A Testable Hypothesis Explaining EVP Formation in Digital Recorders

By Tom Butler. Based on a Hypothesis Proposed by James Jones
Previously Published in the Spring 2004 AA-EVP NewsJournal

IC recorders, such as the Sony ICD B7, have become well-known for their effectiveness as a tool for collecting EVP. This disparity between the effectiveness of digital and analog tape recorders offers an important opportunity for analysis; specifically, what is the active mechanism in IC recorders that is not present in analog recorders?

AA-EVP Member, James Jones, has recently offered a possible answer that is both technically insightful and very testable. If his hypothesis is shown to be correct, it provides a substantial proof that EVP are formed of noise and that they are formed within the electronic components of the recording device.

In IC recorders, the analog input signal is converted to digital information to enable processing and storage. This analog to digital (a/d) conversion is based on a sample rate in which the amplitude and frequency of the signal is converted into a series of digital words that are suitable for processing in the digital circuit. In a hypothetical digital word of “01101,” the last “1” is the Least Significant Bit (LSB). The LSB represents the lowest level of sound that can be accurately represented by the a/d converter.

James Jones contends that the error in accuracy of the a/d conversion between the analog input and the digital output introduces a chaotic electrical environment (noise) within the recorder as the a/d attempts to find a best fit for the incoming signal. The fewer the bits in the digital word (probably smallest in inexpensive recorders) the greater the noise associated with a/d conversion. This a/d induced chaos is conducive to EVP formation.

As James has noted, “Noise plays an important part in the encoding process. By strategically introducing noise, it is possibly to encode signals that are smaller than the LSB or less than one step of quantization.” Externally supplied low volume noise will allow the a/d to process signals that would normally fall below the LSB, thus reducing internal noise. Supplying extra noise is referred to as “Dither.” See http://www.earlevel.com/Digital%20Audio/Dither.html for a tutorial on the subject.

“Mike Sullivan tested this theory by tracking the number of EVP collected with and without the addition of white-noise. According to the theory, the amount of error and distortion allowed into the output of the recorder would be decreased with the addition of white noise, and therefore, the chaotic conditions necessary for the creation of EVP would be diminished. In fact, he recorded fewer to no EVP while supplying external noise when using an IC recorder.”

This theory proposed by James Jones possibly represents one of the first “smoking gun” analyses of where and how EVP is formed within a digital recorder that answers the current evidence. We would like other researchers reading this NewsJournal to consider similar experiments, and as always, we look forward to reports that we can include in the NewsJournal.