Electronic voice phenomena (EVP) are electronically generated noises that resemble speech, but are not the result of intentional voice recordings or renderings. Common sources of EVP include static, stray radio transmissions, and background noise. Some have claimed these sounds are of paranormal origin, while there are natural explanations such as apophenia (finding significance in insignificant phenomena), auditory pareidolia (interpreting random sounds as voices in their own language), equipment artefacts, or simple hoaxes. Recordings of EVP are often created from background sound by increasing the gain (i.e. sensitivity) of the recording equipment.
Classes of EVP......
Voices are very clear and easily understandable by many.
Voices are fairly loud and clear and are sometimes audible without headphones.
Voices are very soft and often indecipherable. These tend to be researcher dependent.
Parapsychologist Konstantin Raudive, who popularized the idea, described EVP as typically brief, usually the length of a word or short phrase.
As the Spiritualism religious movement became prominent in the 1840s–1920s with a distinguishing belief that the spirits of the dead can be contacted by mediums, new technologies of the era including photography were employed by spiritualists in an effort to demonstrate contact with a spirit world. So popular were such ideas that Thomas Edison was asked in an interview with Scientific American to comment on the possibility of using his inventions to communicate with spirits. He replied that if the spirits were only capable of subtle influences, a sensitive recording device would provide a better chance of spirit communication than the table tipping and ouija boards mediums employed at the time. However, there is no indication that Edison ever designed or constructed a device for such a purpose. As sound recording became widespread, mediums explored using this technology to demonstrate communication with the dead as well. Spiritualism declined in the latter part of the 20th century, but attempts to use portable recording devices and modern digital technologies to communicate with spirits continued.
American photographer Attila von Szalay was among the first to try recording what he believed to be voices of the dead as a way to augment his investigations in photographing ghosts. He began his attempts in 1941 using a 78 rpm record, but it wasn't until 1956, after switching to a reel-to-reel tape recorder, that he believed he was successful. Working with Raymond Bayless, von Szalay conducted a number of recording sessions with a custom-made apparatus, consisting of a microphone in an insulated cabinet connected to an external recording device and speaker. Szalay reported finding many sounds on the tape that could not be heard on the speaker at the time of recording, some of which were recorded when there was no one in the cabinet. He believed these sounds to be the voices of discarnate spirits. Among the first recordings believed to be spirit voices were such messages as "This is G!", "Hot dog, Art!", and "Merry Christmas and Happy New Year to you all". Von Szalay and Bayless' work was published by the Journal of the American Society for Psychical Research in 1959. Bayless later went on to co-author the 1979 book, Phone Calls From the Dead.
In 1959, Swedish painter and film producer Friedrich Jürgenson was recording bird songs. Upon playing the tape later, he heard what he interpreted to be his dead father's voice and then the spirit of his deceased wife calling his name. He went on to make several more recordings, including one that he said contained a message from his late mother.
 Raudive voices
Konstantin Raudive, a Latvian psychologist who had taught at the University of Uppsala, Sweden and who had worked in conjunction with Jürgenson, made over 100,000 recordings which he described as being communications with discarnate people. Some of these recordings were conducted in an RF-screened laboratory and contained words Raudive said were identifiable. In an attempt to confirm the content of his collection of recordings, Raudive invited listeners to hear and interpret them. He believed that the clarity of the voices heard in his recordings implied that they could not be readily explained by normal means. Raudive published his first book, Breakthrough: An Amazing Experiment in Electronic Communication with the Dead in 1968 and it was translated into English in 1971.
Spiricom & Frank's Box
In 1980, William O'Neil constructed an electronic audio device called "The Spiricom." O'Neil claimed the device was built to specifications which he received psychically from George Mueller, a scientist who had died six years previously. At a Washington, DC, press conference on April 6, 1982, O'Neil stated that he was able to hold two-way conversations with spirits through the Spiricom device, and provided the design specifications to researchers for free. However, nobody is known to have replicated O'Neil's results using their own Spiricom devices. O'Neil's partner, retired industrialist George Meek, attributed O'Neil's success, and the inability of others to replicate it, to O'Neil's mediumistic abilities forming part of the loop that made the system work. However, there is strong evidence to suggest that the recordings of conversations were falsified by O'Neil, specifically with an electrolarynx. The clearly audible vocal fricatives in the recordings, along with the fact that during the hours of recordings O'Neil's and Mueller's voices never overlap (as would happen in normal conversation), support this theory.
Another electronic device specifically constructed in an attempt to capture EVP is "Frank's Box" or the "Ghost Box". Created in 2002 by EVP enthusiast Frank Sumption for supposed real-time communication with the dead, Sumption claims he received his design instructions from the spirit world. The device is described as a combination white noise generator and AM radio receiver modified to sweep back and forth through the AM band selecting split-second snippets of sound. Critics of the device say its effect is subjective and incapable of being replicated, and since it relies on radio noise, any meaningful response a user gets is purely coincidental, or simply the result of pareidolia.
Grabbed From The Wikipedia