An audio CD is mostly arranged like a single continuous recording. Tracks are added on top of this via the Q subcode channel that gives information about the current location and the ToC stored in the lead-in area (also using the Q subcode channel). In the ToC, each track will have one more indexes that points at a specific location on the disc by minute, second and "frame" (represents 1/75 of a second, basically a sector).

If a CD is properly following the Red Book standard, index 0 will point to a 2 second pre-gap of silentce and index 1 will point to the actual start of audio of the track (additional indices are allowed, but not common). The purpose of the pregap is to make life easier for less sophisticated players that aren't able to seek to a precise frame on the disc. They just have to be able to hit a 150 frame region. However, just because the standard says the pregap is supposed to be 2 seconds and silent doesn't mean it actually has to be. Players generally don't care and by the time the format was popular, even inexpensive players could seek precisely. This allows you to stick audio data before a track that will be skipped by the player when it's trying to seek to that track. If you stick it before track 01, it will be skipped even when just playing the disc through unless you rewind.