components more than together ultimately speaker recognition with human in the loop. that's another term. >> you always have a trained examiner listening? >> that is correct. >> your examiners are trained in the various methodologist available for speaker recognition? >> that is correct. the computer cannot really assess whether, you know, inputting voices produce inference or computer cannot tell whether a person was happy, sad or voice was you know produced under, you know, normal duress. so to detect all of those, we have to depend up on the trained examiners. >> all right. we will step away and sort of like a at least conclude my show with a synopsis of what