Scientists were able to do some experiments because they have a data set of recordings of more than 600 people in Germany Of whom 111 had a confirmed cold. They could then use essentially computers and machine learning algorithms to analyze the relationship between the loudness of those harmonics and The computer was able to find patterns in those relationships that allowed it to distinguish healthy voices from sick voices Hmm interesting David, I don't know if you can tell but I've got a bit of a sore throat myself this morningHmm. So to refer the algorithm is not foolproof yet in these tests It was able to correctly identify a voice with a cold about 70% of the time Now that's obviously better than chance But