I agree with Boris(ch): it completely depends on the task. If students had to answer some comprehension questions, the quality of the recording might have been adequate. If students have to spell Lesotho or Seeiso correctly, then I would hope that they have encountered these names previously.
War das die gleiche Qualität wie hier? Excellent question.
Several people mentioned that you couldn't see Prince Harry's lips. As far as I know, that's not required for a listening comprehension activity. My students rarely see a speaker (unless it's me) when they are being tested on listening comprehension.
I thought this selection of Prince Harry's speech was actually quite understandable for BE; I've heard lots of accents that were much harder to adjust to from the get-go. Presumably students in the 10th grade in Germany would have listened to quite a number of BE speakers before.
The comments in Spiegel about Miriam Makeba seem more relevant. African accents could be much less familiar and thus harder to understand. I myself have had trouble communicating with people from Ghana, say, until I got used to their accent. If Makeba's speech did have music in the background, then that presents an additional challenge. While I would expose my students to this kind of situation in their first four or five years of language instruction, I probably wouldn't select that kind of sample for an important listening comprehension test.
Another criticism of the student mentioned in Spiegel is that the listening text contained vocabulary words that had not been taught. This objection also can't be evaluated without knowing the task set. In my opinion, unfamiliar vocabulary is part and parcel of authentic texts. From the first trimester of German my students are continually exposed to unfamiliar vocabulary in reading or listening texts. I tell them that's the way real life will be. The important thing is do they understand enough. However, if arriving at a correct answer requires students to understand a new word, then they have to be able to determine the meaning using their skills (cognates, context, prior knowledge etc.).
Students who have been exposed to lots of authentic texts should have developed skills for coping with linguistic uncertainty. However, if students have only been exposed to textbooks and the speech of a small number of non-native speakers, then an authentic text will be difficult for them.
That's why I tell my students all the time that it's important they hear lots of different speakers: males (I'm female), teenagers from our partner school in the north, actors, television personalities, singers from Austria, exchange students from Switzerland, whatever. The more different speakers they hear, the better they will be able to cope with a new accent, rate or pitch of speech, etc. Sometimes I play listening samples of speakers who mumble, but then it's not in on a test and there are generally additional clues for comprehension, such as a video, or questions that guide the students' listening.