Hypothesis: EVP and room reverberation

Originally posted on Sunday, February 08, 2009



Hypothesis: EVP and room reverberation


Here is the basic hypothesis for the research project I'm starting on. I'm posting this in an effort to not only share with others what I am working on, but I hope that if there are any inconsistencies in my method that someone will point them out. Also if anyone knows any audio experts that may be willing to take a look, please point me to them. Some basic protocol has been left out because it really doesn't need to be stated over and over again to those in the know.


Hypothesis: The basis for my research project suggests that legitimate Electronic Voice Phenomena(EVP) can not produce room reverberation like the voice of a living person in the same room. Since the idea of EVP is that the anomaly is not heard at the time of the recording, I assume that no air is ran across vocal cords to produce sound waves. Basic physics dictate that it would be impossible for these anomalies to produce any natural room reverberation without moving air to produce the sound waves. The anomalies should at the very least not possess the same reflection time as the control. Overall reflection time is expected remain static regardless of the dB range.


Testing: Controls will be recorded at the beginning of each session for each room where data is recorded. Non-compressing studio quality recording equipment will be used as opposed to low fidelity dictation devices or tape machines(which produce "hiss"). Measurements of the room will also be taken for use with the Sabine Formula. The sessions will take place in a controlled environment as to avoid any false positives. All data will also be collected only at indoor locations where rooms can be measured.


Analysis: Only "Class A" data will be analyzed. "Class A" indicates a clear, intelligible, and audible anomaly with the volume of a normal speaking voice or above (e.g. No whispers or faint
anomalies will be analyzed or reviewed). The Sabine Formula will be used to gather expected reverberation time to compare with the control data. Any anomalies will be reviewed as well and both will be used to look for any inconsistencies. I do however see some problems with the Sabine Formula because it does not take in any consideration for sound absorbing materials such as ceiling tile, carpet or large windows and doors which may allow sound waves to escape rather than to reverberate. Taking control recordings should help to adjust this. In some cases, placement of the living speaker or control may help to avoid any problems with this. Since the anomalies are obviously uncontrolled, placement won't help with them.

Added: In some cases, random anomalous noises(non-vocal) may also be analyzed, since sound waves work the same whether they are vocalized or not. However, it may be wise to use video to back up whether the noises are anomalous or not. It would also be best to use distinct sounds which are identifiable as something not present when data was recorded. For instance, a vacuum cleaner would be analyzed rather than random knocks and the like. It may be wise to remember that spontaneous noises do not indicate the existence of life after death.

I would also like to point out that this theory is pointed towards "intelligent" anomalies. Anomalies of a "residual" nature may very well contain a room's natural reverberation. Since we are unsure how residual anomalies are recorded, it is impossible to assume either way. This will be another side of this research.

Sabine Formula: I have tested the Sabine formula and found that it predicts overall reflection time with a good amount of accuracy. Since it measures overall reflection time, early reflections do not need to be reviewed. In addition, since early reflection times normally vary and are practically impossible to measure, I do not find them necessary. Overall reflection time in an average sized room is difficult enough to measure. For the most part, I find listening with the ears and brain to be the best option. The Sabine formula may still be found to be relevant at a later date and will always be there to back up the research along with the controls.


This project will be long term and I don't expect to finish next month or even next year. The main reason, though, is the tiny amount of "Class A" data I've received in the past. These things don't grow on tree's as some would believe. Please comment, ask any questions you may have or point out any flaws in my logic.


Update (12-7-09): I have managed to record one EVP matching the required criteria. Unfortunately, it was unintentionally 'stepped on' by an investigator so I could not fully analyze the reverberation. Early reflections, however, show no reverberation present.