calculation which I use as a visual check 
to see if the matching process is getting better on subsequent runs. Since
the predicted blast will always be 0.0 due to the fact that this begins
the matching process, this field did not need to be displayed and could be
used to keep a running check on the sum-of-squares calculation. Any value
above 20 shows poor matching. Any value closer to 10 would indicate a much
better match. Anything near 5 would be an excellent match. If you get results
with a sum of squares less than 10, please E-MAIL me at amarsh@quik.com
with the variables you used.