make this better

[bwhitman]

from jimmcgowan

I'm using the EchoPrint Codegen for an OSX based project, and I was
looking through the iPhone example code, and it seemed a bit
overcomplicated as it does an audio conversion to LPCM of the whole
compressed file, writes out the uncompressed version to disk and then
re-reads the uncompressed file to input to the codegen. A more
efficient method is to create an ExtAudioFile for the compressed file,
attach an LPCM format description to the ExtAudioFile as its Client
Format, then read the first 30 seconds of the file into a buffer. The
client format causes the results of the read to be in the LPCM format,
so the buffer can be passed directly to the codegen.

I've posted a code sample at https://gist.github.com/1073528
Hopefully it will help iOS or Mac OS X Cocoa developers get off the
ground with EchoPrint a bit quicker.

Thanks to the EchoPrint team for giving is such a great tool!