Monday, 1 July 2013

Voices on the wind

Sound affected by distanceSounds of music outdoors are not unusual in summer with so many windows open. But what I heard recently was music in bits - fragmented. It was coming from a fair being held a couple of streets away. The music was fragmented because there was a stiff breeze blowing. Such conditions cause turbulent air flow which can break up sound into bits. It made me think of some reports I'd read of people doing EVP recordings outdoors.

Most EVP recordings appear to be made indoors, if reports on the web are a reliable guide. But some people do EVP sessions outdoors, often in reputedly haunted locations. The problems with this are fairly obvious. It will be difficult to identify (and eliminate from consideration) likely 'normal' sound sources, particularly if the session is happening at night, compared to more controllable conditions inside. But my 'fragmented music' observation suggests another important problem.

Suppose there is someone, out of sight, talking loudly enough to be recorded by an EVP investigator. In calm conditions it is likely that the natural origin of the voice will be fairly obvious to anyone listening to the recording. Even then, however, the distance may distort what the voice is actually saying. That's because air absorbs higher frequency sound, so making spoken words less recognisable at a distance. Though the resulting sound may be recognisable as a human voice, the words heard may be different to those actually spoken. There's an example of this in the EVP Gallery (here). You can see the effects of increasing distance on human speech in the audio spectrogram above. Time is the bottom axis, increasing rightward (0 to 28s). The right axis is frequency, increasing from the back (from 45 to 5000 Hz). The vertical axis is loudness. The same word is spoken repeatedly at increasing distances from the microphone. See how the upper frequencies are lost progressively with distance (towards the right).

But what happens if there is a bit of a breeze blowing? The words will become fragmented. There is probably no more potent source of 'formant noise' than bits of actual human speech! Formant noise is a non-speech sound which happens to contain frequency peaks in simple harmonic ratios that a human brain interprets as speech. Many natural sources of noise, like rustling paper, can produce such sounds. There's a lot more to the concept than that so I suggest you follow the link (here) if you are not familiar with it. If real human speech gets mangled by the wind, there is a good likelihood that it will be interpreted as different words to those actually spoken. It's intermittent nature will also make it sound more like typical EVP, which often comes in short bursts, rather than everyday speech.

So, a crucial question when listening to EVP recordings made outdoors must be, was it windy?

PS: Chinese lanterns are being blamed for more serious problems than UFO reports - here.

No comments:

Post a Comment