Terra Cava

Un esame del suono... non ho tempo di tradurlo purtroppo

http://spi.com.sg/spi_files/soh/main.htm

Dissect the sound file
The authenticity of this instance as speculated above ties much on the authenticity of the sound clip. In order to prove whether this is only a hoax or is a true occurrence, we put the sound clips under test by using a sound frequency analysis software program called AnalFreq Spectrum Analyzer Ver. 1.8 by Loyola University Chicago.

How do we test the sound files?
After we loaded the sound files in wave format into the program, we displayed its frequency spectrum using a FFT window of size 1024 & sampling rate of 22050 for optimum display. The frequency resolution was set at about a spacing of 43 Hz (0Hz to 22kHz displayed on screen) down to bin spacing of about 1.35 Hz. Hanning display algorithm was used. The FFT algorithm used by AnalFreq takes approximately 1 ms for the same size FFT.

The original sound clip was split into 3 sections: (1) only the first half where only a man was speaking on the narrative, (2) only the last half where only the "screams" from hell, and (3) somewhere in the middle where it has both a single man talking and then the screams.

With these 3 different sections of the original sound clip, we want to compare the frequency distribution of a single man speaking, and that of millions of souls screaming as claimed. We do this by repeatedly showing the frequency distribution by FFT analysis. Concurrently we run the dual mode of Freerun, in which the (almost) instantaneous FFT plot is displayed, and Long Term Average, in which the time-averaged plot is displayed. Pay close attention to the changes of frequency distribution on the charting plot when running these 3 different files.



Our Findings


Frequency Distribution of a Single Man Speaking
The red line in both figures represents the long term average values in the spectrum, which doesn't fluctuate as much as the blue bars do. We see that the distribution weighs on the ranges of 0.5Khz to 3Khz, also some humps on 4.8Khz and 6.2Khz which are quite normal for a male's voice. High pitches are rare.

Frequency Distribution of a Million Souls Screaming
In Figure 2, again, on those frequency ranges where a male voice would manifested (0.5-3Khz, 4.8Khz and 6.2Khz), the frequency distribution falls. However, there are two major differences. Firstly, addition to the low band and mid band frequencies, some high pitches are recorded in 8Khz, and likewise from 9Khz to 10Khz. Secondly, the long term average values in those ranges seem to be more abundant than those in Figure 1. What do they mean?

Results of a Million Souls Screaming?
The abundance in the distribution can be explained by the fact that there were multiple voices being recorded making a sound. It is like putting many single man's frequency distribution together, so to reach an equilibrium amount. The high pitches frequencies indicate that there were some high-key screaming, possibly by female. Although you can not see the results in motion, the red line in fact has fluctuated very little during the play. As the voices were going kind of on-and-off, they made the line vibrates slightly.

So what does it all mean?
According to our observation when playing the file no.2 (pure screaming), there were "many" voices aggregated together to produce such a result. However, that "many" would never come close to a huge number for example a million. If we blend in a million voices together and we assume these voices come from different individuals who have different voice patterns, one would expect it to have some kind of "smoothing" effect to the long term average curve. That is, perhaps, a bell-shape or a more uniform curve than what is shown in Figure 2, to be observed. The frequencies distribution would get more evened out. On counter-proving our arguments, the souls may have quite a monotonous voice such that the combined frequencies would only be concentrated on the ranges (coincidentally?) same as those in Figure 1. Then these may be like duplicated folks of the same breed. Technically, on the other hand, this may also possibly be the results of artificially mixing only several different tones of screaming so to produce this grand effect of millions souls screaming. The sound is quite "mechanical" when it is played in a long run. You can do a little experiment on your own to test this. Play the sound clip by using Media Player (or anything that gives you a slide-bar). When it is being played, move the slide-bar back and forth randomly. While doing this, you will be surprised to hear that the sound was still playing smooth and continuous. Why? That is because you were listening to an aggregated result of voices that got randomly "pieced" or "cascaded" together using a mixer.

Our conclusion is, either it is true that the voices are from hell but the souls were shouting constantly at about the same tones, same volume and at about the same intervals, or somebody just mixed several voices to achieve this effect. If the answer were the latter, then he succeeded very well in blending many screams together. But the frequencies are too regular and the tones are too rigid.


p.s. le immagini sono sul sito

[Modificato da +Vurdalak+ 09/04/2007 19.45]