Helping improve the voice recognition?
Has there been any discussion or proposals to allow community members to submit recordings of their voice as correct when they are having issues with particular microphone exercises? I can get the majority of these exercises no problem with my laptop microphone, but every now and then I'll have to skip an exercise because Duolingo can't figure out what I'm saying. An example is the phrase "Usted volvió", I tried a dozen times and could not get my laptop to understand usted. The word itself does not seem to be a problem, as I can get the program to hear it for other exercises where it's used.
It almost seems like, for whatever reason, sometimes my voice doesn't seem to fall into any sort of correct buckets for the voice recognition program. It would be nice if we could submit correct answers and have them judged as correct or not like you can for written exercises. They could then be mixed in as more correct examples to improve the speech recognition program - I'm assuming it's some sort of machine learning algorithm that has training examples it has learned from?
What interface are you using? On my iOS app, gives me three tries to get it right and then moves on BUT I don't loose any hearts ( or presumably not go back one step in the Web interface).
Lately I have been using Chrome on my laptop. It has switched to the no-heart, keep trying until you fill up the bar for the lesson version (highly recommend if you haven't updated Duolingo lately!).
I seem to get unlimited tries before it keeps me from continuing, but once in a while I notice it will actually just fail me and make me move on. This could be a difference between practice and doing a new lesson? Could be I'm just not keeping count well? Not sure, but most of the time I have to click "Skip" explicitly to move on if it won't recognize me.