this is a machine we have had for a while, it’s called a PSE, Psychological stress evaluator and it works...
The signals any machine will get will have been gained through initial editing at the source plus the digital reduction algorithms used before uplink transmissions thru satellite and digital fiber optics. ( Addditionally audio compression limiters will modulate the extreme low and high amplitudes of the debaters’ voices to reduce audio distortion so that the sound sources have a narrow bandwidth of say 20 to 30 db’s.} Then the signals are processed again before sending them on thru cable networks to the viewer’s television sets including the equip used to “process” the voices for stress analysis. There will not be anyway these folks can get a direct feed from the candidates mouths so that their “stress” modulation peaks and valleys can be analyzed directly by themachine itself.
In short without a direct feed from candidate to machine, the data will be useless. What’s worse, if the folks at the preprocessing side know about the “lie detector” folks, they might be politically motivated to tinker with the audio feed to introduce false stress artifacts of the hated politcal candidate into the uplink feed.
Both candidates are getting intel updates that they can’t necessarily talk about, so if they are asked questions touching on sensitive data, the machine may pick up on
both candates personal stress when they avoid or spin a question that relates to that sensitive material. (though I suppose an answer such as “I will not talk about that issue since national security may be involved” might be how an otherwise honest candiate gets around such questions)