You do realize that 75% of the "CONS" on your app are demonstrably false, don't you?
1. You can load sounds on-demand from your server (web is easiest, but you can make any protocol you want). If the user wants cow sounds, no need to send horse.
3. If you put your catalog of sounds on your server, your app can check the version of that catalog and update itself. You do not need to send an entire app update because you've added one new MP3 file.
4. An iPhone is ... well ... a phone. It can record sounds.
Now ... your CON #2 is correct - it ** IS ** easy to update sounds, catalogs, and allow information to flow back-and-forth between the phone and your servers (Facebook, Flickr, Twitter - these are just a few of the GIANTS doing this, and ** LOTS ** of small guys do the exact same thing). Dunno why this represents a problem, but haters gotta hate!
If you are going to allow the user to apply effects & filters (reverb, distortion, high pass, low pass) then running it on the phone means your servers aren't taxed with all this nonsense, and there are several DSP's already built-in you can access to keep it as hardware-as-possible.
Turn on iTunes file sharing (UIFileSharingEnabled), and the users can drop files directly into your app's Documents directory. Make collections and let the users mix-and-match.