Manual for
EVPmaker
Program for the generation of acoustic "raw material"
for recordings of "Electronic Voice Phenomena" (EVP)
by means of random controlled phoneme synthesis.
© by Stefan Bion
Freeware
|
Manual for EVPmaker
Program for the generation of acoustic "raw material"
© by Stefan Bion Freeware |
"Electronic Voice Phenomena" (EVP) are recordings of utterances upon sound recording media (recorder tapes, cassettes, video tapes, etc.) which are physically unexplainable. These voices are often reasonable messages replying to corresponding questions, or they are referring to circumstances concerning the experimenter. The voices sometimes seem to show even precognitive or telepathically transmitted contents, i.e. the voices are referring to future events or to things concerning another person of which the experimenter has no knowledge at the time of the recording.
This phenomenon has been discovered in 1959 by the swedish artist Friedrich Jürgenson, who made them public as "voices from the hereafter". In fact, these voices often call themselves the "dead" or the "deceased". Whether or not this is the only valid explanation remains to be seen...
To carry out such experiments, no high-tech equipment is required - standard equipment and a little bit of patience are sufficient to reproduce this fascinating phenomenon. In the simplest case, this can for example be done by putting a cassette recorder into "Recording" mode and capturing some background noise (= acoustic "raw material") via the microphone. From time to time you may ask a question. The voices can be heard afterwards during the playback of the tape.
More information about this topic can be found on the Web site of VTF, the German Association For Transcommunication Research or AA-EVP, the American Association of Electronic Voice Phenomena.
In order to record EVP it seems to be necessary to provide some kind of "raw material" which is used as a "carier" for the voices. Inside of soundproof rooms without any acoustic source, no voices seem to be formed at all. The frequency spectrum of this raw material - no matter if it's a direct sound event or if it's modulated onto various carriers (radio waves, light etc.) - apparently has to be within the audible range. One of the most common views is, that the voices result from paranormal transformations of the raw material. However, up to now this assumption could not yet be confirmed clearly in experiments. Usually the voices are already contained within the raw material. In my article Hypotheses about the origin of the Electronic Voice Phenomenon I have made some reflections on this topic.
There are various methods existing to generate, transmit and record the raw material. Many experimenters use a radio set which is tuned to a foreign-language radio station, or to a mixture of several radio stations, as a raw material source. It is important that the experimenter does not understand the language of the raw material because this would be distracting. I personally don't prefer this method because it's too error-prone since it can be very delusive if you are interpreting something as a message which in fact is a usual word in the foreign language. A raw material that is "unsuspicious" but is as similar as possible to language had to be found!
Already in 1988, when most people knew computers only from science fiction movies, the former president of the German Association For Transcommunication Research (VTF e.V.), Fidelio Köberle, had the following idea:
|
"[...] to provide an artificially generated raw material, like for instance the often used noise from rippling water or Struck's "rubbing method". Ideal could be a continuously produced synthetic raw material which comes as close to speech as possible. As close as possible in order to allow the interlocutors on the other side to form real speech from it using as little energy as possible. Little energy, because we know that this works best (see transformations). The raw material must of course not already be speech but should be transformed into reasonable speech easily. It should bubble without periodicity. It should, like usual speech, contain pauses. Without the use of random generators this won't be managed. [...]". [1] |
Inspired by this suggestion, different means to generate such a raw material had been developed. Electronics engineer Peter Stein (Denmark) for example used two stereo cassette players (walkmen) and switched continuously between their four audio tracks by means of an electronic switch. The switching speed could be varied. However, for me personally this method wasn't flexible enough and the generated raw material wasn't "dynamic" enough. Inspired by the psychokinesis experiments of the parapsychologist Helmut Schmidt, I wanted to use a real random number generator (RNG) for the focus of potential paranormal influences. Each random number should be assigned to a certain phoneme of the German language. While the RNG was running, the corresponding phonemes should be played back via a speaker. With a little "practice" the originators of the voices should be able to synthesize their messages by influencing the RNG mentally. Since it probably would have been rather expensive to realize this by electronic means, I decided to make use of a computer.
In those days (1989), PCs were still expensive, and sound cards weren't available anyway. The Amiga already had fantastic sound capabilities, but at that time I owned the C64 home computer, so I gave it a try. It is true that the C64 had a sound chip which could produce sounds artificially, but you couldn't use it to synthesize speech. Then, by chance, I found in an electronics shop a kit for a so called "Audio Interface" which could be connected to the User Port of the C64 in order to digitize audio signals, store them into the computer memory and play them back later. The Audio Interface used "Delta Modulation", an 1-bit A/D and D/A conversion. It was very unprecise, and sounded totally noisy, but occupied very little computer memory. Nevertheless the only 48 kilobytes of memory were already full after 13 seconds of recording time. But this was sufficient for my purposes; after all it was possible to store about 100 phonemes (if you base on an average phoneme duration of 0.1 to 0.2 seconds), and that was absolutely sufficient to synthesize artificial speech.
To operate the Audio Interface, a little BASIC program was enclosed (of course as a listing for typing in) which POKEd some Assembler routines from DATA lines into the main memory. I used a "Machine Language Monitor" (something similar to DEBUG under MS-DOS) to disassemble the program and to add some further functions to edit, save and load the audio signal. I called the resulting program simply Audio-Ed. The part "Ed" indicates, that this program was able to edit an audio signal: You could cut out short sound sequences by ear and store them as individual segments onto "Floppy-Disk". If you had collected enough phonemes, you could make a phoneme file from them. These phonemes could be used to generate "random speech" which sounded like "gibberish". Additionally, there was a function to control the choice of the phonemes via the Game Port, so that an external random generator could be connected to the C64.
But regarding the quantity and quality of the obtained EVP, it turned out in practice that it is irrelevant if "true" phonemes are used, or if simply small pieces (segments) of equal duration are cut out automatically from the audio signal. Also if "true" random numbers or only "pseudo" random numbers are used to select the phonemes/segments turned out to be insignificant. In this connection, here a little anecdote:
|
Once during an EVP session using female speech as a raw material, I received the quite distinct and striking voice "Computer ist kaputt!" (German for: "Computer is kaput!"). I didn't know what to do with it, because apparently my device worked very well. On the next day I used the same raw material again for an EVP recording (I had stored the sample on a diskette). I was very surprised when I suddenly heard exactly the same voice "Computer ist kaputt" again! What has happened? Since the C64 had no built-in real time clock, the RNG had been initialized with the same starting values each time it was switched on - until then I hadn't been aware of this fact. But how could this be? Such a distinct voice from pseudo random numbers? Moreover, through this incidence the voice obtained a certain meaning by giving me to understand in a coded form (which is often the case with EVP) that the RNG did not work the way I assumed. Since the literature always mentioned a "random" number generator, I wrongly assumed that it would produce "true" random numbers. Therefore my hypothesis was that this RNG could be influenced paranormally. But obviously this did not happen, since otherwise this voice would not have been produced once more. |
The conclusions from this incident were far-reaching: Apparently it was not necessary for the generation of distinct and meaningful EVP to influence the devices in some paranormal way! (Meanwhile my theory is that the EVP phenomenon as well as other paranormal phenomena is based on some kind of "synchronicity".)
In 1990, Audio-Ed had been rewritten for a PC (80286 CPU) under MS-DOS and extended by a "logging" feature. With it, the generated sequence of random numbers could be logged into a logfile during the (pseudo) random controlled playback of the segments. Later exactly the same "raw material" could be played back again in order to "repeat" an EVP session. By this means it was possible to examine if the same voices were obtained again. (In my own experiments, this was always the case - thus no "transformations"!)
Since editing each individual phoneme and compiling a complete phoneme set was always a very time consuming task, and because randomly cut segments work as well as phonemes, the next program version got no more editing function at all, thus it was given the name "Audigit". This version - now written in C - worked with 8-bit D/A and A/D conversion. Sound cards were still nearly unaffordable, therefore I used a self-made "Sound Digitizer" which I built from a construction manual which was printed in the October 1990 issue of the German computer magazine "DOS International" (now "PC Magazin"). Sound card support was added later. Since Audigit was a DOS program, the usable memory was still limited to 640 KB which allowed only a short recording time (about 25 seconds at a sample frequency of 20 kHz and a resolution of 8 bit). Moreover it only worked with 100% Soundblaster compatible sound cards. Therefore in May/June 2000 I wrote a "successor", the program EVPmaker, which runs under Windows 9x and Windows NT, works with any sound card and has no memory limitations.
Compared to Audigit, EVPmaker has several new features. If a sound editor such as Audition (formerly CoolEdit) is used to determine the phonemes contained in the source audio file, then "real" phonemes can be used for the generation of the EVP raw material instead of arbitrarily cut chunks. And if the phonemes are labeled with their phonetic spelling symbols, the phonetic spelling of the EVP can also be displayed. So if you hear an EVP, you can additionally read it in phonetic transcription.
Another new feature of EVPmaker is the possibility to work with "EVP sessions". An EVP session can consist of any number of individual EVP. Each EVP is automatically provided with the current date and time. For each EVP a question and an interpretation can be entered. EVP sessions can be saved to disk and loaded again later. The raw material sequence of each EVP can be repeated as often as you wish, it can be saved as a WAV file, or it can be loaded into an sound editor where it can be processed or examined in any way. Every single raw material segment, from which the randomly generated sequence has been composed of, is stored in a "cue list". If the sound editor is able to display such a cue list, then these segments can be addressed and played back immediately. In this way you can examine exactly, from which fragments a voice has been composed of.
Furthermore there are differend means for the generation of random numbers which are used to pick the segments/phonemes from the source audio file: In addition to pseudo random numbers, also "true" random numbers can now be used by connecting e.g. a radio which is tuned to "white noise" to the line input of the sound card. However, this feature is so new that I can't say which impact it has on the number and the quality of the EVP. Although I wasn't able to observe direct transformations of the raw material (as stated above), I hope for more meaningful voices since I expect that analog devices which are using "true" random processes can better "synchronize" into paranormal processes than a pure logical machine like a computer.
You can imagine yourself EVPmaker as a meat grinder - or better, as a "sound grinder": Like a meat grinder grinds meat, EVPmaker "grinds" sound. It takes a source audio file (WAV audio format), "chops" it into small segments, and then plays the chopped sound back via the speakers connected to the sound card of your PC. Like a meat grinder has different grinding plates to vary the fineness of the ground meat, EVPmaker allows you to vary the segment size between 0.001 seconds and 10 seconds. If you're lucky and listen carefully to this "ground sound", you'll hear hidden personal messages within this "gibberish"...
Actually, there exist other programs called "sound granulators", like for instance GranuLab, which is used by musicians to produce interesting sound effects. Compared to this, EVPmaker is only a very simple sound granulator. But EVPmaker has some special features especially for EVP research. One of these features is the ability to log the sequence of the segments from the source WAV file, together with some other information (date, time, sequence number, question, heard answer), into an "EVP session" file. Additionally, the information in this session file can be used by EVPmaker to reproduce an previously generated EVP at any time, or to export an EVP to a new WAV file (output WAV file). An EVP session can contain several individual EVPs.
As I mentioned, the EVP session feature is optional - you can use it, but you don't have to. If you don't use it, you must record the chopped sound by any other method - for instance by a usual cassette recorder, or by the computer itself (while playing back simultaneously).
The current release of EVPmaker can be downloaded from the following WWW address:
To start the installation, please double-click the .exe file after the download has succeeded. The queries of the installation dialog may be simply confirmed by clicking on the "Next" buttons. Of course, the default values for the "Installation path" and the "Start menu group" can be changed as needed. Finally a click on "Install" will start the installation...
Newer versions can simply be installed over an old one. A previous deinstallation of the old version is not necessary.
If you should want to uninstalled the program later, double-click the "EVPmaker" entry in the Windows standard software install/uninstall dialog ("Start > Settings > Control panel > Software").
Inside the application folder an empty folder with the name "user" will be created which is intended for the source WAV files for the generation of the raw material, the session files, and the WAV files from exported EVP sequences.
To enable access to the computer's parallel printer port, a device driver (hwinterface32.sys) will be copied into the Windows system32\drivers directory upon the first start of EVPmaker. This requires administrator rights, so if you're usually working as a standard (limited) user, you need to run the application as administrator once which should install the driver.
EVPmaker has not been written in commercial interests, but for fun of programming and for scientific curiosity. Therefore, I don't want any money or other material returns for EVPmaker; thus the software is distributed as "Freeware". Nevertheless, I reserve all rights for this program, EVPmaker is Copyright © by Stefan Bion.
The idea on which the program is based - the generation of raw material for EVP recordings by random controlled assembling and playback of individual segments of an audio signal - is common knowledge and as such not subject to copyright.
EVPmaker is supplied on an as-is basis. The author offers no warranty of its fitness for any purpose whatsoever, and accepts no liability whatsoever for any loss or damage incurred by its use. EVPmaker is not a supported product. The author accepts no commitment or liability to address any problems that may be encountered in using it; however, EVPmaker is continually being developed and improved, so he is always interested to hear about any bugs or deficiencies.
The program may be distributed or made accessible to the public in its complete and unchanged form, along with the manual. Concretely this means that only the unchanged installation file evpminst.exe may be passed on.
Your feedback regarding bugs or results in your work with EVPmaker is greatly appreciated; my current e-mail addres can be found here.
EVPmaker is continuously being developed; always the latest version is available for downloaded. Major changes and improvements are marked by an increased version number. For minor bugfixes or for "cosmetic" corrections only an internal version number will be increased, which is displayed under the "Version" tab in the "Properties" window (available through right-click on the program icon).
Here follows a survey of updates, sorted descending by the "major" version number:
New Features:
Bugfixes:
New Features:
New Features:
Changes:
Bugfixes:
New Features:
Changes at the user interface:
Changes at the random-controlled playback:
Bugfixes:
Finally also 3 bug fixes:
After this survey I will now describe the functions of the program. But first of all, I'd like to invite newcomers to open the introducing Step-By-Step Guide via Help > Quickstart Guide in the menu and to replicate the given practical examples.
After EVPmaker has been installed, it can be started either by double-clicking the "EVPmaker" icon on the desktop or by choosing "EVPmaker" from the Start menu.
After the start, the main window of the program appears:
First of all, here follows an overview of the individual areas of the program window:
At the top there is the menu bar (File Session Extras Help). All functions of the program except of those in the menus Extras and Help can be reached by buttons as well as via the menu. In the menu, there are shortcuts displayed for most of the functions. These shortcuts usually consist of letters which must be pressed together with the Ctrl key in order to activate the corresponding functions.
Below the menu bar there are the buttos
(Open),
(Save),
(Record),
(Play),
(Make EVP),
(Pause) and
(Stop).
These buttons serve to load a source WAV file, save it, record sound, play it back, start the random controlled playback,
pause and continue the playback, and to terminate the playback. How this works exactly will be explained in detail
in section 3.2. ("Generating raw material").
Depending on which functions of the program are available at the moment and which aren't, the buttons as well as other elements will appear either activated (colored) or deactivated (grey).
For the buttons there exist shortcuts as well, so that every function can be executed without using the mouse. For the upper buttons these are the function keys F2 to F8, and for the remaining buttons these are the underlined letter which has to be pressed simultaneously with the Alt key. The shortcut letters of the menu functions (Ctrl-...) correspond with the shortcut letters of the buttons (Alt-...).
Below the buttons the two areas Segments and Play EVP are following. Under Segments, the
length of the individual segments which are played back randomly can be entered. Furthermore you can select here,
whether EVPmaker itself shall determine the segments for the random controlled playback, or whether the sections
defined in the Cue List shall be used instead (Use Cue List). The first option is very easy to use, because
any WAV file is suitable. For the second option, first of all a cue list has to be created and stored into the WAV
file using a sound editor. In the area Play EVP you can choose whether the random controlled playback shall
happen continuously until the (Stop) button
is clicked, or whether it shall stop automatically after a certain number of seconds.
Below the two function groups Segments and Play EVP there is the area EVP Session which takes up the largest area of the program window. Here a new session can be started, sessions can be stored in session files, or previously recorded sessions can be loaded. For each EVP a question and the heard interpretation of this EVP can be entered. For each session, as many EVP as you like can be recorded. While skipping forward and backward between the individual EVP, the corresponding questions and interpretations are being displayed. If the cue list of the source WAV file has been used to generate the raw material for the EVP, additionally the cue labels (phonetic transcription) will be displayed in this area. Later in this manual an example will be given which shows how to work with "EVP sessions".
At the bottom of the program window the status line is located which is subdivided into three parts. Here some parameters of the loaded source WAV file are being displayed: The sample frequency (e.g. 44100 Hz), the sample resolution (8 or 16 bit), the channels (stereo or mono), the number of cues, and the length of the recording in hours, minutes and seconds.
This section will describe how EVPmaker is used to generate acoustic raw material for EVP recordings. If you just want to make some EVP recordings, this is all you have to know. First of all, a source WAV file has to be loaded (option 1) or recorded (option 2). If possible, it should be a recording of speech. For example, you could record something from the radio, or you could speak into the microphone.
To record a WAV file, the radio must be connected via a cable to the "Line in" input, resp. the microphone
to the "Mic" input of the sound card. In the Record settings of the Windows mixer
the corresponding input ("Line In" or "Mic") must be activated, and the slider should be in the middle position.
After clicking on (Record), a window opens to let you
select the recording parameters:
For the sample values, 11025 or 22050 Hz, 16 bit resolution and one channel (mono) should be sufficient for this purpose. If the option Normalize to achieve optimal recording level is activated, the level of the recorded signal will automatically be adjusted to the greatest amount of amplification that will not result in clipping. A click on the Start button will start the recording immediately. While the recording is running, the time display in the status bar will be incremented. In addition, the recording level is displayed. To achieve an optimal recording, the level should be as high as possible, but not too high. Adjust the volume control of the radio or the slider of the recording control of the mixer an a way that the green bar reaches as far to the top as possible, but without flashing the red indicator above it - in this case the recording would be overdriven and distorted (what, however, might be intended...). A level between 90 and 100 percent would be optimal - this is the case if the yellow indicator above the green bar flashes as often as possible, but the red one as little as possible.
Left: level to low - Middle: optimal level - Right: level too high
By setting a check mark at Boost by, a weak audio signal can be digitally amplified during the recording already. To do so, slowly move up the slider until the recording level is as high as possible, but not overdriven (as described above). Since the amplification works only digitally, it will also amplify possible digitizing artefacts. Therefore, this option should only be used if a good recording level can't be achieved by any other means, i.e. by adjusting the Windows mixer or the volume of the audio source.
After you have recorded for some minutes or seconds, click on
(Stop) to end the recording.
To save the recording for later use, click on
(Save), browse to the destination folder and enter a filename for the WAV file:
Of course, saving the file isn't necessary for the generation of raw material; if you prefer to generate "fresh" basic material for each EVP recording, you don't need to save the recording.
To open an existing WAV file, click on
(Open), and an "Open" dialog window
will appear, which lets you select the file.
Alternatively, a WAV file can also be opened simply by dragging the file with the mouse button pressed from the Windows Explorer into the EVPmaker window and drop it there ("Drag and Drop" method). By the way, the same works also for the EVP files which will be discussed later.
The name of the loaded WAV file appears now in the title line of the program, and in the status line further information is given (sample frequency: 9395 Hz, resolution: 8 Bit, channels: mono, Cues: 0, Duration of the sample: 51.92 seconds).
By clicking on the button
(Play)
you can listen to the whole WAV file. The playback can be paused by clicking on
(Pause).
By another click on
the playback is continued. The playback either runs until the end and will then be stopped automatically,
or it can be stopped prematurely by clicking on the
(Stop) button.
The random controlled playback can be started by clicking on the
(Make EVP) button. Here the buttons
(Pause) and
(Stop)
are effective as well. While the random controlled playback is running, an EVP recording can be
performed as usual. It turned out to be practical to send out the computer generated random speech signal into the
room via speakers and to receive it by a microphone which is connected to a cassette recorder. The questions of the
experimenter are being recorded along with the acoustic raw material. As a microphone, an amplifier microphone like for
instance the "Hannover-Mikrofon" which is offered by the VTF
or which can be built by yourself is suitable excellently.
It's also possible to record directly into the computer, so that the use of a cassette recorder is unnecessary. Required is a sound card which supports "full duplex", i.e. it must be capable of recording and playing back simultaneously. Most sound cards today meet this requirement. The microphone must be connected to the "Mic" input of the sound card instead of the cassette recorder. While the raw material is played back via the speakers, a sound recorder like for instance CoolEdit is turned to "recording" mode at the same time. Some sound cards require the same values for sample frequency, resolution and channels: If for example the source WAV file which is used for EVPmaker has the values "22050 Hz, 16 bit, mono", then exactly these values must be used for the microphone recording in CoolEdit. Afterwards, the EVP recording can be listened to at the computer. Some sound editors offer the valuable feature to mark selected segments within the wave form by written notes. In CoolEdit, this can be done with the "Cue List" which can be accessed from the "View" menu. In this way, EVP interpretations can be entered directly, and the voices can be accessed later very quickly.
But now back to EVPmaker: By changing the settings within the framed areas Segments and Play EVP, the random controlled playback of the segments can be varied further. First the segments: Here are two edit fields for the duration of the segments in milliseconds (ms), i.e. 1/1000 seconds. By default, only the upper edit field labelled with From is active, while the other one, which is labelled with To, is greyed (= inactive). If the box on the left side of the lower edit field is checked, this field becomes active, too. The values in these fields control how much the signal is being chopped during the random controlled playback. If only the upper "From" value is active, all segments are equally long. If also the lower "To" value is activated, then the segment length varies random controlled between these two values, which results in a more dynamic raw material. The segment duration can be increased or decreased by clicking on the little up/down arrows besides the input field. The valid range covers values from 1 ms to 10000 ms, which is equivalent to one thousandth to 10 seconds. Values around 100 milliseconds (= 0.1 seconds) have turned out to be convenient. If the values are too big, parts of words or even whole words may become audible. If the values are too small, the signal is being distorted very much, which however can lead to interesting results.
Right beside the input field for the segment duration there are the three checkboxes Overlap, Z-Cross ("Zero Cross") and X-Fade ("Cross Fade"), which will be explained now.
Overlap has the following meaning: Normally all segments are located seamlessly one after another at exactly defined positions. With a recording of 10 seconds and 100 ms segments, there are for example exactly 100 different possible starting positions for the randomly played segments. If Overlap is checked, any segment can start at any position within the audio signal, which means that segments can overlap each other. Overlapping segments result in a higher variation of the generated raw material.
The following chart is supposed to illustrate the above mentioned for more clarification:
The above chart shows 5 examples of settings for segment lengths and the arrangement of segments. The thick vertical lines indicate the beginning of segments, the thin vertical lines their ends, and the red arrows indicate the play direction and the length of the segments. The segments themselves are displayed in grey, and areas where two segments overlap are indicated by a dark grey color.
At A, the "From" and "To" segment lengths are equal, or there is no check mark before To. The Overlap option is deactivated. All segments are of equal length and are located seamlessly one after another without overlapping each other. The segments start at positions which are a multiple of the "From" length.
At B, the "To" segment length is less than the "From" segment length. Thus, the segment length randomly varies between these two values. Since Overlap is switched off, the segments again start at positions which are a multiple of the "From" length.
Also at C, the Overlap option is switched off, and the segments again start at positions which are a multiple of the "From" length. However, this time the "To" segment length is greater than the "From" segment length, which - in spite of the deactivated Overlap option - may result in operlapping, which is the case when the end of a segment lies behind the beginning of the following segment.
At D and E, the Overlap option is now switched on. This causes that now the segments don't necessarily start at positions which are a multiple of the "From" length, but they can be distributed across the whole recording at arbitrary positions. Through this, overlappings are possible for both constant segment lengths (D) and variable segment lengths (E).
If Z-Cross is checked, segments will be cut out from the audio signal only at positions where the wave form crosses the center line (zero amplitude). By this, audible pops or clicks during the random controlled playback of the segments are being eliminated. However, a side effect of this function is that the segment duration no more matches the preset value, but usually is a bit greater (for example 102 ms instead of 100 ms - it depends on how far the next zero crossing point is apart). In very unfavorable cases, it might even occur that the next zero crossing point will be found much later, or that it won't be found at all, so that single segments may last several seconds. This can for example occur when the signal has an DC Bias, i.e. when the zero amplitude of the signal has an offset towards the center line of the wave window, so that the wave form crosses the "actual" zero line only now and then. In such a case it might help to load the WAV file into a WAV editor and to "adjust" the signal (in CoolEdit, choose "Transform > Amplitude > Normalize", check the option "DC Bias Adjust", and normalize the signal to 90%).
With activated X-Fade option, temporally consecutive segments will be gradually cross faded during playback. You can choose between different types of transitions between the segments: Linear, Sine and Trapezoid. The following (bilingual) chart illustrates this:
At A, no cross-fading takes place - the segments follow each other immediately. At B-G, the transitions are more or less gradually: While the volume of the new segment (Segment 2) increases, the volume of the previous segment (Segment 1) decreases simultaneously. After reaching the maximum level, the current segment (Segment 2) slowly fades out while at the same time the next segment (Segment 3) slowly fades in, etc. This fading results in a "smoother" sound since the cuts between the segments are no longer audible. However, with the same cutting frequency the length of each segment will increase.
At B-D, the transitions are "trapezoid-shaped" (called this way because of the trapezoid-shaped volume envelope: At first the volume of a segment increases, then it remains constant for some time, and finally it decreases again. The three shown examples differ only in the steepness of the slopes, what results in a diffrent degree of temporal overlap between adjacent segments which also can be specified in percent.
At E, the transitions are completely steady or linear. Basically these are trapezoid-shaped transitions as well, but with an overlap of 100 percent.
At F, the volume envelope is sine-shaped, which leads to a smoother variation of the volume of each segment throughout the entire range, i.e. there are no abrupt changes as with the trapezoid-shaped envelopes.
At G, the volume envelope has the shape of an sine half-wave. This leads to higher sound levels in the
overlapping areas, which for some signals (e.g. recordings with a lot of noise) leads to an even smoother hearing
impression. The reason for this is that the effective amplitude of two non-correlated signals U1 and U2 don't
simply double when they are added up, but it increases only by a factor of 
.
Which of these fading types are most suitable should simply be tried out.
The different transition types can be selected in the context menu which opens upon right-clicking the X-Fade checkbox:
As mentioned, besides the option to use segments of fixed duration, the program can also be instructed to use the "cues" which have been stored in the WAV file to generate random speech. To achieve this, the option Use Cue list inside the area Segments must be selected. This Option is available only if the loaded WAV file does actually contain cue markers. The just mentioned options Overlap and X-Fade take no effect if the cue list is used. But using cues instead of fixed-length segments will be explained in detail later.
For EVP recordings, it is sometimes quite practical if after each asked question the raw material is being
played back only for a certain time and then stops automatically, so that the next question can be asked. For
this purpose, EVPmaker offers the option to preset a duration in seconds, after which the random controlled
playback stops automatically. Normally, the playback runs until the
(Stop) button is clicked. This is the
case if in the area Play EVP the option continuously is selected. To activate the automatic stop
of the playback, for x s must be selected. In the corresponding input field of this option, a
number of seconds can be entered, after which the playback of the raw material is being stopped. By clicking
in the little up/down arrows, this value can be varied from 1 to 999 seconds.
This is all you have to know in order to generate raw material for EVP recordings. Any further description is only interesting if you intend to use EVPmaker's EVP Session feature in order to document and examine the obtained EVP scientifically.
The EVP Session feature is primarily meant to take down the generated raw material into a log file in order to be able to examine and to analyze it later. Thus it serves mainly scientific purposes and is less intended for all-day use, particularly since experience shows that most EVPs are audible only in the acoustic recording of the raw material and not already in the original raw material itself. By the unadulterated recording of the original raw material with the help of the function EVP Session one has, however, the possibility of comparing both with each other and that way documenting possible "transformations".
In the session file (EVP file) not the result in form of a new WAV file is saved, but merely all parameters which have led to the generation of this raw material are recorded. This includes the name and the location (folder) of the used source WAV file as well as for each segment its position and length within the source WAV file. With the help of these details EVPmaker then is able to reproduce the raw material exactly. However, an "export" function allows also to store the acoustic final result as WAV file in addition (more on this later in this manual).
In order to record a session, the recording mode must be activated by clicking the button Rec in the area EVP Session. While the recording mode is active, the label Recording appears in red letters on the right side of the upper border of the frame of this area:
The recording mode can only be activated if either a WAV file has been opened, or a previously made recording has been saved as a WAV file.
If a session file is already open, new EVP will be appended to the session.
If during activated recording mode the random controlled playback is being started via the button
(Make EVP),
at first a dialog box will be opened which allows you to enter an question:
The question or the comment which may be entered here has only documentation purposes and doesn't have any effect on the generation of the raw material. However, with regard to the documentation of possible references of the EVP to a previously asked question it would be important to note the (mental or verbal) asked question.
After clicking OK, the dialog box is closed, the entered text is stored into the session, and the playback
of the raw material is started. Cancel will cancel this function. In recording mode, the numbers of the randomly
played back segments resp. of the cues are being logged in the session file. After playback has been stopped (either by
clicking the upper (Stop) button, or automatically after a certain
number of seconds), the next EVP can be recorded immediately by clicking on
(Make EVP) once more and entering a new question.
At any time you can listen to any EVP which has been recorded so far, or to the entire session, by clicking on the
lower Play resp. EVP
button (explained below). To finish a session, the recording mode has to be terminated by clicking on Rec
again. The red label Recording will then disappear. The recording can be continued
at any time later, but it is recommended to create a new session file for each EVP session. To start a new session
while a previous session is still opened, click on the New button. If the previous session had not yet been
saved, you will be asked to do so. The current session can be saved at any time by clicking on Save (explained
below).
Every single EVP that has been recorded in the session file is automatically being provided with the current date and time (Date/Time) and a consecutive number (EVP No.). The number before the slash is the number of the currently displayed EVP (sequence number), and the number after the slash is the total number of recorded EVP in this session. You can skip between the individual EVP by clicking on the buttons labelled - and +. Instead of clicking these buttons, you can alternatively use - and + on your keyboard. The previously entered question now appears in the field labelled Question:
By clicking on the button Play, the entire session including all EVP will be played back completely, whereas the button EVP playes back only the currently displayed EVP. Stop terminates the playback prematurely. You can enter your interpretation of an EVP into the field EVP Answer. This also is important with regard to a documantation of references of the EVP to a previously asked question.
After a session has been recorded, it should be saved into a session file. Any saved session can later be loaded into EVPmaker again in order to process it (for example to edit the interpretation text) or to listen to it. To save a session, click on the button Save. A dialog window will be opened to save the session file:
The suggested filename is a combination of date and time of the first EVP which has been recorded in this session file. In the above example this is the session of 12 June 2000, 19:43:10 h. Of course the filename can be changed if it appears too "illegible" to you. The file extension is normally ".evp".
Saved sessions can be opened by a click on Open or by using the "Drag and Drop" method. At the same time also the corresponding source WAV file which has been used to generate the raw material will be loaded automatically. Thus the used WAV file should be preserved well; without it, the session cannot be opened later!
If the source WAV file can't be found anymore when opening a session file because it has been renamed or moved to another folder, EVPmaker complains and asks if you would like to search for the matching file:
If you answer Yes, an "Open" dialog will appear which lets you browse to and open the missing WAV file.
After a session has been saved or an existing session has been opened, its filename will be displayed in the field labelled Session file.
Perhaps you would like to save individual sequences of EVP raw material in order to archive them or to send them by e-mail. For this purpose, EVPmaker offers the export function which allows you to export the currently displayed EVP as a WAV file. After clicking on the button Exp, an dialog window is opened to save the file:
The suggested filename is composed of the session file name, followed by a dash (-), followed by a three-digit number. This number is the consecutive number of the current EVP which is displayed behind EVP No.
To examine an EVP more exactly, it's recomended to open it in a suitable sound editor. My favorite one is Audition (formerly CoolEdit) from Adobe. By clicking the Edit button, the currently displayed EVP will be temporarily saved as a WAV file and loaded into the editor. By default, this is the program which is predefined in Windows as the default application for WAV files, i.e. which is opened automatically when double-clicking WAV files. If this is not your favored application for analyzing EVP, you can configure an alternative editor in EVPmaker. This is done by the context menu which appears upon right-clicking the Edit button:
After choosing the Select editor item from the context menu, an open dialog appears which lets you browse for the executable file (EXE file) of the desired editor. The executable files normally reside in a subfolder beneath the "Program files" folder (e.g. C:\Program files). By selecting the context menu item Use Windows default, the Windows default is restored.
By a normal left-click on the Edit button, the currently displayed EVP is loaded into the predefined editor. This might look like this (here still shown at the example of the old "Cool Edit 96" version):
The blue triangles above and below the wave window are the markers of the individual segments of which EVPmaker has composed the EVP sequence. These markers have been stored as "cues" in the WAV file and are now being displayed by CoolEdit. To show the cue list, choose "View > Cue List" from CoolEdit's menu. The list of the individual segments might for example look as follows:
The individual segments are being displayed one below the other. By using the scrollbar on the right margin, you can scroll through all of the segments. Double-clicking a segment selects that segment in the waveform. In the first column of the cue list you see the exact positions of the segments within the loaded WAV file in seconds. The next column shows the labels of the segments. All segments are numbered consecutively, starting with 1. The label is also being displayed in the field "Label", if you click one of the cue list entries. Right beside the label there is a "Description" field which displays for each segment the name of the original WAV file (that one which EVPmaker used to generate the EVP sequence), the number of samples, and the first and the last sample within the original WAV file.
An EVP which was previously discovered in EVPmaker can now be examined much more precisely. For example you can zoom into the signal or switch to spectral view in order to analyze the frequency components of the signal:
BTW: Not only those EVP which are loaded in the WAV editor are provided with a cue list, but also those EVP which have been exported as WAV files.
As mentioned above, EVPmaker can not only use arbitrarily choosed segments for the random controlled playback, but also individually selected parts which have been stored in the cue list of the source WAV file. By means of the cue list, a set of phonemes can be created which can be used to synthesize random speech. Phonemes are the smallest sound elements speech is being composed of; this comprises all consonants and vowels. To learn which phonemes are existing, look into a dictionary where the pronunciation of the language is explained. Now record a suitable text - either spoken by yourself, or an existing one -, select every single phoneme and add it to the cue list. For a label, you should use a short name. Since you can't use phonetic transcription symbols in the cue list, you have to manage with a kind of "pseudo phonetic transcription". In the Appendix I have listed the phonemes which I have used, along with their names and some pronunciation examples. [2]
Next, zoom into the signal and select the phonemes by ear and by sight:
Before you add a phoneme to the cue list, choose "Edit > Zero Crossings > Adjust Selection Inward" from the CoolEdit menu to align the selection to the nearest places where the waveform crosses the center line in order to eliminate clicks and pops. Then click on the "Add" button of the cue list window and enter a name for this phoneme into the field named "Label":
Optionally you can enter a description. By the way, the labels can be used multiple, i.e. they needn't be unique, so phonemes can be stored in different variations (durations, pitches etc.) using the same name for it. To simulate a natural flow of speech, you could additionally include a few pauses.
To save memory, you could open a second CoolEdit window and copy each phoneme you have selected in the first window into this second window so that all phonemes follow one after the other without gaps.
When saving the phoneme set to disk, pay attention that the option "Save extra non-audio information" in the Save dialog is checked! If you forget this by mistake and overwrite an already existing phoneme file and then close CoolEdit, your complete cue list will from now on reside in Nirvana... ;-)
The phoneme file can now be opened in EVPmaker. Below in the status line, after Cues, the number of the loaded phonemes are now being displayed. In order to use these phonemes instead of arbitrarily segments for the generation of the raw material, the option Use Cue list must be selected:
The further procedure is the same as described in section 3.3. ("Recording a session").
In addition to the heard interpretation of an EVP, you can now read the phonetic transcription which is displayed in the field Cue Labels. To change the font which is used to display the phonetic transcription, right-click into this field and choose another font from the Font dialog. Individual sections or the complete line of phonetic transcription can be copied to the clipboard and pasted for instance to a text editor. Use the mouse to select a portion or the entire line and press the keys "Ctrl" and "C" simultaneously to copy the selected part to the clipboard. In your text processing program, press "Ctrl"+"V" to insert the text from the clipboard.
In the example which is displayed above, for example the word "beschieden" (German for: "notified") was heard. In phonetic transcription this corresponds to the sequence "b t 3 3 sch: i i: m". At first sight this sequence has no big similarity to the word "beschieden", but as everybody knows the human hearing is pretty flexible...
To examine this sequence further, the EVP is being loaded into a wave editor by clicking on Edit:
Here the section which includes the heard phonetic sequence "beschieden" has been zoomed. In the spectral view you can clearly recognize the various phonemes: From 2.13 to 2.20 the plosive "b" with a preceding pause before suddenly the air deflates between the lips, then from 2.20 to 2.25 the inappropriate "t" - also a plosive with preceding pause -, followed by two "3" vowels (called "schwa" vowels), followed by a "sch" which can be recognized in the spectral view as hissing. Next the two "i" which result in a long "i:", then the "m" which sounds like am "n" with a preceeding plosive because of its suddenly (in fact unnaturally) beginning.
In the cue list, the individual phonemes are listed in detail:
The labels of the phonemes (= the "pseudo phonetic transcription") were added in brackets and between quotes after the segment labels. In the cue description the time and the original number of the segments resp. cues from the source WAV file is specified. wird zu den einzelnen Segmenten die Zeit angegeben, sowie die Nummer des Segments bzw. der Cue aus der Ursprungs-WAV-Datei, aus denen sich die Datei mit der EVP-Sequenz zusammensetzt.
Well, this was only a little insight into the possibilities of how EVPmaker can help to objectify EVP interpretations.
Instead of a previously loaded WAV file, also the signal of an audio source (such as a radio etc.) connected to the sound card input can be "chopped" in real time. For this, the signal is being recorded continuously into a storage area of predefined length (e.g. 10 seconds), and simultaneously played back random-controlled. Since the recording into this buffer storage constantly overwrites the previous recording, the sound of the chopped material will change all the time.
The live audio mode is being activated by selecting the menu item Extras > Live audio source. If a WAV or EVP session file is loaded, it will be closed; an unsaved recording will be discarded. However, there will be a possibility to save the changed files.
Next, a window is shown for entering the sample parameters to be used, and the buffer time:
The sample rate, number of channels, and the resolution affect the quality of the recording; the buffer time specifies the recording time of the signal of the audio source in seconds. If this time is very long, then it will also take a long time until the buffer is full and the playback can start; if it's too short, there will be too many repetitions of the same segments of the audio signal. 10 seconds seems to be a good value.
After clicking OK, the recording of the signal of the audio source connected to the sound card input into the
buffer will start immediately until the buffer is full. During this process, the message "Initializing Live Audio Mode"
appears in the title line. As soon as the buffer is full after the predefined time, the caption text will change to
"Live Audio Source", and the sample parameters and the length of the audio signal in the buffer will be displayed in
the status line at the bottom. The programm is now ready for the random-controlled playback which can be started as usual
by clicking on (Make EVP).
Other playback or recording functions than "Make EVP" aren't possible during the live audio mode, and also WAV or EVP session files can't be loaded. Recording EVP sessions doesn't work, since this requires a loaded WAV file. However, the speed control will work, i.e. you can slow down or speed up the chopped signal of the live audio soure.
To terminate the live audio mode, select the menu item Extras > Live audio source again.
By using effects, the source WAV file which is used for the generation of the raw material can be defamiliarized further. In the Extras menu you can find the following entries:
This function will turn over the entire audio signal so that it will play backwards. This could be used for EVP recordings to make a recording of speech in a language you understand more unintelligible, if you want to use very long segments, in which otherwise you would hear individual words.
With this function reverberation and echo effects can be produced. By this the audio signal will get more "volume". Two parameters will influence this effect: Delay in milliseconds and Decay in percent:
Short delay times (less than 100 ms) will yield a reverberation effect, long delay times will yield echo. The duration of the reverberation or the number of echoes is determined by the decay: A small value will let the reverberation or the echo cease faster than a large value.
This function will amplify the amplitude (i.e. the recording level) of the audio signal to the greatest possible amount that will not result in clipping. By this means, recordings which are too faint can be adjusted to an optimal level subsequently.
* * *
All three effects, Reverse Signal, Add echo to signal, and Normalize signal, will change the loaded WAV signal. Therefore if a new source WAV file is opened, or if EVPmaker is exited, you will be asked if you would like to save these changes:
If the answer is Yes, the changes will be saved; if you want to keep the original WAV file, you should enter a new filename. You should do so if this WAV file has already been used for other EVP sessions!
By clicking on (Save), the WAV file
can also be saved before.
In a normal speech signal, the volume is changing continuously in the rhythm of the speech. With this function, the loaded WAV signal can be automatically provided with cutting marks at the positions of the volume minima. The idea behind this is that for the random-controlled playback this will result in a more natural speech rhythm than it would be possible with cutting marks set at arbitrary positions or at equal distances. (Perhaps this will make it easier for the assumed 'EVP partners' to form their intended messages.)
For a better understanding of what this function exactly does and how it works, I will now give a short explanation using the following illustration:
The program first analyzes the audio signal (green waveform) regarding its volume progression, also called amplitude envelope. At the positions where this envelope deceeds or exceeds an adjustable threshold value (indicated by the two horizontal red lines), markers are set (symbolized by the red dots). In the center between each two of these markers, where the envelope narrows (which means that there is a volume minimum), cutting marks (the yellow vertical lines) are placed.
After selecting this menu item, the following window appears:
At Threshold a threshold value can be specified, to which the volume envelope is compared. At Set threshold to ... percent, the threshold value can be entered directly in percents; with Determine threshold automatically selected, the program will set the threshold to a calculated mean value of the volume of the audio signal by itself. At Amplitude envelope, the Frame size parameter can be specified. This value (in milleseconds) defines the size of the time windows which are used for the determination of the volume progression; it should be chosen neither too small nor too big. If the time window is so small, then the "fringes" of the audio signal shown in the illustration above are interpreted wrongly as volume changes, and too many cutting marks will be set. If the time window is too big, then the program may "miss" fast volume changes, and too little cutting marks will be set. The preset value of 15 ms should yield suitable results in most cases in practice. However, it's unavoidable that some cutting marks will be too far apart from each other, if for instance the volume alteration of the signal is not big enough. In such a case, it's possible that a section between two cutting marks comprises several syllables or even an entire word!
After clicking the OK button, the auto-cut functon is applied. For each automatically created cutting mark, a cue will be inserted into the cue list. If there are already cues in the WAV file, they will be deleted first. Since the WAV file is changed by this, EVPmaker will ask if you want to save the file before you close it. To use the automatically created cues for the random-controlled playabck, thr Use Cue list option must be activated.
Opens a slider which lets you adjust the playback speed. It can be adjusted to values between 50% und 200%, which is bisection or doubling:
Since this is yielded by changing the sampling rate of the WAV signal, it depends on the sound card it the variation of the speed is possible or not. If the resulting sampling rate lies beyond the capabilities of the used sound card, an corresponding error message will appear upon playback. It also depends on the capabilities of the sound card whether or not a change of the speed affects the playback immediately or only after an interruption and continuation of the playback.
By the 100% button the playback speed can be reset to the original value by a single click.
This function is based on an idea someone suggested to me by e-mail. Here is an excerpt from it:
|
Assuming that communication can be obtained through the random generation of information, what if, in addition to sound you also added visual images to enhance and clarify the verbal communications? This would not only make communication clearer (assuming we are communicating with other dimensional entities), it would make verification of this communication much more palpable if the images correlated or complemented the words. Let's say you had about five hundred or so images or symbols or colors. Here's a few ideas: a ufo, a cross, a dragon, a smiley face, a pentagram, a peace sign. These images of course would be randomly chosen in the same way the sound is on your program. Going by one theory, if these other dimensional intelligences actually can "see" the screen, there might be a way to program the software in which they have a real time interaction with it to see their "work in progress" so to speak. |
This idea sounds interesting. In order to be able to check it, EVPmaker has a so-called "Visual Feedback Window". It can be activated by choosing Visual feedback window from the Extras menu:
If now the random-controlled playback is started by clicking on
(Make EVP),
a certain symbol with a certain color for each segment will appear in this window
during the generation of the raw material.
As symbols, the 223 characters of the Windows font "Wingdings" are used by default:
Combined with the 15 colors Black , Maroon , Green , Olive , Navy , Purple , Teal , Silver , Gray , Red , Lime , Yellow , Blue , Fuchsia and Aqua these are 223 x 15 = 3345 possible different symbol/color combinations. This means of course that with more than 3345 segments a clear assignment of segments to such a symbol/color combination is no more possible. This is for instance the case if a WAV file of 8 minutes is subdivided into segments of 100 milliseconds: In this case there are 480 seconds / 0,1 seconds = 4800 segments, which is more than existing symbol/color combinations. For this reason a corresponding note will be written at the bottom of the "Visual Feedback Window" if such a case occurs. Also it doesn't make sense to use overlapping segments, since because of the arbitrary starting positions there is virtually an "infinte" number of different segments. Therefore a corresponding note will appear here as well, if during random-controlled playback the Overlap option is activated.
By choosing Settings > Font in the menu of the "Visual Feedback Window", another font than the default "Wingdings" or a font size different from the default of 300 can be chosen:
There are many Web pages where symbol fonts can be downloaded. Just pay attention that actually all 223 characters are used, since otherwise invisible "blanks" will be desplayed. "Webdings" for instance is another suitable font.
A different font size is appropriate if the size of the feedback window is changed to fit to one's own screen resolution. The characters being displayed should fill the window as well as possible.
As mentioned in the Introduction, the original idea behind the means of random controlled phoneme synthesis was a supposed paranormal influence on random number generators (RNG). Experiments have shown however, that meaningful EVP also occur by using program generated pseudo random numbers which can in no way be influenced. Similar observations have been made during psychokinesis (PK) experiments in parapsychology labs: PK influences of a display by test persons occured not only by use of "true" RNGs on the basis of radioactive decomposition, but also by use of computer generated pseudo random numbers. Even by use of so called "pre-recorded targets" (PRT), i.e. series of random numbers which have been recorded some time before the actual beginning of the experiment, this PK effect did occur, which by common sense should not be...
In order to be able to conduct experiments with these different possibilities for the random controlled phoneme synthesis in EVPmaker, you can choose between several ways of random number generation:
Each random number source delivers numbers within a certain range of values:
| Random number source | Range of values | Number type |
| Pseudo random numbers | 0 ... 2,147,483,647 | Integer (32 bit) |
| Fractal equation | 0.0 ... 0.999… [1] | Floating point (20 digits) |
| Sound card | -32,768 ... 32,767 [1] | Integer (16 bit) |
| File | -32,768 ... 32,767 [1] | Integer (16 bit) |
| Internet service | 0 ... number of segments/cues (without Overlap) resp. samples (with Overlap) | Integer |
| Manual sweep | 0 ... 2,147,483,647 [2] | Integer (32 bit) |
| Parallel printer port | 0 ... 4,294,967,295 | Unsigned integer (32 bit) |
[1] Takes effect only if the Bitwise convert 32 RN to 1 RN option is deactivated. If this option is activated, each 32 of the delivered values are, according to the procedure described there, converted into unsigned 32-bit integers within the range of values of 0 ... 4,294,967,295.
[2] Which numbers within this range of values are actually generated, depends on the width of the slide control. Only as many numbers can be generated as slider positions exist. For example, if the slider has a width of 1000 pixels, 1001 diffrent numbers between the smallest and the biggest number (including them) are possible.
The range of values of the selected random number source is distributed equally among the existing segments or cues, i.e. the smallest possible random number corresponds to the first segment, and the biggest possible one to the last segment.
You can select one of these random number sources in EVPmaker in the Random source submenu which can be found in the Extras menu:
In the following section, all of these options will be explained in detail:
This is the program-internal generation of pseudo random numbers, or more exaxtly, the "random()" function from the program library of the Borland C++ Builder 4.0 compiler which I have used. According to the documentation, this is a multiplicative, congruential random-number generator with a cycle length of 2³².
With this setting, the random numbers are generated by means of a nonlinear equation which delivers nonpredictable results (deterministic chaos). At present only the the Growth Equation by Verhulst has been implemented:
f(x) = r * x * (1 - x)
This formula is for example used in biology to describe the growth of a population inside a closed system. This equation is non-linear and produces chaotic results when r is above 3.57. For more information, see for instance here.
Experiments by Prof. Dr. Johannes Hagel and Margot Tschapke [3] have shown, that the chaotic results of the iteration seem to support the EVP. Voices occur both within the noise which becomes audible when the x values are interpreted as sound samples, and when these values are used for the selection of phonemes or speech segments.
If you select Fractal equation from the EVPmaker menu, the following mask will appear:
In the upper area (above the separation line) a Feigenbaum diagram is shown. Here you can experiment with various x and r values in order to explore the chaotic areas. By clicking on the Draw button the diagram will be drawn using the current values. A left-click into the diagram sets the graphics values: To enlarge an interesting area, first click to the upper left, then to the lower right edge of this area, and then click the Draw button. With Reset, the graphics values are set to default.
In the lower part of the mask, the settings of the parameters r and x can be done, which are used for the generation of the random numbers. You can either enter these values directly into the two input fields on the left, or you can increase/decrease them by moving the sliders, or you choose a random value by clicking on the Random buttons, or you can simply right-click into the above displayed Feigenbaum diagram. With the Iterate! button, you can simulate some iterations to observe the behavior of the x parameter. If this mask had been opened during the random controlled playback (using Random source = Fractal equation), you can make these settings immediately audible by clicking the Apply button. The OK button does the same but additionally closes the mask. Cancel closes the mask without changing the r and x values for the program.
To generate 32-bit random numbers, sample values are being read continually from the sound card. Here an external noise source connected to the Line input of the sound card serves as an analog random source. In the simplest case this can for instance be a radio tuned to "white noise". In some cases even the internal noise of the sound card is sufficient - however, in this case usually the Bitwise convert 32 RN to 1 RN option must be activated, since the signal strength is to low to achieve a distribution of the generated random numbers across the entire WAV file. If EVPmaker seems to "stutter", or if only a few segments are being repeated again and again, this usually means that the signal provided by the sound card to low or not "randomly" enough.
If the sound card is being selected as the random number source, a level meter pops up to display the signal level at the selected input of the sound card during the random-controlled playback. To make sure that the played segments which are selected by the random numbers are distributed among the entire WAV file as uniform as possible, the recording level must be as high as possible - however not too high, since otherwise the signal peaks are clipped, causing the smallest and the biggest sample value to appear unproportionally often, which would lead to a preferred playback of the segments at the very beginning and the very end of the WAV file. A level between 90 and 100 percent would be optimal - this is the case if the yellow indicator above the green bar flashes as often as possible, but the red one as little as possible.
Left: level to low - Middle: optimal level - Right: level too high
By setting a check mark at Boost by, a weak audio signal can be digitally amplified. To do so, slowly move up the slider until the signal level is as high as possible, but not overdriven (as described above). Since the amplification works digitally and thus only "stretches" the sample values, but doesn't increase the signal resulution, this option should only be used if a good recording level can't be achieved by any other means, i.e. by adjusting the Windows mixer or the volume of the audio source. "Stretching" the sample values results in the coverage of a wider area within the loaded WAV file, but - from a certain point on - not in an increased number of played segments.
The setting Sound card does work only with "Full Duplex" sound cards, i.e. the sound card must be capable of recording one signal while playing back another signal simultaneously. Otherwise, an error message will appear.
Here the integer values needed for the generation of random numbers are being read from a file. This can be either a text file or a binary file. Files with an ".txt" extension are regarded as text files, everything different as binary files. For binary files, the default extension is ".pcm", but this can be changed in the Open dialog which appears upon selecting this random source, by selecting the file type "All Files (*.*)".
A binary file can in fact be any file; however, the values which are read from it will be interpreted as signed 16-bit integer numbers with a range of values of -32,768 to +32,767. This is equivalent to the PCM data format of WAV files with a resolution of 16 bit. Suitable as a random source are therefore for instance sound files of white noise, recorded with a sample resolution of 16 bit and in mono; the sample rate doesn't matter. The sound file should be saved in the "PCM Raw Data" format, which means that only the "raw" sample data without any header is written to the file. Of course, not only sound data but also any other type of data may be used, as for instance the digits of the number Pi, or the results of fractal iterations.
In the text file used as a random source, the numbers must be written one below the other in one single column, i.e. one number per text line. The numbers in the text file must be integers from -32,768 to +32,767 as well. The numbers mustn't contain any separators (commas or dots, etc.). Lines beginning with a semicolon are allowed and will be ignored; this way you can insert comments into the text file.
By selecting this random source, the required random numbers are delivered by an Internet service. The service must provide the random numbers in plain-text format, separated from each other by line breaks (i.e. one below the other) in one column. The format of the caller URL is usually described on the website of the service provider. Here's an example for the service of random.org:
http://www.random.org/integers/?num=10&min=1&max=100&col=1&base=10&format=plain&rnd=new
This call delivers 10 random numbers in the range from 1 to 100. You can test this by clicking the above URL!
To let EVPmaker tell the service how many and which random numbers it is supposed to deliver, the parameter values in the URL can be substituted by valiables:
| Variable | Parameter |
| %min% | Smallest random number |
| %max% | Biggest random number |
| %num% | Number of random numbers |
Example of an URL with the parameters replaced by EVPmaker variables:
http://www.random.org/integers/?num=%num%&min=%min%&max=%max%&col=1&base=10&format=plain&rnd=new
The service URL can be selected or entered in the following window which automatically appears when "Internet" is chosen from the menu as a random source:
The service provider www.random.org is preset. If more services are defined, they can be selected from a drop down list which is opened by clicking the "down arrow" button right besides the input field. By entering an URL into the input field you can add a new provider to the list. By clicking the Delete button, the currently shown service is removed from the list. Clicking OK selects the chosen service as a random source. The Cancel button cancels the selection.
The random numbers will be retrieved from the Internet service in blocks of 1024 numbers. If they are "used up",
a new block of 1024 random numbers will be requested. This procedure repeats until the playback is stopped. When
(Make EVP) is invoked for the first time, at first such
a block will be requested before the random controlled playback will start. During the playback, new blocks of
random numbers will be requested if required. While new random samples are received from the service, the title
of the EVPmaker window shows the text Retreiving Random Numbers via HTTP. If an error occurs, an
error message will be displayed. Errors might be caused, for example, by an interrupted Internet connection,
if the the service is unavailable, or if the quota of available random numbers is exceeded. For random.org,
the available quota can be requested here. (Tip:
To "refresh" the quota, retrieving a new IP address by reconnecting to the Internet might help.)
Since the range of values of the 1024 received random numbers are based on the segmentation parameters (From and To segment lengths, Overlap and Use Cue List options) as they were at the time of the request, new random numbers should be retrieved over the Internet each time these parameters are changed. This can be achieved by selecting the random source "Internet" again, or by activating the Retrieve new RN for each EVP option.
This random source converts the movements of a slider by the user into numbers which are then used to choose the segments resp. cues to be played back. The slider position corresponds to the relative position of the segments or cues within the loaded WAV file.
After selecting this random source from the menu, the following window will be opened:
After the randomized playback has been started by clicking
(Make EVP), you can use the mouse to drag the slider back and forth - the more irregular the better!
If the Mute when slider not moved option is checked, the playback will be muted as long as the slider
remains on one position. Otherwise, one and the same segment/cue would be repeated over and over, which might
be undesired.
Although EVPmaker works very well with software-generated pseudo random numbers, one might prefer using "true" random numbers generated by an external hardware device. If this random source is selected, EVPmaker will query an input line of a parallel printer port (LPT port) in order to retrieve random bits which are then assembled to 32-bit integer numbers.
After choosing this random source from the menu, a window will appear to let you specify the port address, the input line, the sample rate, and the unbiasing method:
If the address of your LPT port is not in the Port address drop-down list, you can enter it manually. To find it out, go to "My Computer > Manage > Device Manager", choose "Parallel Port" from the device tree, double-click it and select the "Resources" tab - the base address is the first stated number.
A standard LPT port has 5 input lines. Select the line your random bit generator (RBG) is connected to from the Input line dropdown box. The numbers before the dashes refer to one of the 25 pins of the printer port connector:
The Sample rate value specifies the number of port reads per second. To ensure that the random bits are uncorrelated, the sample rate must not be too high. I recommend value of not more than 1/10 of the average output frequency of the RBG. So if the RBG outputs 1 MHz, set the sample rate to 100 kHz (= 100000 Hz).
An ideal RBG should output the same amount of digital 0s and 1s over a longer time period, giving an average of 0.5. If this is not the case, the resulting 32-bit random numbers will be distributed unevenly, and there will be accumulations of certain numbers. If the used RBG delivers a biased signal, it can be corrected by selecting one out of two available Unbiasing methods: The XOR method reads 2 bits and outputs a 1 if both are different and a 0 if they are equal. The Von Neumann method also reads 2 bits but outputs the first one if both are different and throws them away if they are equal. Example:
| Raw output | 10 | 11 | 00 | 10 | 10 | 01 | 00 | 01 | 10 | 10 | 01 | 01 | 11 | 00 | 10 | 00 | 10 | 10 | 01 | 01 | 11 |
| XOR correction | 1 | 0 | 0 | 1 | 1 | 1 | 0 | 1 | 1 | 1 | 1 | 1 | 0 | 0 | 1 | 0 | 1 | 1 | 1 | 1 | 0 |
| Von Neumann | 1 | 1 | 1 | 0 | 0 | 1 | 1 | 0 | 0 | 1 | 1 | 1 | 0 | 0 |
If "None" is selected, the unaltered raw output of the RBG is used.
For more information refer to the online article "Building a Hardware Random Bit Generator for EVPmaker".
Usually the random values read from the sound card or from the file or delivered from the fractal function are used as they are to select the segments. With this option activated, they are converted into new 32-bit numbers bit by bit first. If the values are used as they are (the option is deactivated), the delivered values are distributed proportionally to the range of values of the random numbers among all existing segments resp. cues of the loaded WAV file: The smallest possible value is assigned to the first segment, and the biggest possible value to the last segment. This leads to a strong correlation between the random numbers and the individual segments. This method is especially suitable for random sources which deliver a signal with a lot of variation, as for instance white noise with a high amplitude.
For very weak signals (such as for instance the internal noise of the sound card), this method would yield only a few diffrent random values which wouldn't be sufficient for the random-controlled playback of the segments. Because of that, with an activated option, the numbers are converted bit by bit into new 32-bit values according to the following rule: Starting with the first bit (bit 0) of this new number, the bit is set to "1" if the current random value is greater than the last one, to "0" if it's less, and to the value of the previously set bit if it is equal. This is done for all 32 bits. With this method, 32 values from the sound card, from the file, or from the fractal equation are needed for one random number. Using this mehtod, also signals with a very low amplitude can be converted into sequences of random numbers. Even a sequence of zeroes and ones is sofficient for this.
If you aren't sure, try both methods and judge by your hearing what sounds better.
This option is related only to the Sound card setting. If it is activated, EVPmaker attempts to use the best possible sample rate of 44100 Hz for the recording, in order to miminize the recording time and thus the period in which no random numbers can be produced. (If the recording time exceeds the playback time for 20 segments, which may occur when very short segments are used, then there will be audible gaps, since no more random numbers can be supplied at this moment.)
Some sound cards however can only use one sample rate at the same time for the simultaneous recording and playback. If for instance the WAV file used by EVPmaker is sampled with 11025 Hz, then a noise signal can only be recorded with 11025 Hz as well with those kinds of sound cards. If this isn't successful, the sample rate of the input WAV file will be used instead. Since this attempt costs some time, this option should be deactivated with these sound cards.
This option is only effective if Internet is selected as the random source. If this option
is activated, new random values will be retrieved from the Internet service each time the
random controlled playback is started by clicking the (Make EVP)
button. Without this option activated, new values are only read if the previously read 1024 values are "used up".
Assuming a paranormal influence on the noise source, the use of "fresh" random numbers for each new EVP would probably
be more effective...
This option is only effective in conjunction with the File setting. After reading all integer values
from the file, it is "used up", and no more random numbers can be generated from it. The EVP generation will then be
halted and cannot be proceeded. Of course, the same file can be used as often as you wish by simply opening it again,
or by restarting EVPmaker, but this can be done easier: With the Loop file option activated, the file
will automatically be "rewinded" after reaching the end, and the whole process will start over. The random controlled
playback will continue until (Stop) is clicked.
However, it should be considered that through this "infinite loop" the series of generated random values will be
repeated continuously, and in conjunction with the same WAV file and the same settings for the segment length, this
will result in a continuously repeated "raw material". Based on an average segment duration of 50 ms, a 1 MB file
will yield approximately 7 minutes of continuous raw material before it is "used up" or will be repeated.
To prove a possible mental influence of test persons on random events during PK experiments, the used sequences of random numbers can be analyzed statistically. If this analysis shows a statistically significant deviation from a mean value over a longer period of time with numerous individual runs, this result can be interpreted as an PK effect.
To examine the random number sequences for such deviations, EVPmaker offers a function which allows to log the used random numbers into a text file during the random controlled playback. Another function allows to save large blocks of random numbers to a file, independent from EVP gereration, for comparison purposes (this is the "reference material").
Here follows a description of these two functions:
Here you can enter the name of a text file, into which the used random numbers are written during the the random controlled playback within their range of values. All numbers will be written into the text file one per line. After entering a file name, the program is ready for logging. The logging is switched off again by selecting this menu item once again.
Larger blocks of random numbers can be written into a text file as a "reference". To be suitable as a basis for a comparison to the random values logged during an EVP recording, of course the same random source should be used as for the EVP recording, and also the same settings for the Bitwise convert 32 RN to 1 RN option should be made. Also all other settings as well as used files etc. should be the same. After selecting this menu item, you are frist asked to enter the number of random values to be written into the file:
The number can be increased/decreased in thousands using the two up/down arrows. Using the two buttons besides it, the displayed value can be divided or multiplied by 10. Of course, the number of the random values can also be entered into the input field.
After clicking "OK", another dialog box appears to enter the file name. Writing a large number of random values may take a while...
The Windows mixer lets you set the recording and playback levels of the sound card. It can normally be opened by double-clicking the speaker or mixer icon in the task bar on the lower right. However, the recording mixer can be reached only inconveniently via the mixer's "Options > Properties" menu. For a quicker access to the recording and playback controls, the Windows mixer can also be opened via Extras > Audio Mixer > Playback resp. Recording in the EVPmaker menu.
Various presettings of the program can be made in a "Settings" window which can be reached via Extras > Settings in the menu:
If this option is marked, EVPmaker will remember for each of the following file types the last recently used folder and suggests it in the corresponding "Open" or "Save" dialog:
Without the activation of this option, all "Open" and "Save" dialogs will use the same last recently used folder.
If this option is checked, the program window will appear at the same position on the desktop the next time it is started as when it was quit previously.
With this option checked the current settings will be saved automatically each time when the program is terminated. By clicking on the "Save now" button all current settings including the current position of the program window on the desktop will be saved, so that these values are immediately available when the program is startet the next time.
When this option is checked, the recently loaded WAV file will be automatically opened when the program is started.
If there is more than one sound card inside the computer, you can select here which one will be used for Playback, for Record, for Random Source, and for Live audio source. The milliseconds value for Recording buffer determines the size of the cache memory for the record function. This value is set to 100 ms by default and should only be increased if there are audible gaps or clicks during the recording on slow computers.
In the Help menu there are four items: Manual displays this software manual. Quickstart Guide opens an introducing Step-By-Step guide with practical examples. Go to Website / Download loads the WWW page into the Web browser where you can find current information about EVPmaker or download the latest version. This will require an open Internet connection. About EVPmaker displays the version number and copyright information of the program.
[1] Source: VTF-Post P 51, issue 2/88 - 1.4.1988, page 42. - Return
[2] Phonemes for speech synthesis, displayed using common symbols from the Windows ISO-8859-1 character set:
| Vowels | ||
| Symbol | Comment | Example |
| a | hat [hat], Tat ['ta:t] | |
| ä | hätte ['hät3], wählen ['vä:l3n] | |
| e | Methan [me'ta:n], Weg [ve:k] | |
| 3 | turned "e" | bitte ['bit3], engl.: girl [g3:l], learn [l3:n] |
| i | vital [vi'ta:l], viele ['fi:l3] | |
| O | opened "o" | kommen ['kOm3n], engl.: before [bifO:] |
| o | closed "o" | Advokat [atvo'ka:t], Boot [bo:t] |
| Ö | opened "ö" | öffnen ['Öfn3n], plötzlich ['plÖtslic] |
| ö | closed "ö" | Ökonom [öko'no:m], schön [Sö:n] |
| u | kulant [ku'lant], Buch [bu:x] | |
| y | Glück [glyk], knüpfen ['knypf3n], Füße ['fy:s3], Krümel ['kry:m3l] | |
| Nasals | ||
| Symbol | Comment | Example |
| a~ | nasaliertes "a" | Abonnement [abOn3'ma~:] |
| ä~ | nasaliertes "ä" | Timbre ['tä~:br3]] |
| O~ | nasaliertes "O" | Fondue [fo~'dy:] |
| Ö~ | nasaliertes "Ö" | Parfum [par'fÖ~:] |
| Diphtongs | ||
| Symbol | Comment | Example |
| ai | weit [vait], Beispiel ['baiSpi:l] | |
| au | Maus [maus], Haut [haut] | |
| ei | engl.: date [deit], play [plei] | |
| Oy | Beute ['bOyt3], Heu [hOy] | |
| Word endings
(For words ending on "-r", the "r" is often not pronounced.) | ||
| Symbol | Comment | Example |
| äa | Berg [bäak], ordinär [Ordi'nä:a] | |
| ea | schwer [Sve:a] | |
| iO | Wirt [viOt], Tier [ti:O] | |
| oO | dort [doOt], hervor [häa'fo:O] | |
| ÖO | Wörter ['vÖOtO] | |
| öO | Gehör [g3'hö:O] | |
| uO | kurz [kuOts], Kur [ku:O] | |
| yO | Türke ['tyOk3], Tür [ty:O] | |
| Consonants (vocal) | ||
| Symbol | Comment | Example |
| l | lassen ['las3n] | |
| m | Maus [maus], Kummer ['kumO] | |
| n | nein [nain] | |
| N | singen ['ziN3n], lang [laN] | |
| w | engl.: whisky ['wiski], wild [waild] | |
| Plosives | ||
| Symbol | Comment | Example |
| b | voiced | Ball [bal], haben ['ha:b3n] |
| p | voiceless | Paß [pas], Weib [vaip], obgleich [op'glaic] |
| d | voiced | du [du:], dann [dan], finden ['find3n] |
| t | voiceless | Tee [te:], Stadt [Stat], Fund [funt] |
| g | voiced | gut [gu:t], Gast [gast] |
| k | voiceless | kalt [kalt], Tag [ta:k], Kuchen ['ku:x3n] |
| Fricatives | ||
| Symbol | Comment | Example |
| c | voiceless | Licht [lict], Mönch [mÖnc] |
| j | voiced | ja [ja:], jung [juN] |
| f | voiceless | finden ['find3n], Hafen ['ha:f3n] |
| v | voiced | Vase [va:z3], Gewinn [g3'vin] |
| h | voiceless | hat [hat], hin [hin], Hund [hunt] |
| x | voiceless | Achtung ['axtuN], Loch [lox], Buch [bu:x] |
| r | voiced | Rasen ['ra:s3n], Regen ['re:g3n], Rose ['ro:z3] |
| s | voiceless | Glas [gla:s], Masse ['mas3], hast [hast] |
| z | voiced | Sohn [zo:n], Rose ['ro:z3], Hase ['ha:z3] |
| S | voiceless | Schiff [Sif], Charlotte [Sar'lOt3] |
| Z | voiced | Genie [Ze'ni:], Journal [Zur'na:l] |
| T | voiceless | engl.: thin [Tin], path [pa:T], method ['meT3d] |
| D | voiced | engl.: there [Dä3], father ['fa:D3] |
| Affricates | ||
| Symbol | Comment | Example |
| dZ | Dschungel ['dZuN3l], Gin [dZin] | |
| pf | Pfanne ['pfan3], Kopf [kOpf] | |
| ps | Psychologie [psycolo'gi:] | |
| pS | hübsch [hypS] | |
| ks | Hexe ['häks3] | |
| kS | Eckstein ['äkStain] | |
| ts | Katze ['kats3], Zahl [tsa:l] | |
| tS | klatschen ['klatS3n], Wettstreit ['vätStrait], deutsch [dOytS] | |
| Directions | ||
| Auxiliary symbols | ||
| : | the preceding phoneme is pronounced long | |
| ' | the following syllable is stressed | |
| | | "glottal stop", e.g.: Verein [fäa'|ain] | |
| ~ | the preceding vowel is a nasal (see examples above) | |
| Annotations | ||
| Before plosives there is usually a pause of 50 to 60 ms. | ||
[3] These experiments which used the Verhulst Dynamics for the generation of EVP have been discussed in detail in 1999 and 2000 in the TBS-List (a German EVP mailing list). Unfortunately, a summarizing research report about that isn't available, but the articles can still be read in the archive of the mentioned mailing list. - Return
