EVP-AVP Analysis, Based on Observations and Experiments of MNPARA-SCI
February 24 2013 - MNPARA-SCI
** Also note this page is still under development and will be edited as I spot grammar errors and gather more information about my observations. (Thanks MNPARA-SCI)
To understand how MNPARA-SCI’s (my) observational research applies to paranormal phenomena; it is highly suggested that you become familiarized with MNPARA-SCI’s Primer on Raudive Phenomena; and Paranormal Theory Primer. With that being said let’s get started. In the course of researching Raudive Phenomena (aka: electronic voice phenomena & acoustic voice phenomena) I have made some observations in regards to the identification and analysis of these events. If you are looking for a collegiate level physics lecture with supportive mathematical formulas; or advice from a certified audio engineer, I’m probably not going to meet your expectations. Please note that I am not a certified sound engineer, nor do I hold any degrees in physics or related sciences. I may have some of the technical jargon askew, but the concepts I am going to cover appear solid.
Lets start with the most obvious observation to make, that raudive phenomana are voices. Just like your voice or mine voice, they have certain characteristics that can be observed recorded and measured. So the first logical step in the location and analysis of raudive phenomena, is to know how to spot and recognize all voices as they appear on recorded media.
To spot voices on your recorded files, whether they are human or from some other source, you need to find a simple visual way to examine sound. This first way to examine sound is to use the most common feature found with most audio software packages. That being the waveform display. Let’s take a look at a human voice using the waveform view on audacity. Note: If you click on any of the images here, a full-sized image should open on your screen.
WAVEFORM VIEW – ANALYSIS OF VOICES or VOCAL EVENTS
This is me speaking in a normal speaking voice. This clip is a sound file that was captured using a dynamic microphone, jacked into an ART Pro Series Pre-amp set at 23db pre-amp gain, and ported into a two channel Sony digital audio recorder. You can see several distinct features.
First, let’s examine the horizontal central blue line located at (0.00 db) which literally cuts the image in half, this is the noise floor. Anything that is captured by you recorder, in the form of sound or EMF (Electromagnetic Fields), will emerge from the noise floor. The thickness of this line depends on the amount of ambient noise there is in the environment. Second, let’s look at my voice. It emerges from the noise floor like a sideways pine tree. Using the recording setup I am using, this waveform appears to emerge equally both above and below the noise floor. In some instances this will not be the case depending on your recorder and software package used. In this simple view, you get a visual impression on the amplitude, power of the captured event. There are ways to figure the frequency of the sound or event using this view but it is complicated (Icky, requires math….)
(Text Source Wikipedia) In signal theory, the noise floor is the measure of the signal created from the sum of all the noise sources and unwanted signals within a measurement system, where noise is defined as any signal other than the one being monitored. In radio communication and electronics, this may include thermal noise, blackbody, cosmic noise as well as atmospheric noise from distant thunderstorms and similar and any other unwanted man-made signals, sometimes referred to as incidental noise.
If you want to further examine the event you captured, there is a solid tool available on audacity which does not require mathematical formulas. Use the frequency analysis tool on audacity to do the work for you. By selecting a small portion of a waveform, and using the frequency analysis function of Audacity, you get this plotted cool frequency plotting diagram. You read this diagram on two axis, start from left to right and that gives you the frequency range. You read it from bottom to top to get the amplitude or power registered.
Looking diagram plot you can see the most powerful portion on my speech is around 350-500 htz at 35db; and as you proceed farther to the right in the plot of my voice it loses power (amplitude); but it still produces additional resurgent amplitude spikes. These additional spikes harmonic sound frequencies and are also evidence of phonemes and formants (word sounds) being formed; as the sound is being manipulated by the larynx, pharynx, palette, nasal passages, glottis, tongue, or some other mechanism.
In many cases the plot of a vocal event will form a sagging “power line” or “laundry line” appearance. In the cases of some consonants and vowels, the appearance could be less pronounced or rounded, but there will still be activity spikes that give a jagged or noisy appearance.
This image is a frequency plot of the noise floor. As you can see in this frequency plot, it is just like the blue line in the wave form and is pretty homogenous. There are time when things are hidden in the noise floor, watch for unusual spikes of patterns in that area. If you spot frequency plots that look like vocal events you will want to examine that area of the recording closely.
Note: Due to the white noise factor, this area is heavily prone to apophenia and pareidolia. However it should be noted that raudive phenomena can be found inside the noise floor; but they are typically weak and could need extensive noise removal and audio processing to clear up. Be careful with noise removal, as it can create artifacts of “sculpted” white noise, which can look and sound a lot like a voice. I advise leaving these weak phenomena out of evidence level presentations, but feel free to play around with them to learn the limits of your software.
(Source Wikipedia) Pareidolia (pron.: /pærɨˈdoʊliə/ parr-i-DOH-lee-ə) is a psychological phenomenon involving a vague and random stimulus (often an image or sound) being perceived as significant. Common examples include seeing images of animals or face in clouds, the man in the moon or the Moon rabbit, and hearing hidden messages on records when played in reverse. The word comes from the Greek words para (παρά, “beside, alongside, instead”) in this context meaning something faulty, wrong, instead of; and the noun eidōlon (εἴδωλον “image, form, shape”) the diminutive of eidos. Pareidolia is a type of apophenia, which is seeing false patterns in random data.
(Source Wikipedia) Artifacts: In sound and music production, sonic artifact, or simply artifact, refers to sonic material that is accidental or unwanted, resulting from the editing or manipulation of a sound. Because there are always technical restrictions in the way a sound can be recorded (in the case of acoustic sounds) or designed (in the case of synthesized or processed sounds), sonic errors often occur. These errors are termed artifacts (or sound/sonic artifacts), and may be pleasing or displeasing. A sonic artifact is sometimes a type of digital artifact and in some cases is the result of data compression (not to be confused with dynamic range compression, which also may create sonic artifacts). Often an artifact is deliberately produced for creative reasons; for example, to introduce a change in timbre of the original sound or to create a sense of cultural or stylistic context.
When I visually inspect recordings for possible events I start with the waveform. For me it has been a good method to quickly examine recordings for major events; but I strongly recommend listening to the complete audio file, because sometimes there are things in the noise floor that will surprise you! When I find suspect events, I use another feature of most audio software packages, that being the spectrogram view.. With that lets make the jump to the next step in my observational research.
SPECTROGRAM VIEW – ANALYSIS OF VOICES or VOCAL EVENTS
A spectrogram, or sonogram, is a visual representation of the spectrum of frequencies in a sound. Spectrogram can be used to identify spoken words and to analyze various calls of animals. They are used extensively in the fields of music, sonar, radar, speech analysis, seismology, geology and many other areas. I find this function of most software packages to be invaluable in raudive phenomena research. Using this view I can look at various structural aspects of a possible vocal event.
Note: Use the above image as reference to the following screening questions. In this recording there are only human voice sounds. There is no evp/avp present in this recording. You can see normal volume speech, whispered pre-words, and breathe sounds. The following questions should be used to screen your own suspected events when examined from a spectrogram standpoint.
WHEN YOU ARE EXAMINING A VOCAL EVENT USING A SPECTROGRAM ASK THE FOLLOWING QUESTIONS.
1. Does the spectrogram image show an event with vertical columnar structure(s); that are travelling upwards indicating a frequency range? Most human voices will start somewhere around 85hz and progress upwards, sometimes fading out at 10-15khz (10,000-15,000 hz), however the usable portion of the voice, where formants and phonemes (aka- word sounds) occur, will be found below 5000 hertz.
2. Are there fine frequency structures, within the columnar structure, that appears as “fine horizontal stacked bands or chips?” This is a fundamental structure of a vocal events, think of this as the raw material that makes up of the event. The more defined and dark these features are, typically represents a higher order of volume or proximity to the recorder.
3. Are there darker bands within the fine frequency structures, or in their absence, within the columnar structure? This indicates formants and phonemes (word sounds) are present. They also indicate the fundamental and harmonic frequencies of the vocal event. These dark bands are representative of manipulation of the sound by the larynx, glottis, tongue, nasal chambers, and mouth… or some other mechanism of sound shaping.
Now, in the case of breath sounds, pre-spoken vocalizations, whispered events, distant events, and events where multiple voices are speaking at the same time; the spectrogram image can appear blurry, faint or fuzzy. However you should be able to spot some traces of at least one of the above characteristics.
** Some EVP/AVP events, will appear above the noise floor, as if they just emerged from nowhere. They appear “blob-like” and not similar to other vocal events… In screening these events, look very close at the event itself. You should still be able to make out at least large-scale formant – phoneme structures. The “blob-like” structure itself, is in all likelihood is a large formant-phoneme structure. Sometimes these structures will appear above the normal range of human formant range of 5000 hertz!
SO YOU FIND ALL THE ABOVE TO BE TRUE, WHAT DO YOU HAVE THEN?
That’s Great! What you have at this point is a probable vocal event. Vocal events include incidents of you talking, out loud playback of recorded speech, text to speech playback voices, audible sounds of speech coming from active radios or other electronic devices, other investigators talking, and audible voice phenomena & electronic voice phenomena. To narrow it down farther requires some work on your end. This brings us to the second most obvious observation made, that raudive phenomena emerge from an unknown source. Since no one has been able to absolutely lockdown that source, you are left with having to eliminate all other sources before you can conclude that what you have captured is raudive phenomena.
You have to eliminate the possibility that the event was not the product of a radio or tv turned on, someone talking nearby, animals in or near the recording site, you or another investigator accidently speaking out loud, grumbling stomachs, flatulence, and a litany of other possible sounds. Once you can eliminate these possibilities, you can start suspecting that you may have an electronic voice phenomena or an acoustic voice phenomena event. The best methodology of doing this, is to establish solid protocols for your recording sessions and investigations! Some suggested protocols include the following:
PROTOCOL – SITE SECURITY: This is not as simple as it seems. You have to ensure there are not contaminants that are entering your investigation site. Whether that is people, pets, cell phones, radio’s, or other environmental factors that can lead to misidentified sounds on your recordings! If you can’t eliminate them; then you have to account for them by identifying them by marking them on your recordings. Marking is a common practice of verbally announcing when a real world event occurs that is audible to the investigator, example: “Dog barking outside…” “I coughed…” “Stomach growled…” Etc.. etc…
PROTOCOL – ACCIDENTAL SPEECH & CHATTER: All of us tend to mutter or mumble stuff inadvertently, or whisper responses unconsciously; it’s part of how we vocalize thoughts, and you must plan to deal with that. I would recommend video recording all investigations, and make everyone’s faces are in full view of the camcorder… You can eliminate most instances by referring back to the recording when you have doubt. In fact it’s not a bad idea to actually review the video whenever you have a possible EVP/AVP captured. If you are working with a group of investigators… make sure the chit chat is out of the way prior to the investigation. That way you avoid the whispers and laughs of folks trying to play catch-up with each other since the last investigation!
PROTOCOL – LEARN INCIDENTAL SOUNDS Listen to recordings of breath sounds, animal noises, bird chirping, water heater noises, furnaces, your friends voices, and other common misidentified sounds. Examine them in waveform and spectrogram. That way when it comes down to it, you can eliminate the false positives. Example: Naval Sonographer’s have to learn the various sounds of aquatic wildlife and other natural events, to identify and eliminate them as false hits, when using passive sonar to locate nearby objects. As an EVP/AVP researcher you have to enlist the same line of thinking.
PEER PANEL REVIEW – THINGS YOU HAVE TO DO TO AVOID PAREIDOLIA!
Ok, now that you’ve set up protocols, inspected your files in waveform and spectrogram, and are ready to make the next step. This is perhaps the most important step of the process. You need to send the sound file to a blind panel of people for review. Typically you should email it to 10 people whose opinions you trust; or if you really want to put it out there, post it on trusted Facebook group for your peer’s evaluation.
Many researchers are afraid of doing this. The first dozen or so times you do this, expect to get frustrated when the returned results don’t match what they expected. Don’t be… This is a required part of any research process. Sometimes it comes from the audio quality of your posted clip, or the equipment that the peer used to listen to evaluate it. Always encourage your peer panel to use quality headphones when reviewing your clips to assure the best chances of a solid review.
Note: After you have reviewed your own material, and believe that you have captured something. Especially after you think you have a voice saying a particular word. You are subject to subjective influence. Your brain can trick you after that point and you will always hear that specific word, even if it isnt. Case example how many times have you caught yourself singing the wrong lyrics to a popular song, but you were positive you had it right, until someone pointed it out to you. Its a common bias and error in research as well. That is the reason peer panel reviews of your captures is essential!
Note: Please avoid asking the trigger question “Is this an EVP?” It’s like dumping blood in the water around sharks. You’re opening up a can of speculative conjecture and personal opinions. Remember that there are true believers out there who will say everything is an EVP/AVP. There are those who are hyper skeptical and who will shoot down all EVP/AVP. Most of these folks mean well, but can skew your panel if you don’t regulate the questions and response options. Then there are the trolls, the bullies out there that look for openings on a newbie, and will verbally attack once the open ended question is posed.
The only thing your panel should be allowed to determine is whether the event captured sounds like a voice. Was the event on the recording easy to locate. Do they hear any specific words or phrases being spoken? Remember, it’s your job to determine if the capture was an EVP or not. Unless your panel is comprised of team members who were actually present at the time, no one on your panel was there to observe your investigation. They were not there to verify you followed proper protocols. They will not know your abilities to conduct a solid investigation. So when you solicit your peers for a review, regulate the responses. If someone goes out of there way, to explain something to you during a peer review, don’t necessarily throw it out… but use your own best judgment. Never… and I mean never… get into an argument over a panel review, it is unprofessional and hurts your own image.
When you solicit a panel review, I would use something like the following. (Email Example)
Hi there, (Insert name)
Could you please review the following audio clip, and select the best answer that applies to what you hear on it?
___ I Heard Something, It was easy to locate, It clearly sounds like a voice, It clearly says a specific word or phrase. (Please return comment with what you think it said!)
___ I Heard Something, It was easy to locate, It clearly sounds like a voice, but I can not positively identify a specific word or phrase.
___ I Heard Something, It was not easy to locate, but it does sound like a voice.
___ I Heard Something, I do not know what it is.
___ I Heard Something but it is not a voice.
___ I Heard Nothing.
Thanks for helping me in this matter!
Sincerely, (Insert your name and email address)
USING PANELS TO CLASSIFY RECORDINGS: There are those that get hung up on classifying recordings into groupings by there clarity qualities. If you are planning on using the results of your panel review to classify your recordings, I would use the following rule of thumb.
CLASS A Recordings, require a significant majority of the panel responses to come back telling you the event captured: 1) is clearly found. 2.) is clearly a voice. 3.) is saying the same word or phrase. I would recommend that the majority be 80% in agreement or greater. Which would boil down to at least 8 people out of 10 polled in agreement.
CLASS B Recordings, require a simple majority of the panel responses to come back telling you the event captured: 1) is clearly found. 2.) is clearly a voice. I would recommend that the majority be 70% in agreement or greater. Which would boil down to at least 7 people out of 10 polled in agreement.
CLASS C Recordings, require a simple majority of the panel responses to come back telling you the event captured: 1) it can be found. 2.) it could be a voice. I would recommend that the majority be 60% in agreement or greater, be the panel threshold. Which would boil down to at least 6 people out of 10 polled in agreement.
ALL Other Recordings, If your clip does not meet the above requirements after a panel review, I would just call the clip “interesting noise” or an “anomalous recording” and move on to the next clip or investigation.
SO THEN, HOW DO I ESTABLISH, THAT I HAVE AN ACTUAL EVP or AVP?
Remember, it’s your job to determine if your capture was an EVP or not. So lets look at what you’ve accomplished to this point.
* You’ve established and used protocols to eliminate contamination and misidentification of event source.
* You have recorded an event, that you know for a fact, was not said by you or anyone else on location.
* You have examined it via waveform and spectrogram, and forensically established it as being consistent with characteristics of a vocal event.
* Then you sent the recording out to your peers for review. They have validated it is a clearly audible voice, maybe it is saying something specific, and you’ve eliminated the chance of pareidolia or subjective influence.
So what does that sound like to you? Sounds like you have enough evidence at hand, to make our own decision! Now if you are looking for a specific litmus test to hold your recording up to. Maybe an expert who can blindly look a recording and tell you one way or the other if its real. You’re going to be disappointed. There is no magic test or accepted standards. There are no absolute certified experts in this field. You have to learn from your peers, and become your own expert.
Note: You of course can start outsourcing your files to certified audio engineers to look for indications of RF based interceptions like radio transmissions, just remember that this is pretty rare when using newer and undamaged equipment. You could send your recordings to a speech pathologist, to absolutely establish that you have a voice on your recording… Just remember unless they are your friends, you are going to end up paying for these services. Plus in the end, they can only tell you if it’s a voice… or if its radio (rf) contamination. They will not be able to say it is a EVP or AVP. That’s still your job.
Now For Some Fun
A Basic AVP/EVP Analysis Example.
In 2010 my son and I went to a secluded turn of the century homestead site. There we captured one of the best suspected EVP/AVP I have on file to date. Please listen to the file before proceeding; as the images and descriptions taken are based on this case file. Can you guess what the possible voice in this clip is saying? Feel free to leave a comment on the bottom of this page! (Call it Practice Peer Review!)
You can see the base line sound floor cutting through the middle fo the screen. The first large spike area is my voice followed by an unregistered whisper by my son. There is a crackling (believed to be an emergent event) followed by me talking again. Shortly there after, a voice announces itself. Which when you play the YouTube video for yourself, you can try to figure out what it’s saying. The large sound event on the far right portion of the screen is an amplified version of the EVP/AVP.
Here is a frequency plot of my voice. You can clearly see the declination from left to right. The base frequency, harmonics, and the “clothing line” effect showing formants (word sounds) being clearly observed.
Here is a frequency plot of me whispering. You can clearly see the declination from left to right. The base frequency, harmonics, and the “clothing line” effect showing formants (word sounds) is clearly observed.
This is the EVP/AVP in frequency analysis. The base frequency, harmonics, and the “clothing line” effect showing formants (word sounds) is clearly observed.
Using Audacity Spectrogram view in false color mode, going from left to right. First you can see my voice, then my son’s whisper, then my whisper, then the clicking or “precursor” event, finally followed by the EVP/AVP. My voice is pretty easy to pick out the three elements of a vocal event. (Columnar structure, fine structure, and formants) My whisper you can observe two of the elements clearly. (Columnar structure and formants) The EVP/AVP you can observe two of the elements clearly as well. (Columnar structure and formants).
Using established protocols such as video recording the session, not allowing media electronics to remain on, and ensuring the sites isolation. I have been able to eliminate outside speaking contamination. I did submit this recording to numerous peers for their opinion on the voices’s content and clarity. These netted back 100% positive results for it sounding like a voice, and being easily found on the recording. The meaning of the specific words came back around 60%, for it being a specific two word combination; not enough to call it a Class A recording, but a strong B Class.
So when it all comes down to doing my job… I know, that I have captured an AVP/EVP. That’s when the WOW FACTOR sunk in…