A strange voice on a recording can be compelling—but “unexplained” does not automatically mean paranormal.
A quiet room. A recorder running. Nobody remembers speaking. Later, during review, a voice appears. It may sound like a name, a warning or a direct answer to a question. These moments can be among the most compelling parts of a paranormal investigation—but they are also among the easiest to misinterpret.
At Hide and Seek Paraquest, curiosity is not enough. If a recording deserves to be treated as possible evidence, it must survive questions about contamination, equipment, handling and human perception.
The goal is not to prove every unusual sound paranormal. The goal is to determine what remains unexplained after reasonable possibilities have been examined.
What is an EVP?
Electronic Voice Phenomenon, commonly shortened to EVP, describes a voice-like sound discovered on an electronic recording that was not consciously heard or identified at the time. That definition does not establish its source.
An apparent voice could originate from a person speaking quietly elsewhere, movement or clothing noise, radio-frequency interference, plumbing, animals, wind, structural sounds, recorder processing, overlapping ordinary sounds or auditory pareidolia—the brain finding meaningful speech in ambiguous noise. Its source may also remain unidentified.
“Unidentified” is an honest result. It is not a failure—and it is not automatic proof of spirit communication.
The investigation begins before recording
Good EVP work begins with control of the environment. Before a session, investigators should identify who is present, note nearby roads and occupied rooms, and listen for machinery, wildlife, plumbing, ventilation and other repeating sounds.
Mobile phones and radio-transmitting equipment may need to be moved away from recorders or placed into an appropriate offline mode. Investigators should also avoid loose keys, jewellery or clothing that can brush against microphones.
Tag every sound
During recording, every known sound should be identified aloud. If someone coughs, say “cough”. If an investigator changes position, say “movement”. If a car passes, a door closes or somebody whispers, mark it verbally.
Investigators should also identify themselves before speaking. A simple name tag—“John speaking”—can prevent an ordinary comment from becoming mysterious during review. Nobody will remember every sound several hours later. Real-time tagging creates a record that memory cannot provide.
Protect the original file
The original recording should never be enhanced, trimmed or overwritten. Make a separate working copy and perform all review or enhancement on that copy. Preserve the original filename, format and metadata wherever possible.
Forensic-audio guidance recommends obtaining the original recording or earliest available generation. Native files can contain information about the recording system, timestamps, metadata and settings. A strong evidence record should include the original file, an untouched backup, a working copy, recorder details, date, time, location, investigators present, session notes, relevant video and details of every alteration made to a presentation copy.
Enhancement can help listeners hear a quiet section, but it can also change how noise is perceived. Any noise reduction, equalisation, amplification or filtering should be documented. The enhanced version is an interpretation aid—not the original evidence.
Review without leading the listener
One of the biggest EVP mistakes happens after capture: telling everybody what the recording supposedly says. Once a listener is shown a suggested phrase, it becomes difficult to hear the clip without that expectation.
A better review process is to play the clip without subtitles, ask reviewers independently whether they hear a voice, have them privately record any words they perceive, and compare answers only afterwards. Disagreement should be documented rather than forced into consensus.
If five listeners hear five different sentences, the recording is highly ambiguous. If independent listeners consistently hear the same phrase, that makes it more interesting—but it still does not establish the speaker or source.
Use multiple recordings when possible
A second recorder can provide valuable context. If a sound appears clearly on one recorder but not another positioned nearby, investigators should consider microphone direction, distance, sensitivity, processing and local interference.
If the same sound appears across multiple independent devices at corresponding times, it becomes easier to confirm that an acoustic event occurred in the environment. Video may also reveal movement, whispering or equipment handling at the critical moment. Audio should not be reviewed in isolation.
Classify the result honestly
Not every unusual recording should be posted as paranormal evidence. HSPQ may classify a clip as explained, likely contamination, ambiguous, unexplained or potentially significant. Even the final category should be presented as possible evidence—not proof.
What makes an EVP worth sharing?
A responsible public presentation gives people enough context to form their own view. Where privacy and venue agreements permit, that means providing an unlabelled clip first, the proposed interpretation afterwards, relevant questions, who was present, known environmental conditions, whether the sound was heard live, whether other devices captured it and any enhancement applied.
Subtitles can be useful—but only after an unprompted listen. HSPQ’s responsibility is not to make a faint recording sound more dramatic. It is to show what was captured, explain how it was assessed and remain honest about what cannot be known.
The value of uncertainty
An investigation does not become less compelling because some recordings have ordinary explanations. Removing footsteps, whispers, passing vehicles and equipment artefacts protects the integrity of anything that remains.
A clip may still resist explanation. It may be unusually clear. It may appear to respond directly to a question. Those moments deserve attention—but not certainty they have not earned.
The most credible conclusion is sometimes the simplest: we captured something unusual, we tested what we could, and we do not yet know what caused it.
Expect Nothing. Question Everything.