What Camelot notation actually tells you

Every track in a prepped library carries a key. Most software will show it two ways: as a musical key like F minor, and as a Camelot code like 4A. The second one looks like a serial number, which is a shame, because it is doing something the first one cannot.

The two facts in the code

A Camelot code is a number from 1 to 12 and a letter, either A or B.

The letter is the mode. A is minor, B is major. That is the whole of it.

The number is a position on a wheel of twelve, and the wheel is ordered by the circle of fifths rather than alphabetically or by pitch. That ordering is the useful part. Two keys that sit next to each other on the wheel share six of their seven notes. Two keys on opposite sides of it share almost nothing.

So 8A and 9A are adjacent, and a phrase from one lands inside the other with a single note of disagreement. 8A and 2A are six steps apart, and layering them gives you two chords fighting over the same bar.

What that buys you in practice

Three moves are safe from any starting code, and they are safe for reasons you can state:

• Same number, other letter. 8A to 8B is the relative major and minor. The notes are identical. Only the note the ear treats as home has moved, which is why the mood changes and nothing clashes.
• Plus or minus one. 8A to 9A or 7A changes exactly one note in the scale. This is the ordinary mix, and it is the reason the wheel is ordered the way it is.
• Stay put. 8A to 8A is not a cop out. Two tracks in the same key can sit on top of each other for as long as you like.

Everything else is a decision rather than a default. Moving three or four steps is a real modulation, which is sometimes exactly what a set needs at the point where it should stop being comfortable.

Three things the code does not tell you

A track has one code and may not have one key. The label describes the file as a whole. Plenty of records change key partway through, and a breakdown that lifts into a different mode is common enough that the tag on a nine minute record is better read as “mostly” than as “throughout”. The part you are mixing out of is the part that matters, and it is not always the part the number describes.

Detection is inference, not measurement. Tempo can be counted. Key is estimated from which pitches are present and how much weight they carry, and records that are mostly percussion, or mostly one sustained note, give an estimator very little to work with. Ambiguity between relative majors and minors is the usual failure, because those two share every note. If a code looks wrong, it may be.

Compatible is not the same as good. Two tracks can share a key and still be a bad idea back to back, because energy, texture and arrangement all live outside the code. The wheel tells you what will not clash harmonically. It has no opinion at all about whether the transition is worth making.

That is still worth a great deal. Knowing which pairs cannot hurt you means the part of the decision left over is the part that was always going to be yours.