I have a lot of live material that I have ripped & encoded using EAC + LAME. While listening to it on my iPod, there is a noticeable switch between tracks. From what I've read, this can't be entirely eliminated using the current LAME (sounds like at the least there will be 100 of so samples missing at the beginning of the track) - it even sounds like perhaps it can't be done with any MP3 encoder (using individual files for each track).
I would like to write a program that will optimize the WAVs to reduce the gap between tracks before running them through LAME. For example, I can at least make sure that each WAV has a # of samples that is an even multiple of 1152 by taking the needed # of samples from the following track. What other optimizations can I make? For example, it sounds like something goes on with the last frame or two that might benefit from borrowing a few frames worth of samples from the following track when encoding, then perhaps cutting them back out from the resulting MP3.
... or will compressing with --nogap fix all of my problems?