MP3 recordings treat the audio in much the same way… they toss out as much as they can, and leave just enough pieces that your brain can fill in the gaps without too much trouble. MP3 coding can leave out 80 – 90% of the original material!
This is the reason that bit-reduced audio (mp3, aac, ogg, and others) should be used only for the final step in the audio chain. The “reduced” audio is specifically designed to be understood by human ears… not necessarily by another piece of audio editing software.
If you record an interview on your portable machine using mp3, it’s already been reduced within an inch of its life. Then you re-inflate the audio to edit the file on your PC. When you’re done editing, if you save the file as mp3, the algorithm again tosses out as much data as possible. The problem is, it was already “reduced” as far as could be reasonably tolerated… now you want to take out more? This won’t be pretty.
Storage is cheap these days. (A 2TB – that’s 2000 Gigabytes – hard drive can be had for less than $100) You should always record new material in full linear mode (PCM, WAV, AIFF) and at least 44.1KHz, 16 bits. Keep the audio “full-size” throughout the editing process. Then play the non-reduced file on the air, if possible. This applies to music, too. Don’t play an MP3 if you have access to the CD.
If your playback system requires that the audio be stored as mp3, use the highest bit rate you can… at least 256K.
The file you play is going to be manipulated several more times before the listener ever gets to hear it. Give the audio a chance to make it through the maze with minimal battle-damage.
TK U FR UR ATNTN