The analogue method.

One of my very early experiments was a repetition of Belling and Lees investigation. The unmodulated hiss was replaced by distorted speech from a radio-receiver. This distortion was obtained by "wobbling" the radio-receiver during my experiments. (Wobbling is a term taken from electronic engineering. It's a quick movement up and down of the receivers frequency). During my experiments the receiver was electronically tuned up and down over one or more close situated foreign language transmitters. When listened to the output from the receiver, this strange "wobble-mixture" was almost unintelligible, but after having been recorded on a taperecorder, a careful listening at reduced speed could reveal several voices. Also the number of voices had increased considerable and in very very few cases it happened to get music converted to speech. Nevertheless, there still was a week point - the totally unpredictable radio-receiver. So I made up my mind and discarded all radio based experiments once for all. What I was looking for was a method which allowed a simultaneous recording on a multitrack-recorder of both the input and output material. A material which had to be speech like - but not speech. 

The new method used an electronic switch to alternately cut out short segments (about 100 ms) from two independent tape-recorders, each running a three minutes "endless" tapecassette, pre-recorded with foreign speech, mostly Serbo-Kroatic or Russian. The output then consist of an alternating flow of segments from the two recorders, normally with a short pause between the single elements. The whole apparatus was controlled by a four- track recorder (picture below) which simultaneous recorded: 

Track 1 and 2: The separate speech-segment output from each of the two tape-recorders running the pre-recorded endless cassettes - the input.

Track 3: The flow of the alternating segments - the output.

Track 4. In a later apparatus the output of track 3 was used to modulate a small radio-transmitter which then was demodulated (received) and registered. The improvement was very small - if any at all.

What happened was that the output from track 3 (or 4) on several places now was changed and not only single words but also whole sentences could be heard. 

Through a comparison of this new content with the original signal it was possible to find the changes that had occurred. 

The apparatus was build in three equal models, the two used by German experimentators. Their results were the same as mine. 

 Analogue apparatus 1997 - the last before the computer took over. 

From the left: 1.The YAMAHA MT100 four-track recorder. 2.The two small cassette-players with endless cassettes (the material for the generation of voices). 3.The electronic switch with modulator and transmitter stage. 4.The power supplies. 5.On the top of electronic switch: Two input meters. Not shown: Cable control box below. 

contents next