<div>I'm planning to write an application called Katch, used to (manually) write down text that's "dictated" to you from an audio file. Since I use xine in my day-to-day private multimedia activity, I was at first inclined to use it for Katch. However:
</div>
<div>1. What if the user is in a country where Helix is needed to play MP3s legally?</div>
<div>2. What if xine turned up to be badly KDE-integrated or whatnot?</div>
<div>3. What if, when Phonon came in an year, I'll be forced to (non-trivially) rewrite all my code to get it Phononized?</div>
<div> </div>
<div>For issue one, a semi-sane solution seems to fork amaroK, stripping it down to the engine abstaction part. In fact, I already started that, but gave it up. Doesn't seem worth it.</div>
<div>For issue two, aRts comes to mind, but I've heard all kinds of bad things about it.</div>
<div>For issue three, I could either put up with the low-quality (?) aRts (so that the code would be easily ported to Phonon), or wait an year (which I won't).</div>
<div> </div>
<div>My only major requirement for the framework is that it should have pitch/speed control. Noise removal would also be nice.</div>
<div> </div>
<div>Any advice?</div>