Amgine, 13/03/2013 20:24:
The benefit of a web app/service is the extremely low hurdle for users to become contributors (the original reason for wikipedia, remember.)
Sure. However, main loss of time is IMHO: 1) configuring your recording software to work with your hardware (Linux or Windows, I never managed to do it – not enough incentives of course); 2) cutting and possibly converting and renaming the recording into an audio file suitable for upload (and this for many files of few seconds – can't even imagine starting); 3) uploading it to Commons with all that this means. Finding words to pronounce and adding the audio files are minor problems. In short, I think that a tool integrating steps 1–3 would be so wondeful that the small burden of installing it wouldn't be an obstacle, in short it's not worth being a requirement in my opinion. However, Google Hangout shows that sometimes web applications are smarter at interacting with hardware than local applications, and my competence is close to zero, so my point is probably moot. (Sorry for wasting even more of your time by elaborating on it.)
Nemo
jQuery has a range of modules for simplifying uploads, but presumably it would be better as an api call[1].
A simple search on programmableweb.com finds a rather large range of tools which might be mashed into this project, including transcription (though I did not see an IPA transcription). There is a related project[3] found through pw which connects words to video of contextual use and pronunciation (en-only, I think.)