Is anyone else having problems understanding the audio exercises?
Is it just me or does the TTS engine pronounce certain phrases really unclearly? I don't remember having had this problem to the same extent with other Duolingo courses.
It does in some cases, and we're disabling the listening exercises for those sentences as we find them.
However, there is always a learning curve when it comes to distinguishing sounds in a new language that you aren't used to having to distinguish between in your own. Some nuances that are very clear to the native ear, can initially be hard to pick out for a learner.
We get -a lot- of audio reports for the first few skills in the tree. Not because there is more faulty or unclear audio in those skills, but because the learners at that early stage of the learning process haven't had the chance to attune their ears to the new language yet. Fortunately, that is something that improves with time and exposure. :)
I see that you've also taken the Spanish course here, and the TTS there does indeed sound a bit better. This has to do with how phonetic and regular the pronunciation of Spanish is. It's very robot friendly!
Interestingly, the English TTS is reasonably good, even though no one would ever describe English as phonetic or regular. That's presumably because an enormous amount of money has gone into developing English TTS systems. Norwegian is considerably more regular, but not quite as popular.
At a secret Norwegian army base somewhere north of the arctic circle the team developing the robot viking army is hard at work on the problem.
It has been done: http://www.telenor.com/wp-content/uploads/2012/05/T03_2.pdf
Telenor used quite a bit of resources on TTS and voice recognition in the nineties/early twenties, but I think this has ground to a halt as more and more private enterprises has come into the same market.
Yes, I think some of the problems are with how "flexible" Norwegian can be about pronunciation. I still don't recognize every word I know how to write.