EVP RECORDING  

How do you know when you have “over processed” a sound file?

Answer: The short answer is that you may not be able to know. It is best to follow a few rules so that you might avoid the problem in the first place.

Over the years, more and more sophisticated programs for managing audio files have become available. Audacity, for instance, is an excellent, free program that has a growing number of important features helpful for analyzing a sound file for EVP. The only problem is that it has become too tempting to use as many of these new editing tools as possible to try and make an indistinguishable bit of audio into a Class A EVP.

As an engineer, I have tried my share of tricks to improve the quality of a sound file, and I am guilty of the occasional over processing. It is just too tempting. However, our policy is to do as little as possible to a sound file. We reason that the communicating entities can do better, so it is part of our experimental procedure to ask them to repeat what we think might be important but unintelligible statements. They do not always comply, but it seems reasonable to put the burden for communication on both ends of the circuit.

In voice formed by our physical body, the fundamental frequency (F0) produces harmonic frequencies (formants known as F1, F2 and so on) on the way out of the mouth. It is the intensity of these frequency groups, and how our mouth modulates them that form our voice. The formants are in predictable groups, and a voice print will show this relative order as clusters of frequency that have more intensity.

Research has shown that the voices in EVP are formed out of available audio frequency energy. See http://aaevp.com/articles/articles_about_evp10.htm. This means that, in some cases, the voice will be formed of odd groupings of sound, rather than the orderly array of biologically initiated formants. When a physically formed voice is filtered, we can expect it to respond in predictable ways, but when an EVP is filtered, it is possible to unknowingly delete frequencies vital for the meaning of the word. As an example, the word “love” has a strong beginning sound that, in an EVP, might be formed by a noise burst initiated in the amplifier by the entity, rather than a biologically formed “L.” If the EVP were passed through a noise filter that reduced the intensity of that noise burst, the word could be changed to sound like “of.” A meaningful but poorly heard EVP would be changed to well-heard nonsense.

A second hazard of over processing sound files is that it is possible to amplify a sound file many, many times with a press of the mouse button. Hand-held audio recorders have a “floating ground” which does not very effectively exclude external electrical signals such as a nearby radio station. When they are being handled, the experimenter can function as an antenna to further exasperate the problem. Finally, the connection between a computer and the recorder makes the recorder even more of an antenna. The zero-signal or quiet state of the recorder while in the record mode will, in normal circumstances show as no sound being recorder, or at worst, a small background sound. But if this zero level signal is amplified a great deal, all sorts of modulation can be discovered, and it often sounds a little like EVP. In fact, it is probably a composite of all of the local radio broadcasts in town.

There are other hazards, most too exotic to be concerned with, but the bottom line is that it is possible to change the meaning of a phenomenal utterance by using too much filtering, sound reduction and/or amplification. This is especially evident when many tools are used, such as noise reduction, filter, normalize and amplify. Even changing the sequence of how these tools are applied can change the outcome.

As policy in the AA-EVP, it is recommended that no more than ten or twenty decibels of amplification and only passive filtering of frequencies above about 3,000 Hz be used if the EVP is to be shared with others or used in research. An alternative policy is to provide the original EVP as a raw file, followed by about a half-second of silence, and then the same example as it has been modified. In this case, there is probably no reason to limit the processing, except it is a consideration to tell the listener what processing has been applied.

If the recording is only for personal use, then of course there are no limits, but be sure to understand the hazards of self-deception by making the utterance something it may not be.

One point to consider here is that we recommend converting the example to mp3 for posting on the Internet or emailing, but we do not recommend processing an mp3 sound file in any way. Well, perhaps a little amplification. So, save, process and store your files as Windows PCM (*.wav) files or whatever is equivalent for your operating system, and convert a copy to mp3 for distribution. Also, we have had some difficulty with *.wav to mp3 conversion at very low sample rates, say 8,000, so listen to the resulting file before distribution.

As a final note, other than as a demonstration that it is possible, we feel that EVP found on the reverse of a sound file may be more problematic than useful. There are too many sound combinations in normal speech that, when played backswords, make other normal words. As a rule of thumb, we feel that unless the utterance is a Class B+ or Class A that is clearly an appropriate response to a question or comment about the circumstance, it is best to disregard reversed examples. As always in trans-etheric phenomena, the content of the message should be a guide as to whether the message is ... well, is trans-etheric.

I was wondering how the ITC/EVP arena views apophenia or pareildolia as a science or if in ITC research allowances are made for these psychological observations. Can ITC research and apophenia and its subsets coexist? Do your researchers take pareildoia into consideration when determining an ITC event? If not, why? If yes, how do you separate the two by definition?

Answer: There can be no doubt that some reported EVP are mundane sounds mistaken as a paranormal voice. Pareidolia is a real enough phenomena that we naturally must deal with at all times, just as we deal with apophenia which is the visual form. Part of the dogma of those terms is that human brains are hardwired to recognize voices or faces in noise; so much so that we find these things even when they are not there. In deer hunting, they call that "buck fever." However, later research shows that this assumption has been greatly overstated. As it turns out, the mind has a pattern recognition center that is trained to recognize faces from cultural upbringing, just as we are trained to find understandable words in sounds that would be gibberish for someone from a different culture. Those same centers can be trained to recognize any pattern, showing us that we should as easily find a favorite song in random audio noise or maybe a favorite movie star in visual noise.  

In fact, we do find this in situations where people are relaxed, and there is a relatively steady-state noise in the environment, such as a refrigerator fan or heater fan. Many people report "almost" hearing a distant conversation or radio station. They even report almost hearing music. We think that EVP are formed by a physical process known as stochastic resonance. In that process, a weak signal will act as a seed signal to be amplified in a field of broad-spectrum noise. The weak signal emerges as a pattern energetically stronger than the field of energy and recognizably the same signal. Thus, a weak telekinetic signal might be amplified by the noise we use in audio phenomena and visual phenomena. Our thought is that the neuro-electric field in the brain might act on the broad-spectrum energy from the fan to allow familiar or expected hearing or seeing. We do not consider such faces in clouds or voices in fans to be ITC and such reports fall way outside of the usual characteristics of EVP or visual ITC. (I have attached a document that has a list of characteristics.)

There is also a very real problem with stray radio signals producing "unexpected voices in recording media." Since we use mostly digital recorders, and the majority of RF these days is digital, RF contamination is less of a problem than it was a few years back. On the other hand, hand-held digital voice recorders are excellent antennas and there is not much of an effort to eliminate the antenna effect in their design. EVP has been successfully recorded in mil-spec screen rooms. Armatures can do the same experiment by nesting metal containers with their digital recorder isolated in the center.

See http://aaevp.com/articles/articles_bill_weisensale_experiment.htm

A common problem is the recording of unnoticed people talking in the area. This is especially a problem in field recording when a ghost hunting crew is at work. The very first Best Practice we proposed was the use of a control recorder of higher quality, such as a video camera. EVP seldom forms in higher quality recording circuits. Also, we have no record of the identical EVP showing up at the exact same time in more than one recording process. These two characteristics enable us to say that it is reasonable to discard an utterance that is found in more than one recording unless there are extenuating circumstances.

See http://etheric-studies.aaevp.com/best_practices_development/index.php?title=Using_a_second_audio_recorder_as_a_control

So yes, we are aware of these things. They are real and techniques must be designed to account for them. A problem though is that anyone can attempt to collect examples of these phenomena, yet few are well-trained in the subject. For instance, orbs found in photographs are almost always particulates illuminated by the camera flash or light reflecting from a bright surface. Yet, people persist in thinking that orbs are etheric entities, such as the local ghost. The difficulty is that there is some evidence indicating that some orbs are phenomenal, and unless the orb is clearly in motion, there is no current technique for distinguishing reflected light from a self-luminous orb. One of the reasons for the Best Practices part of the Etheric Studies initiative is to enlist experienced researchers to establish techniques that will guide people with lessons learned.

See http://etheric-studies.aaevp.com/

Research can be designed to expand our understanding. For instance, we have conducted double-blind experiments to determine whether or not a person can use EVP to collect information that is unknown to the experimenter.

See http://aaevp.com/articles/articles_about_4cell_experiment.htm. The 4Cell results were in the range of 60% correct responses.

We have conducted two online EVP listening trials that showed "average" people can correctly understand about 34% of the words in EVP. A very experienced person would have heard more like 100%, and the difference between the website visitors and an AA-EVP member is about 40% word recognition. (there is an apples and oranges problem with the way this figure was arrived at, but it agrees with observations.)

We are still seeking "peer input', but the report at http://etheric-studies.aaevp.com/best_practices_development/index.php?title=EVP_online_listening_trials shows that EVP contain real words but that the words are formed in a novel way that requires training to be understood--much like learning a new language.

So, I would say that we are developing a good foundation of research that will enable us to properly study these phenomena.

When speaking to a participant in the Big Circle (see the video by clicking on the image below), we do not ask for scientific proof, but when we are talking to just about everyone else, we require good reason to accept the example of phenomena. You will see that there are a lot of people who make claims that cannot be supported with good research, but I must say that it is the responsibility of the listener to recognize the difference between a person trained in the field and an amateur before concluding that all of these phenomena are imaginary. This is also true of what scientists say about EVP. Are they qualified?