(Note! This is not a direction for use but just an example to give an idea of a computer aided creation of a para-voice. The whole process is illustrated by screen-shots.)

 Step 1

The first of the two start files:
20 seconds of brown noise.

 Step 2

The other of the two start files:
20 seconds of white noise.
No additional material is used!

 Step 3

Through several mutual manipulations (filtering, adding, multiplication etc.) the two files are combined to one and contains now the "voices" but hidden by a large amount of noise which has to be removed.

Step 4

Noise removing 1 

 Step 5

Noise removing 2

 Step 6

Noise removing 3

 Step 7

Noise removing 4 - The last noise removing.
When you have to stop the removing process is a question of balance between voice quality and the rise of a new (different) background noise.

Now you can zoom in the selected section and examine the voice.