OK. I double checked again and you're right: The standard pronunciation is indeed [fæn] (English) vs [fɛn] (German). However, many people (including myself) do use the English pronunciation with a somewhat shortened vowel sound. So, as long as you don't stretch the vowel sound too long, I guess it's fine to use the English pronunciation. People won't notice. As to the TTS engine: The audio I'm getting is clearly incorrect. However (and very unfortunately), I just learned in another discussion that the sound people are getting depends much more strongly on the local audio setup and equipment than I'd have imagined. This goes as far as the difference between an English "W" and a German "R". This is very unfortunate since learners can't fully rely on the opinion of native speakers any more if there is such a variety in the audio itself. See here for our discussion of the audio issue: http://www.duolingo.com/comment/821339
@idshanks: yes, I totally agree with you. Inconsistency between users is a very serious issue because it basically means that you can't even trust native speakers. I'm still completely puzzled about what's going on here. In the "Röcke" example, the sound is clearly and absolutely incorrect if I use my laptop speakers but kind of ok (although not perfect) on my headphones. My guess is some strange interaction between compression algorithms and audio devices. (E.g. an audio device can't clearly output some Fourier components that some compression algorithm heavily relies on – just a wild guess). For learners this situation is most unfortunate.
Sound like it's saying "fin" on my end - definitely seems to be an issue with the text to speech synthesis being inconsistent from user to user. I've heard things that were clearly completely wrong that people I trust insist are being said right.
I hope Duolingo comes up with some manner of solution for this. I'd never put too much faith in TTS anyway, but it's certainly less than convenient. Perhaps if the TTS engine synthesised the sentences server-side and sent the audio to the user? I'm sure someone who knows better could come up with something better, but the increased consistency, so that people know for a fact if something is wrong or not, would be useful. Users hearing things incorrectly that natives are insisting are correct because they hear something different will lead to some people learning the wrong thing.
Kinda rambled, but yeah, just my thoughts :)
It's baffling to me. I don't think it's the audio output devices, or else all of my devices are coincidentally incapable, but they're of highly varied quality (5.1 speakers of decent quality, gaming headset of fairly poor quality, and studio monitors of excellent quality). I hear the same errors through all three devices, but all on the same computer. I can confirm that I hear the Röcke error too. Hopefully they can figure it out. :)