I really like the WAV idea. I suppose one could also fiddle with the MP3 sampling and/or compression rate so the MP3 processor does not have to "work" as hard! The MP3 module says:
"Supported sampling rates: 8-48K and Bit rates 8-320Kbps audio files"
but no guidance on how this relates to startup time. But I think the WAV has the most promise for reducing the gap. I'm not sure if these startup gaps are of any concern to the bulk of applications and hence not worth cluttering the manual. Of course it's a simple matter to "Save As" WAV format in an Audio Editor. There is so much available memory in microSD cards that I can't imagine anyone needing the compression to save memory when you have 1GB for 4 or 5 sounds! Audacity example:
And that the MP3 processor would "prefer" to decode 16-bit vs. 32-bit floating point. 
