Interestingly, there is now another poster who is using the TDOA method based on MatLab (link below). Does he also use a mislabeled source which only nominally comes from DT mic?
In such a context, one has also be aware of the possibility of deliberatly misplaced or even completely falsified audio tracks.