CIA-RDP96-00792R000701010001-8

March 1984
PSI RESEARCH
March 1984
PSI RESEARCH
Approved For Release 2000/08/15 : CIA-RDP96-00792R000701010001-8
SOME FINDINGS RELATING TO THE ELECTRONIC
VOICE PHENOMENON
Alexander MacRae
Skye, Scotland
.
Born in Skye, Scotland, MacRae studied electronics
in London, for a time under Sir Geoffrey
Housfield, FRS, joint Nobel laureate. Working at
Palo Alto, he designed communications systems for
NASA - one design being used as recently as the
first Space Shuttle. He currently lectures in
Scotland on Information Technology, and is a
director of a small electronics company. His
studies of "paranormal" voices on tapes represent
a scientific approach to one of most controversial
phenomena of our time. - Ed.
Introduction
This paper relates some recent findings in EVP by the
author, and suggests that the Electronic Voice Phenomenon
(or paranormal voices) is a psi phenomenon which - with
current equipment - can be produced frequently and
predictably by anyone trained in its use. Because of the
availability of results it has been possible to come to some
preliminary conclusions, and these are the subject of this
paper.
EVP - the electronic voice phenomenon - first came to
my attention in 1979, and having read about it, I decided to
try some experiments - but without any noticeable success.
In 1982 I got in touch with some of the few remaining EVP
investigators left in the UK.
The "field" seemed to proceed in the most unscientific
manner, nothing was ever measured, although the words
"research" and "expert" were bandied around like tokens in a
game of "let's play scientists." Instead of measurement,
judgment by recourse to reputation was the rule, indeed
"reputation" was the name of the game, it seems. There was
a sort of subjective scale for rating the "quality" of
utterances, although it is doubtful if it was actually
quality in the Hi-Fi senses of bandwidth, and absence of
distortion, that was being assessed - instead there was a
vague idea of signal/noise ration involved.
One of my first suggestions was to try to assess the
probability of recording an utterance in a given time
period, from which one could then derive a measure for the
quantity of "communication" being produced, by which various
systems - then being adjudicated, sometimes quite viciously,
by reference to reputation - could be objectively evaluated.
I suggested also that simply establishing that there was a
correlation between a stimulus (a question from the
experimenter) and a presumed response, was sufficient
evidence - if replicable - of a phenomenon. But all this
fell on deaf ears.
I virtually abandoned the whole thing, but one factor
kept nagging me. I had observed that almost all EVP
utterances had a duration of between 0.7 and 2.2 seconds,
with a pronounced probability peak around 1.75 secs.
Now this was against nature. If the phenomenon was
purely a random event, a chance recording of bits of
telephone conversations, radio plays, or the like - then the
spread of durations for the utterances should show this
randomness. There is no good reason why an utterance should
lie in this slot around 1.75 secs. Various time constants
were considered - but none of them could provide a plausible
explanation for all circumstances.
It seemed to be a definite phenomenon, extending even
to longer statements, which might consist of one 1.75 sec.
segment, then a pause, finally another 1.75 sec. segment.
When all looked black, I clung to this one little bit of
objective truth, ignored by the "experts."
In early 1983 I started producing my own EVP, using a
piece of equipment designed for another purpose. The
"quality" was very poor, but results were being obtained
predictably. At first I could not believe it and felt sure
it was CB breakthrough, or ship-to-shore radio, or somesuch.
But these were gradually eliminated. The first voice
recorded used a characteristic phrase of my late father - a
factor suspicious enough to make me very careful of my own
Approved For Release 2000/08/15 : CIA-RDP96-00792R000701010001-8
37

March 1984
judgment, in case desire for a result should lead to
un justifiable approximations.
Following my own rule, which I had tried to convey to
the EVP field, I ignored what was said and simply counted
the number of voice-like utterances per unit time.
Initially this was running at 0.05 utterances per minute
(U/M). Within a few weeks it had improved by a factor of
10, to about 0.5 U/M. By March of 1983 it peaked around 2
U/M, and with that the next surprising development occurred.
This was a direct response. One afternoon I was about to
explain who I was, and began with the rhetorical question,
"who am I. . . " The immediate response was "voice radio
operator." In succeeding experiments other direct responses
were obtained, on one occasion I found I was the responder!
Voice:
"This is the voice of Cass Evitt"
Me:
"Say again"
Voice:
"This is Sugar Roll's voice"
Me:
"Who?"
Voice:
"Sugar Roll's - voice"
(Later)
"What is that"
(Later)
"How can I communicate to you better?"
Voice: "Just talk."
1. A General Phenomenon.
Approved For Release 2000/08/15 : CIA-RDP96-00792R000701010001-8
Me:
"It is a - (thinking desperately for a suitable
term) - a 'Voice-Radio'."
This was also the most prolonged sequence obtained. By this
time my own research was brought to a halt. A growing
demand for units, and an extensive trials program to prove
that I was not the only person able to use the equipment,
meant that all my units were out in the field, and anyway I
did not have sufficient time to conduct research - hazards
of the one-man, underfunded research efforts. Nevertheless
the following preliminary findings are offered.
The results obtained by others, in other locations, at
other times, using the equipment, are at least as good as,
and sometimes better than, my own.
38
March 1984
PSL RESEARCH
2.
Results not due to Radio Broadcasts:
2.1.
Utterances generally fall into the 1-2 seconds
time slot - an improbable event.
2.2.
A large percentage of utterances contain names -
unlikely if these were random snatches of radio plays and
the like.
2.3. A significant percentage of utterances contain
one's own name.
2.4. The number of responses is in some way propor-
tional to the number of stimuli (questions, requests from!
the experimenter).
2.5.
Different experimenters at different times and
places pick up, on occasion, identical phrases.
2.6.
Some utterances use a non-standard form of
English - a sort of "slick-talk," or slang, that creeps
in
now and then.
2.7. In some utterances, unnaturally prolonged vowelo
sounds occur - with a greater frequency than might be
expected if radio plays were their source.
2.8. Synthetic voicing occurs too frequently to be
natural.
2.9.
A percentage of utterances consist of relevant.
comment, or direct response.
2.10.
Words such as "voice" and "message" occur with
abnormal frequency.
3. Speech Formats generally are of a non-Glottal Typed
Speech forms which have been recorded include the
following:
(a)
Natural voiced speech - infrequent.
(b) Synthetic voiced speech - more frequent - this breaks
down into two categories,
(i) A type in which the synthetic voicing is relative
ly periodic, giving the "robot-like"
OLL
"Mechanical" sound of present day computer speech
(ii) A type in which the synthetic voicing is relative
ly random, but within frequency constraints, gives
a sound much like hoarse whispering, with the
random low frequency perturbations giving thes
utterances an unwarranted sinister aspect.
39
Approved For Release 2000/08715 "CIA-RDP96-00792R000701010001-8

March 1984
00 1 01 1204
---
...........
(c)
Whisper speech - fairly frequent.
(d)
Whistle speech - most frequent of all, this type
requires skilled listening, as the modulation of the
whistle (both in terms of frequency and amplitude) is
fractional compared with the normal deviations.
What is significant is that apart from the relatively
Infrequent type (a) all the above formats are independent of
the possession of vocal cords. The afficianados of the
radio play" and "CB breakthrough" schools might care to
explain why almost everybody coming through on our equipment
seems to have lost their vocal cords!
4. Changes in Acoustics
The following effects depend upon one factor: an echo;
Sechoiness, reverberations, acoustic liveliness, choral
Peffects. The factor in question is time delay.
4.1. A significant proportion of the utterances
Lexhibit one or more of the above effects, indeed there are
examples of changing acoustics - e.g. reverberation - during
an utterance - a rather improbable event. Furthermore,
normal broadcasting studio, radio room, or vehicular
broadcasts are made from an acoustically "dead" environment.
5. Evidence of Intelligence
Because of the low quality of the voices, and
principally because of the quite exaggerated dependence of
earlier researchers on the significance of what was said, I
have decided to concentrate, rather, on measurable factors.
D My main effort, therefore, in looking for evidence of
intelligence, has been to look for evidence of communica-
tions theory being applied. In particular, communications
theory (after Shannon et al.) states that the more
predictable a message becomes, the greater the probability
of its being correctly received. There are many ways in
which this can be applied.
5.1. By sticking to a regular schedule, same place,
same time each day, the U/M is found to progressively
increase.
5.2. Evidence of "redundancy" being applied.
(a) Undue prolongation of vowel sounds is sometimes found
(the sort of thing one does when hailing against the
---
wind, or across a valley, which gives the listener's
pattern recognition means a better chance),
(b)
Repeated messages. The best example was 10 successive
repetitions of one simple phrase.
5.3. The rather too regular periodicity of some
synthetic voicing, suggesting - but not necessarily implying
- the use of technology.
5.4. The time correlation between the initiation of a
stimulus and the reception of a response.
5.5. The fact that certain questions produce not just
no response, but such a diminution of background noise, to
almost complete silence, that the change in minus dbs is
itself a signal of noticeable magnitude! Such questions as
"How can I improve this equipment?" produce negative
"signals" of this type, implying an appreciation of the
question, and thus, intelligence.
6. Evidence of Integrity
Integrity is here used in the sense of comprising a
whole, as in structural integrity. The fact that different
experimenters can pick up identical phrases; and that on
occasion, the same names for the correspondents have been
used, does tend to indicate a degree of integrity. Against
that, the probability ot being able to access the same
correspondent repeatedly is low, but this may be due to-
system defects or our own ignorance of the rules involved.
7. Evidence of Purpose
7.1. This, at one level, may be thought of as the
desire to communicate, and this is supported by sucho
evidence as there is of communications theory being applied, U
deliberately or intuitively.
7.2. Generally, at least 90% of all utterances are
comprised of messages, warnings, instructions and pronounce
ments - that is of directed statements uttered, apparently
with some purpose. It should be mentioned, however, that
there seems to be a considerable conflict, and no single
mutual purpose is as yet evident.
41
40
Approved For Release 2000/08/15 : CIA-RDP96-00792R000701010001-8
Approved For Release 2000/08/15 : CIA-RDP96-00792R000701010001-8

March 1984
PSI RESEARCH
PSI RESEARCH
8. Other Rules
&1. The system itself consists of a radio frequency
spectrum, the positioning of the lines being related to a
Omeasure of the dielectric constant in the vicinity of the
Tobserver, compared with line positions of a dense spectrum
Dextending from the audio to radio frequency range.
8.2. Evidence was gradually accumulated indicating
that rules of as yet unknown communications theory may
Dexist. This would seem to be a more advanced, broadly based
Stheory than our late 20th Century communications theory.
The communications theory of Shannon et al., may, by
"comparison, be rather like, "A handyman's short guide to
practical communications," admirable in its attention to
details and instructions for using tools (mathematical), but
hardly a full treatment. The rules of Shannon and others
owould be exemplifications of those more general laws, yet
"undiscovered. But one of those laws is just coming into
focus, and it would seem that Sheldrake's Morphogenetic Laws
would actually be communications laws, although he has
[tended to concentrate in just one area of implementation.
9. Implications
9.1. That EVP, as a psi phenomenon, is available for
frequent and predictable observation, using the right
equipment.
9.2. That "there is something there," as yet un-
recognized by science.
9.3. That communication rules exist which, when known,
? may appreciably alter our world-view.
Additional Comments
After receiving Alexander MacRae's article we
wrote him, asking for a circuit diagram and a
detailed description of his equipment which would
enable interested readers to attempt to replicate
the findings. We received the diagram (Fig. 1)
and an explanation of operating principles of the
device (called by MacRae MK1 Alpha) which we
present below. - Ed.
42
March 1984
There are two main inputs. One is a monitor of
Galvanic Skin Response or skin resistance.
What happens
here is that current is applied to the skin of the hand and
the potential difference between the electrodes is applied
to the non inverting input of an operational amplifier which
has 100% DC feedback so that it is always conductive within
any reasonable limits. But what we are really interested in
is not so much the potential difference across the hand but
changes in potential difference.
That operational amplifier also has an AC gain which is
variable according to the setting of another control. By AC
I am talking about low frequency perturbations or broad
blips. That output, which consists of a standing level on
which from time to time are superimposed long term blips, is
taken to the voltage control oscillator (VCO) input of a
phase-locked loop (PLL) and it is then capacitively coupled
to the VCO input of another PLL so that the output frequency
of the first PLL is proportional to the standing level and
the output of the second PLL is proportional to the change
in standing level. The first PLL is set at about 100 Hz
which is about the Glottal frequency and the output of the
second PLL is about 5 KHz range.
Those two outputs from the PLL's are then taken to two
of the three inputs of a 1 of 8 Data Selector Chips. The
third selector input is taken high. So what happens is,
depending on the combination you select, that determines
which output pin of the Data Selector Chip is connected to
the input. Both output pins on that chip are taken high and
the input is taken to ground so every time the input gets
connected to, for instance, output Ill - then that output
goes low. The pullup resistor goes directly to VCC. Now on
the other output, let's call it 110, the pullup resistor
goes instead to the full wave rectified version of the long
term blips on the output of the operational amplifier.
Getting back to the output of the operational amplifier
- that was capacitively coupled to one PLL. That capaci-
tively coupled output is taken also to two comparators.
Both comparators are biased such that their outputs are
setting at ground.
The output from the operational
amplifier is capacitively coupled to the non-inverting input
of one and the inverting input of the other. This means
43
Approved For Release 2000/08/15 : CIA-RDP96-00792R000701010001-8
Approved For Release 2000/08/15 : CIA-RDP96-00792R000701010001-8

March 1984
PSI RESEARCH
Approved For Release 2000/08/15 : CIA-RDP96-00792R000701010001-8
PSI RESEARCH
that when the signal is a positive going blip one comparator
output of the other comparator will go high. Those two
outputs then feed into the bases of two emitter followers
which join together to drive a tri-color LED.
Now, the full wave rectified version of this is picked
off via two diodes which are commoned together and that
through a pullup resistor goes back to the output of the l
of 8 Data Selector. So what is happening is that the two
relevant outputs of the Data Selector go to emitter
followers and each emitter follower goes to a socket output.
Now you can plug a coil into either of those outputs and
take that coil up to a radio, or simply place the unit
beside a radio. Tune to, for example, 260 KHz.
If you plug the coil into the "D" output what you will
get is a continuous tone indication of what is happening to
the two PLL's. It is continuous because the pullup resistor
at that output goes to VCC. If you plug it into the other
one then you won't hear anything until a blip comes along
which gets full-wave rectified and you get a positive
voltage appearing on that other pullup resistor and so for a
moment you will get the tones coming through. So in one
case you get continuous tones - or, if you don't want those,
you can get momentary tones indicating that something has
changed.
In the Mark One version the blip was further fed to a
555 to turn it on momentarily and if the blip was positive
then you ran the 555 at one rate and if it was negative you
ran the 555 at another rate. This was done through
selection diodes. So what you've got for a positive blip
was, shall we say, three low frequency pulses which sounded
like three raps and in the other case you got one rap. That
earlier system is now dispensed with. It's really just
duplicating what shows up in the LED. None of that is
really the interesting part.
Remember I mentioned that the PLL's were running at
different frequencies. The frequencies of the PLL are set
not just by the resistors, but by the capacitor on each one.
Now, you unplug from the R input and you are still
monitoring Galvanic Skin Response or body capacitance,
whichever you prefer, and you plug it in now and this C
input is actually in series with the timing capacitor on the
44
March 1984
first PLL. So what you now get is a frequency in the 100
KHz range which is proportional to body capacitance, if you
like, and this swoops all over the place and varies a lot.
The other frequency from the other PLL is fixed and because
its down in the Kilo-Hertz range, it has a set of fixed
harmonics at, let's say, 5 KHz Spacing all the way up intoc
the 3 MHz range. So that's what's happening in the Alpha.
As to the exact means, whereby "EVP" enters the system,
research must continue. My original premise was that a
variation in either permeability (M) or permittivity (E )
by altering the "electrical space" - could accomplish phase
or frequency modulation. Interestingly, the article in thec
December 1983 issue of Psi Research, "The Physical Fields of
Biological Systems," mentions remote monitoring
permittivity on the bases of UHF radiation. Indeed, the
first phenomenon noted on the Alpha was not the EVP but9
pulse-rate. The USSR researchers have remotely monitored
heart rate by monitoring change in skin surface charge, but
have to use a Faraday cage - equipment not needed with the
Alpha (as using narrow-band radio pick-up the noise power is
considerably reduced).
Where Do We Go from Here?
(Concluding Editorial Remarks)
We asked Alexander MacRae to send us a sample EVP tapes
We received it but were not greatly impressed: much noise,
whistle and something barely audible which (with a certain@
degree of imagination) can be interpreted as words
However, we still believe that the field is worthy of@
inquiry. If, through continuing research, we do not obtain
a method for recording "voices from beyond," but rather a
reliable method for monitoring force fields of biological
systems without using a Faraday cage, or even evidence that
the process of concentration on "white noise" enhances ESP,
the endeavor seems to be worthwhile.
SKYETECH
7-2 Industrial Estate
Portree, Skye, Scotland IV51 9HI
45
Approved For Release 2000/08/15 : CIA-RDP96-00792R800701010001-8