James Perrett wrote:I'll bet Hugh has already written a detailed article explaining all this far more clearly than I have.
No need -- you've done a great job of it. :D
All I would add is that the loss-less formats (FLAC and ALAC) reduce the original WAV or AIFF file size by roughly 2:1, so the are data-reduction schemes, but not very efficient ones. However, the files can be unpacked and perfectly bit-accurate versions of the originals rebuilt.
In contrast the lossy formats (MP3, MP2, AC3, aptX, AAC, ATRAC, Ogg, WMA, and others) manage much smaller files sizes -- some will reduce the file size by as much as 20:1, and 12:1 is common. But they achieve that by throwing away audio elements that the designers think the average listener won't notice when hearing the file replayed in the intended conditions.
Passing a lossy-coded file through further stages of lossy codec will degrade the signal quality very quickly -- much like copying a cassette tape from one machine to another and back again (for the older generation ;-) ) ...
With one or two particular exceptions, lossy codec files are all intended for end-user applications: you get the file and listen to it. They are not intended for further signal processing of any kind, or subsequent re-encoding.
Loss-less files can be re-processed and re-encoded without problems...
And that's why you should always use original WAVs/AIFFs or lossless files in production and post production. Lossy files are generally fine for consumer listening (assuming a sensible bit-rate is chosen), but for quality applications loss-less files are appreciated.