Close this Window to Return to the Paranormal Page
With digital recorders you have a couple additional concerns that impact the quality of the recording. It is beyond the scope of this report to give all the technical reasons why the following conditions apply. The mathematics involved in the Nyquist Point, DSP, FFT, aliasing, and intermodulation would put most readers to sleep, so I won't include all that here. Those interested can do a search on these terms. I will only cover the results of tests I conducted on a high quality digital recorder as they apply to EVP work.
The biggest limitation is the sample rate. I tested various sample rates and found that the minimum rate for EVP use is 96 KBPS. At 48 KBPS there was some degradation in high frequency response. Below that the audio was unreliable for analysis. It should be noted that most digital voice recorders use anywhere from 8 KBPS up to around 20 KBPS. From this it is seen that for reliable EVP work the use of compact digital voice recorders is not recommended.
Another factor is the Analog to Digital conversion system. A to D conversion is based on working with a digital value which corresponds to the level of the analog signal at each sample point. The more bits utilized, the more accurate this conversion becomes. Each bit you add doubles the number of digital values that can be used to represent the analog signal. Digital recorders use anywhere between 16 bits and 27 bits. The 16 bit systems generally apply to the small digital voice recorders, most good digital recorders use 24 bit systems. A few really expensive units I have seen go to 27, it may be that some go even higher. In any event, satisfactory operation can be obtained using 24 bit systems. If yours goes above that there is no problem, your resolution is well above the minimum specification. Any recorder that uses 24 or more bits of resolution is satisfactory for EVP work. The small voice recorders using fewer than 24 bits are not suitable for EVP recording.
A final concern with digital recording is the format. Recorders should not alter the original signal while making the recording. Most small recorders, in order to save memory, use various compression techniques. This is unsatisfactory. Any digital method that compresses the data is considered "lossy", meaning that certain elements of the original are lost. MP3 is a lossy compression method. WAV and BWF is not considered lossy, therefore your digital recorder should use one of these two formats when creating the files. Also when you retrieve the files you should save them in one of these two formats if you plan to use them for analysis. Keep in mind as you save them the same rules apply to your computer as to the digital recorder. That is your A to D system should be 24 bits or greater, and the sample rate you select should exceed 96 KBPS. That is determined by both your settings and your sound card in your computer.
If your intent is to simply make the audio available for others to listen to it is acceptable to convert them to MP3. Just be aware that as MP3s they are not suitable for analysis, and should someone wish to do in-depth study they should obtain the original WAV files.