Pundabaya wrote:
I might be talking out of my arse, but I think that the reason that Audacity has to re-export the MP3 (rather than just savethe two bits) is that it can't. Because MP3s are compressed, there's information about how the MP3 was intially compressed saved somewhere in the file itself, without that it won't work.
Think of an MP3 as being a Word .doc in a zip file. You can't just cut the zip file in two, and hope you get page 1&2 in the first half, and 3&4 in the second. You have to expand the archive, open up Word, manually separate the pages into two files, then re-compress them into zip files.
Audacity is going through exactly the same process when it loads your MP3. It converts it back to a .wav file, so you can fiddle with it and has to recompress the two halves when you've done.
Hmmm no that doesn't make sense because it's possible to record an MP3 in real time, so the MP3 recorder has no means of knowing when you're going to start and stop recording, so the file has to be complete at any moment in time. i.e. It's entirely self contained and any chunk thereof has to be self contained as well.
In the form of compression you're talking about (traditional zip files for example), there's already a defined 'end result' to be compressed, so the program involved knows it won't change and it has everything to work with before it starts.