Early pioneers grossly underestimated the challenge of machine translation but nowadays the size of the task is widely acknowledged, as a leading algorithm developer at China’s Alibaba did recently with the comment that even just for text, never mind speech, it is still not even close to the level of professional humans in the field. This seems to denigrate unfairly recent advances that have been made since the addition of machine learning based on deep neural networks has been added to the armory. Indeed, recent progress made on the various underlying components has enthused the European Broadcasting Union (EBU) sufficiently to tackle the final frontier, full speech-to-speech translation with its ongoing project called EuroVOX (not to be confused with the…