Historically, Electronic Voice Phenomena (EVP) have been thought of as unexpected voices found in recording media. The primary technique for recording EVP has been with an audio recorder and the voice is thought to be formed of available audio energy (background noise).

Recent developments have made it practical to work with real-time, two-way EVP. With this development, an expanded definition is required. As one has not gained wide acceptance, the following should be considered a draft definition:

EVP Formation

EVP can be collected under circumstances which preclude the possibility of such mundane sources as radio frequency contamination, unnoticed background conversations, contamination from previously recorded speech and mundane sounds mistaken as voice. People around the world have demonstrated this fact since before Friedrich Jürgenson made his famous EVP recordings fifty years ago June 12, 1959. The important point is that much is known about ordinary EVP formation which allows practitioners to recognize what is and is not phenomenal.

As shown in Figure 1, the phenomenal utterance is formed in the recording device. Current theory is that noise is used for voice formation and the dominant physical process is thought to be the amplification of a weak telekinetic signal by the action of stochastic resonance on the broad-spectrum audio-frequency noise. The "frequencies" available in the sound stream at any moment will determine the richness of the resulting voice and no new audio energy is produced.

There is no evidence that the psi processes involved in the trans-etheric communication produce additional physical processes. Once the information is in the physical, it is completely subject to physical principles. Keep in mind that stochastic resonance was not documented as a process prior to 1981. The fact that a physical process needed to explain some aspect of EVP has not yet been defined does not mean that a paranormal solution must be used. It is important to this essay to note that the preponderance of evidence indicates the requirement of physical processes. There is no need to say that a phenomenal event "somehow" occurs or is caused by by divine intervention.

The characteristics associated with traditionally formed (transform) EVP that are part of the theoretical model used by the AA-EVP include:

• EVP are not acoustical phenomena, meaning that they are formed in the recording mechanism, and can be collected with an acoustically isolated recorder.

• The voice is formed by transforming audio-frequency energy available in an electronic circuit into an approximation of human speech.

• EVP are energy limited, meaning that their formation is in short packets of a few words so that shorter packets tend to have greater volume, longer packets tend to have less volume and very long utterances are seen as several concatenated packets with brief pauses between each one.

• Any audio-frequency energy is apt to be transformed into voice as EVP.

• EVP formation occurs in analog circuits, and once digitized, tends to be stable.

• The utterances are formed by selecting available audio-frequency energy. Audio-frequency energy is not created for this purpose by the etheric communicator and may be in the form of relatively random noise, fragments of voice or synthesized voice.

Because of the many years of study by practitioners around the world, much is known about EVP recorded by traditional means. Small details of the characteristics listed here may be arguable but there is little doubt that this form of EVP occurs in audio recorders in a predictable way.

Stefan Bion has developed a computer program named EVPmaker which has been shown to produce EVP based on random selection of segments of a sound file. The input audio file may be recorded "live voice" but the program is able to use synthesized human voice fragments known as allophones and/or the smaller phonemes.

Precedence for psi influence of random processes has been established by the parapsychological community. For instance, the Princeton Engineering Anomalies Research (PEAR) Lab reported a small but significant effect in psi studies. Depending on presets made by the practitioner, the randomly selected segments of input file are typically too short to contain recognizable speech, especially when using synthesized speech. Once selected by the random process, the segments of sound are combined with others to produce the desired words. As such, EVP formation is thought to be controlled by control of the random process.

Audio output for EVPmaker is a staccato sound as if the system has developed a severe stutter. Recorded speech is considered "live voice" and is inherently suspect as a sound source for EVP formation; however, when using synthesized voice fragments, the input file can be controlled to minimize pre-existing recognizable words. As such, words in the output should be clearly random or they are phenomenal.

A recent development in the study of EVP is the introduction of speech synthesis technology. Bill Chappell has shown that it may be possible to produce meaningful utterances by detecting changes in the environment to control a speech synthesis process. The working assumption is that the communicating entity will change the environment to cause synthesis of the intended words. More research is needed, but the implication is that detection of presumably easily influenced chaotic processes such as room temperature and electrical charge can be used to initiate EVP formation by controlling a speech synthesis process. A second possibility is that the communicating entity is able to directly influence the speech synthesis circuit; however, more research must be conducted to determine this.

Chappell’s Paranormal Puck is programmed with words (speech mode), and a library of English-language phonemes (phoneme mode). Phonemes are the smallest segment of human speech, and when appropriately combined in an audio file, will produce meaningful speech.

An important point to consider in the analysis of environmentally controlled speech synthesis for EVP formation is that the only physical process necessary to explain observed results is a change in ambient energy detected by the device. The library of words or speech fragments is present in the designed capabilities of the technology.

Substantial research and field observation has shown that the processes involved in trans-etheric phenomena are probably knowable. Physical processes such as propagation of sound or light are measurable and their understanding can be applied to devices and other processes. In the same way, the processes governing subtle energy and the behavior of personalities after transition are knowable and able to be applied in technological solutions. A few of the assumptions used in this study of how EVP are formed are:

• EVP are formed via the influence of subtle energy on physical processes according to knowable principles. With this assumption, the explanation for a phenomenal event that it "somehow" occurred cannot be used for research. The "somehow" should be quickly replaced by an established or modified theory.

• There is consistency in how subtle energy influences are able to affect physical processes. For instance, both visual forms such as images found in video feedback noise and EVP formed from background noise appear to depend on something like stochastic resonance.

• The influence of physical processes by subtle energy is energy limited and the most energy-efficient processes should be considered first. This also means the physical energy required for phenomena can be accounted for.

• Available physical processes can be initiated by an etheric influence to cause observable phenomena, but matter and energy are not seen to be created. A physical person might be convinced to do something in support of phenomena. An important implication of this assumption is that physical energy and objects may be transformed but they are not created to form phenomena. For instance, if sound is not present, it cannot be transformed into the words of EVP. The required sound will not be created but a physical process may be initiated which in turn causes sound for voice formation.

The observed characteristics of EVP suggest how they are (typically) formed, and by extension, how they (usually) are not formed. The average person can use this information to evaluate examples to help determine if they are mundane or paranormal. It also means that a deviation to the "rule of thumb" should be considered rare and the example should be held in reserve by practitioner and researchers alike until more data is available.

EVP practitioners listen very closely to sound files in an attempt to find meaningful sounds. This is because the audio signal normally thought of as just background noise is considered a likely place to find a paranormal message. However, audio files thought to be needed for voice formation are usually very noisy and meaning is potentially everywhere. It is common for a distant but ordinary conversation to be detected by a recorder and then be reported as phenomenal. When a practitioner is expecting a meaningful sound, the next sound is sometimes given that meaning no matter what intelligence the sound actually carries.

Typical sources of sound mistaken as EVP include:

• Ordinary variations in background sound mistaken as voice: For instance a burst of static is sometimes heard as an angry voice.

• Stray radio signals mistaken as EVP: Hand-held, battery-powered recorders are excellent antenna for radio reception and the electronics are capable of making stray AM signals understandable.

• "Morphed" foreign-language sound source: Naturally occurring foreign-language words are sometimes reported as having been "morphed" into EVP but then are recognized by a person who speaks the foreign language as not being EVP at all.

• Reversing a sound track: Some words sound like other words when heard in the reverse.

• Ordinary sound changed with an audio editor to simulate EVP: Changing the tempo of a sound file can change its intended meaning. Since voices of EVP are usually formed out of the background noise, filtering the noise can sometimes change how the words are heard. Excessive amplification can make detected radio audible.

There is some evidence that a sound file might be altered while it is in an analog form, but sound files are thought to be stable once they have been digitized. As such, many people can hear the same digitized sound file and should hear the same sounds. If there is an anomalous utterance in the sound file, then it will be the same each time. This is an important characteristic of EVP if online listening tests are to be trusted to produce meaningful data.

With that said, it is common to see a doubtful EVP example shared on the Internet with a surprising number of people saying they hear it as the practitioner reported. In subsequent tests, it is possible that even the person who reported the example will not be able to understand the utterance. Assuming the practitioner and listeners did hear the example as reported even though the words were not actually present, what would explain the confusion?

The most common problems include:

• Suggesting what should be heard: Especially when the example is of marginal quality, the listener is easily guided to hear what is expected by how the example is reported even though the words may not be present.

• Fabricating a "likely story": The practitioner has asked a question and expects the next sound to contain the answer. Whatever the next sound "sounds like" is taken as part of the answer even though it may not make logical sense. To make it sound correct, the practitioner applies meaning to subsequent sounds to tell a story that provides a plausible answer to the question.

• Setting listener expectation in "real-time" dialogue: Two-way conversations via EVP are very difficult, and while shown to be possible, the process can also be misleading. It is reasonable for the practitioner to ask the entity to repeat a word by saying something like, "Did I hear you say ‘Tom?’" However, in doing so, the listener is preconditioned to hear "Tom," even though the word may not be present.

• Selectively relating utterances to questions: The usual recommendation is that the words of an EVP just before or just after the practitioner’s question can appropriately be applied to the question. As such, asking a question and selecting a word out of a string of possible utterances would be considered selective reporting.

This is a rapidly evolving field of study. People once assumed that we will never know how spiritual things work, but today, we have come to realize that we can know how trans-etheric phenomena manifest; we can quantify them and we can build devices based on our understanding to apply them in our lives.

Subtle-energy phenomena such as EVP are subject to the very conceptual influence of the etheric aspect of reality, but at the same time, they are subject to the constraints of physical processes. Their study requires close attention to how people process information. It is not safe to say that an example is EVP unless the human factor is considered alongside the physical processes. At the same time, this study requires an understanding of the physical processes that might be involved.

This is a complex subject and the average practitioner cannot be expected to be an "expert" in this field without considerable study. Recording EVP is probably a craft, and any good craftsperson knows to depend on the foundation of knowledge made available from previous experience and the science that has come before. Our field has little science, but quite a lot of experience. It is for practitioners to take it on themselves to become familiar with the best practices.