When I record, I use a gibberish sound file. This consists of phonemes (approximately 44 in the English language), that is chopped up speech representing every part of our language without actually making sense. Over the last year I have used half a dozen files. The files are abandoned by me once they are corrupted by previous recording sessions burning answers onto the file (I don’t know the cause for this). I most recently attempted to use a female Portuguese file, and though the team made it through eventually (all male) the sound was disagreeable so I returned to my male gibberish and will continue with it till I find an alternative.
Admittedly I haven’t tried too hard, but I have so far failed to receive EVP using traditional methods such as silence, or radio. I’m also aware that sound can travel up to two miles given the right circumstances, so prefer my method.