What is electronic voice phenomenon?
Electronic voice phenomenon, usually shortened to EVP, means sounds on an electronic recording that seem to be voices but weren’t heard by anyone present when the recording was made. In ghost hunting and parapsychology (the study of claimed psychic and paranormal events), these sounds are usually taken to be voices of the dead. Some enthusiasts suggest other sources too: nature spirits, beings from other dimensions, extraterrestrials, or the living mind pressing thoughts onto the recording by psychokinesis.
The classic EVP is short, often a word or brief phrase. It usually sits inside static, hum or background noise rather than in silence. Konstantīns Raudive, the Latvian psychologist who brought the idea to a wide audience, described the voices this way, and most modern recordings match that pattern.
The scientific mainstream has a different name for the experience: auditory pareidolia, the brain’s habit of hearing familiar speech in random sound. Researchers also point to radio interference, equipment artifacts and outright hoaxes. Electronic voice phenomenon is seldom studied inside mainstream science, so most of the work has come from amateur investigators. That’s a big part of why the debate has gone on so long.
EVP sits alongside a wider family of paranormal claims covered in our paranormal accounts and hauntings section, and it’s become a standard tool at nearly every location on lists like the most haunted places in America.
Where EVP came from: Spiritualism, Edison and the first tapes
The roots go back to Spiritualism, the religious movement that thrived from roughly the 1840s to the 1940s and held that mediums could contact the spirits of the dead. Spiritualists picked up every new technology of their era, photography most famously, hoping to show the spirit world was real. As sound recording spread, mediums tried it as well.
Thomas Edison became part of the story when an interviewer from Scientific American asked him whether his inventions could reach the dead. He answered that if spirits could only exert faint influences, a sensitive recording instrument would have a better chance than table tipping or Ouija boards. The record after that is genuinely mixed. His diary from the 1920s contains speculative notes on the subject, and in 1926 he told the New York Times his earlier remarks had been a joke. On the other hand, a final chapter of his memoir discussing the idea was cut from later editions and turned up again in 2015 in a 1949 French translation. No evidence shows he ever built a working device. Whether he was serious is still disputed.
Attila von Szalay and the cabinet recordings
The first sustained attempts to record spirit voices are generally credited to the American photographer Attila von Szalay, who was already trying to photograph ghosts. He started in 1941 with a 78 rpm record. By his own account he didn’t succeed until 1956, after switching to a reel-to-reel tape recorder.
Working with researcher Raymond Bayless, von Szalay built a microphone into an insulated cabinet wired to an outside recorder and speaker. They reported sounds on the tape that hadn’t come through the speaker live, some recorded while the cabinet was empty. Among the phrases they reported were “This is G!” and “Hot dog, Art!” Their work appeared in the Journal of the American Society for Psychical Research in 1959. Bayless later co-wrote the 1979 book Phone Calls From the Dead.
Jürgenson, Raudive and the Breakthrough era
The modern history of electronic voice phenomenon usually begins in Sweden in 1959. Friedrich Jürgenson, a painter and film producer, was recording bird songs. On playback he heard what he took to be the voices of dead family members, including one he believed was calling his name. He kept recording and became convinced he was picking up messages from the dead.
Konstantīns Raudive, a Latvian psychologist who had taught at Uppsala University, worked with Jürgenson and then took the work much further. He made more than 100,000 recordings that he believed were communications with discarnate people, meaning people without bodies. Some sessions were held in a laboratory shielded against radio frequencies, which was meant to rule out broadcast interference. He invited listeners to hear the recordings and write down what they heard, and he argued that the voices were too clear to explain by normal means.
His book Breakthrough: An Amazing Experiment in Electronic Communication with the Dead came out in 1968 and in English in 1971. It made EVP an international topic, and for years the voices were often called “Raudive voices.”
Organizations and a broader idea
In the 1970s the German researcher Ernst Senkowski coined Instrumental Trans-Communication (ITC), a wider term for supposed contact with spirits through any electronic device: tape recorders, televisions, fax machines, computers. One famous ITC claim says Jürgenson’s image appeared on a colleague’s television, tuned to an empty channel, on the day of his funeral.
In 1982 Sarah Estep founded the American Association of Electronic Voice Phenomena in Severna Park, Maryland, a nonprofit set up to raise awareness and teach standard recording methods. Estep had begun recording in 1976 and said she had captured hundreds of messages from dead friends and relatives, and from beings she speculated came from other planets or dimensions.
Spiricom, Frank’s Box and how a ghost box works
As the field grew, some investigators stopped waiting for voices to turn up in noise and built machines meant to produce them.
The Spiricom
In 1980 William O’Neil built a device he called the Spiricom. He said the design came to him psychically from George Mueller, a scientist who had died in 1967. At a press conference in Washington, DC on April 6, 1982, O’Neil’s partner George Meek, a retired industrialist, said the device allowed two-way conversations with spirits, Mueller among them, and released the plans to researchers for free.
Nobody is known to have reproduced O’Neil’s results with their own Spiricom. Meek’s explanation was that O’Neil’s mediumistic ability was part of what made the system work. In 2020 the skeptical investigator Kenny Biddle published a detailed critique of the device’s origins and claimed science after it reappeared on paranormal television.
Frank’s Box and the modern ghost box
In 2002 EVP enthusiast Frank Sumption built what became known as Frank’s Box, or the ghost box. He too said the instructions came from the spirit world. The idea was simple: combine a white noise generator with an AM radio modified to sweep back and forth across the band, catching split-second fragments of sound.
Today’s spirit boxes work much the same way. A radio scans stations continuously without muting between them. The user asks questions out loud, and the stream of clipped words, syllables and static is treated as the reply. Paranormal researcher Ben Radford has called the ghost box a modern version of the Ouija board. Critics say the effect is subjective and can’t be replicated, and since the audio is real radio, any fitting answer is chance plus the listener’s mind.
What controlled tests have found
Only a handful of serious attempts have been made to test electronic voice phenomenon under controlled conditions. Two stand out.
In 1997 psychologist Imants Barušs of the University of Western Ontario ran an experiment modeled on Raudive’s methods and on the work of ITC researcher Mark Macy. A radio was tuned to an empty frequency. Over 81 sessions the team collected just over 60 hours of recordings while a participant either sat silently or tried to make contact. Barušs did find a few sounds that resembled voices, but they were rare, random and open to interpretation. His conclusion, published in the Journal of Scientific Exploration in 2001, drew a clear line. EVP had been replicated in a weak sense, since voice-like sounds did show up on tape. It had not been replicated in a strong sense, since nothing was clearly anomalous, let alone traceable to the dead.
In 2005 the Journal of the Society for Psychical Research published a report by investigator Alexander MacRae, who used a recording device of his own design. He asked seven listeners to match selected clips against a list of five phrases he provided. MacRae said the panel’s agreement pointed to a paranormal origin. Critics note a weakness in that design: giving listeners a short list of candidate phrases tells them what to listen for, which is exactly the kind of priming that shapes what people hear.
| Documented | Claimed or disputed |
|---|---|
| Von Szalay and Bayless published cabinet recordings in 1959 | That the sounds were voices of the dead |
| Raudive made over 100,000 recordings and published Breakthrough in 1968 | That their clarity rules out normal explanations |
| The Spiricom was publicly presented in 1982 | That it carried two-way conversations with a dead scientist |
| Barušs’s 81-session study found a few voice-like sounds | That any of them were anomalous |
| Ghost boxes sweep real radio broadcasts | That spirits pick the fragments heard |
Natural explanations: pareidolia, radio and the recorder itself
Skeptics don’t claim people are imagining that they hear something. The sounds are on the recording. The question is what made them and what turns them into words.
The brain that hears speech everywhere
Human hearing is tuned hard for speech. Feed the brain enough noise and it will find words. That’s auditory pareidolia, and its cousin apophenia is the broader tendency to find meaningful links in random information. One telling detail: EVP voices almost always speak the investigator’s own language. Researcher Joe Banks has gathered these effects under the name “Rorschach Audio,” after the inkblot test. Psychologist James Alcock has put it bluntly: in his view EVP comes from hope and expectation, and the claims don’t survive scrutiny.
Expectation matters a great deal. In a 2019 investigation of a supposedly haunted portrait at a West Virginia museum, Kenny Biddle looked at an EVP said to speak the name “Annie,” the name written on the back of the painting. The audio came from a radio modified to scan stations without muting. Biddle pointed out that ordinary broadcast words like “any,” “many” or “company,” or a phrase like “and he,” can easily sound like “Annie” to someone already listening for it. The same priming that shapes memory, which we’ve looked at in the Berenstain Bears Mandela Effect, shapes hearing too. Our consciousness and the mind section has more on how perception fills in gaps.
Radio, circuits and capture errors
Electronics leak. Broadcast radio, CB transmissions and wireless baby monitors can all bleed into recording equipment, and circuits can even resonate with incoming radio signals without any power of their own. Cross-modulation, where two signals mix inside a device and create new ones, and faulty ground loops add more layers.
Then there’s the noise floor, the faint electrical hiss every device makes. Skeptics including David Federlein, Chris French, Terence Hines and Michael Shermer note that many EVP sessions raise the gain until that hiss is loud, then filter it. Federlein compares the result to a guitar wah pedal: a sweeping filter across noise produces open, vowel-like sounds that can resemble speech. Over-amplification at the moment of recording creates its own artifacts, and so does later “enhancement” through noise reduction, resampling or isolating frequencies, which can sculpt something voice-like out of very little.
Old tape had its own oddities. When parapsychologist David Fontana suggested in 1995 that a poltergeist might have interfered with a researcher’s tape recorder during the Enfield case, media expert Tom Flynn offered a mechanical explanation: tape sticking to a guide post in humid conditions and folding into a loop inside the machine.
How to judge an EVP recording
Whatever you believe, a recording is only as good as the conditions it was made in. If you investigate a location, get permission first, and never enter abandoned or unsafe buildings.
Can you use your phone, or do you need an EVP recorder?
Plenty of electronic voice phenomenon recorders are for sale, but they’re essentially ordinary digital voice recorders sold to ghost hunters. A phone works too, with one big caveat. Phones are built to make speech clear, so they apply noise suppression, automatic gain and compression. On a near-silent recording, that processing can pump and smear background noise into syllable-like shapes. Some cheap recorders are also poorly shielded, which makes them more likely to pick up radio. The device matters far less than the method.
A worked example
Picture a team in an old farmhouse. Someone asks, “Is anyone here?” On playback there’s a breathy sound two seconds later that one member hears as “yes” and another as “leave.” Here’s how to test it:
- Check the log. Was anyone moving, whispering, or wearing a jacket that rustles? Good teams call out (“my stomach,” “car outside”) so those sounds are tagged.
- Compare a control. Record the same spot in silence with the same settings. If similar sounds show up when nobody is asking anything, the “answer” is probably room noise.
- Listen to the raw file. If the sound only appears after filtering or amplification, the processing may have made it.
- Do a blind test. Play the clip to people who don’t know the question or the place, and ask them to write down what they hear. If their answers are all over the place, the clip doesn’t hold a clear message.
- Look for radio. Music, a station ID or a phrase in another speaker’s rhythm points to interference.
Common mistakes
- Telling listeners what the voice “says” before they hear it.
- Looping one short clip dozens of times until a word forms.
- Stacking enhancement filters and treating the result as the original.
- Deleting the files where nothing happened, so the hits look more common than they are.
- Treating ghost box fragments as replies when they’re broadcast audio.
What remains unexplained about electronic voice phenomenon
The strongest ordinary case is solid. Pareidolia is well established, interference is documented, enhancement artifacts are easy to reproduce, and the one well-known controlled study found nothing clearly anomalous. The Spiricom was never replicated, and ghost boxes are made to produce word fragments.
The case for something stranger rests mostly on individual experience. Investigators describe replies that seem to fit a question precisely, names that match a location’s history, and voices recorded in shielded rooms. Some enthusiasts say hearing EVP is a skill you learn, like a new language. Skeptics answer that a trained ear may simply be a more primed one. Supporters say no one has ruled out every case, and that’s true. No one has shown that any single case needs a paranormal explanation, either.
What really persists is the human side. The wish to hear lost voices is ancient. It runs through traditions like the banshee of Irish folklore and the weeping figure of the La Llorona story, and every new technology has been asked to carry it. Electronic voice phenomenon is that old question asked through a microphone.
For long-form conversations about hauntings, strange signals and the edges of perception, the Troubled Minds Vault holds the radio show’s archive, and the full journal has more cases weighed on the record.
Questions and answers
Can you use your phone as an EVP recorder?
Yes. Any phone with a voice memo app can capture the kind of audio EVP investigators study. The catch is that phones process sound heavily. Noise suppression, automatic gain and compression are built to make speech clear, and on near-silent recordings they can smear, pump or shape background noise into something that sounds a lot like a syllable. If you use a phone, record a silent control clip in the same spot and listen to both before reading anything into a sound.
Can spirits come through electronics?
Many EVP enthusiasts and Instrumental Trans-Communication researchers believe they can, and they point to recordings, radio sweeps and odd phone calls as evidence. Nobody has shown it under controlled conditions. The best-known controlled attempt, run by psychologist Imants Barušs in the late 1990s, picked up a few voice-like sounds but nothing clearly anomalous. Known causes like radio interference, cross-modulation and auditory pareidolia explain the cases that have been looked at closely.
How does a ghost voice box work?
A ghost box, also called a spirit box or Frank's Box after Frank Sumption's 2002 design, is basically a radio that sweeps through AM or FM stations without muting the sound in between. It plays split-second fragments of broadcasts mixed with static. Users ask questions and listen for answers in the fragments. Critics point out that the sound is real radio audio, so any word that seems to fit the question is chance plus the listener's expectation.
What does electronic voice phenomenon mean?
Electronic voice phenomenon, or EVP, is a sound on an electronic recording that seems to be a voice and wasn't heard when the recording was made. Ghost hunters and parapsychologists usually take these sounds to be voices of the dead or other unseen beings. Scientists generally class them as auditory pareidolia, which is the brain hearing familiar speech in random noise, along with interference and equipment artifacts.
Do you need a special EVP recorder?
No. Plenty of EVP recorders are for sale, but they're ordinary digital voice recorders, sometimes with extra gain or marketing aimed at ghost hunters. A special device doesn't make a recording more convincing. What counts is control: knowing where your mic is, logging every sound in the room, avoiding heavy noise reduction, and keeping untouched original files. Cheap recorders with poor shielding can actually pick up more radio interference than better ones.
Stay curious. Keep the question open.
Follow the evidence wherever it leads.
