Music starts early

I have noticed this for some time but never got around to asking the question. Some LP's that I play, I will hear faintly the music start. Then all of a sudden the music does start. Does that make sense? In other words. At the start of an LP. You will hear very quietly the song start. Then the real song starts at the correct level. What is up with that?
In most cases it's the result of the original tape having bled through to the blank lead in portion of the tape. Some have offered that the grooves are spaced too close together and you are picking up information from an adjacent groove. I don't buy into this opinion because the information you hear would be one channel only and reversed from the real music. Also, if grooves spaced too closely caused this it would be audible during silent passages in that cut. Every example I've heard in my library is both channels in their correct relationship. This phenomenon is called pre-echo. I wish I knew the definitive answer to this myself.
Right on the money, Lugnut. Tape 'print through' was and still is a problem for any magnetic tape storage media. Makes you wonder if remasters from old analog tape could get any better than the fresh pressings...
Having spent 35 years as a Producer/Recording Engineer, I think that can shed a little light on this phenomena?

It is, indeed, known as 'pre-echo' and one of the main causes has to do with how the original master tape was wound onto its reel and stored. It is common practise in the recording industry to 'tail wind' tapes to ensure that the results of this pre-echo phenomena are not as much of a problem. I'll explain:

What is actually happening is that the magnetic information (music) is being transferred between the thin layers of magnetic metal oxide of the recording tape. Professional recording tape is commonly a mylar or plastic composite 1.5 mil thick, with some of it - for longer recording times et cetera - being only 1 mil. I used the latter extensively for live recording and the former for studio recordings. This transference of magnetic information is a natural occurance, given the nature and physics of recording tape. If the tape is rewound, after use, back to the beginning of the music (head wound) and stored this way, any transferred information will be printed through (known as print through) onto the tape, appearing before the original information from which it was copied. This is what you hear as pre-echo.

In order to minimize the effects of this, tapes are usually wound 'tails out' or 'tail wound' and stored this way. In order to replay the tape it, of course, must be rewound to the beginning. Information that magnetically transfers through the layers (print through) is, for the most part lost in the decay of the music that proceeds it. The volume and frequency of the information, as might be expected, is also an important factor, with loud passages of music being the most obvious ones to transfer between layers.

Another trick to circumvent the phenomena was to edit the leader tape (which is just a plastic or paper tape of the same width as the recording tape being used) right up to the downbeat of the track. This works, but it is always a good idea to leave just a little bit of tape - 1/8" to 1/4" is enough as the tape is travelling at 15ips or 30ips (Inches Per Second) - ahead of the first note, just in case the tape curls or otherwise becomes damaged at the point of the splice. The problem was that sometimes this master tape was copied as part of the disc mastering process in order to document the various changes that might have been made - equalization, limiting/compression, volume and fades and/or re-sequencing et cetera. This new version of the master tape would then become the official Mastering Tape for that project and, at that point, you once again have the possibility of pre-echo.
Sorry to disagree but LP groove echo, not tape print-through, gets my vote as the most commonly perceived occurence by the home listener. Like many, for years I thought groove echo was a playback induced event. It's not. Groove echo is an artifact created in the LP cutting and plating process, independent of the master tape. See Bob Ludwig's explanation:

Certainly tape print through pre-echo exists and causes problems, but I doubt it's the one that we're most accustomed to hearing.
I'm completely in agreement with Rockvirgo on this. Here are my reasons. It is standard, no mandatory practice in the recording industry to store tapes tails out and yet all vinyl recordings show this effect to some degree or another. The degree to which it is audibly apparent is proportional to the degree that the cutting is modulated. The end of the argument is this: I have many recordings on vinyl that were tracked and mastered digitally and never "saw" an analog tape. This is conclusive evidence that LP groove echo is at work. Plus the music is heard exactly one revolution of the LP before the main music kicks in, and there is no way that one revolution of a tape reel corresponds exactly to one revolution of LP every time.

And Lugnut, you are incorrect in the assumption that the music would be reversed. After all the LP, and hence the modulated adjacent revolution is still rotating in the same direction.
It would also be one channel "only" if your argument held up. Every track has a blank lead in groove. The left channel would be picking up right channel only information and nothing from the blank lead in. So, again, if your argument held up the right channel information would be a weak signal in the left channel speaker with nothing in the other channel. Sorry, I can't buy into anything other than tape bleed through as so thoroughly explained by Scouser.
Lugnut, the echo isn't coming from the stylus improperly picking up the adjacent groove, the stylus is tracking the info that the lacquering and metal mastering process created and left behind. As you note the entire groove echoes the one next to it. The of necessity soft nature of the lacquer encourages this. Like a sound echo you make with your voice this one is made by the cutting head deforming its surrounding environment, however slightly. The blank lead in grooves are especially affected because they have no signal cut into them to override the echo. As noted in the Ludwig article, the longer the lacquer sits before making the metal master the more the echo cures into the adjacent grooves.