You have a loop. You need its key before you can write anything under it. Four ways to find out, roughly in order of effort.
Method 1: find the tonic by ear
Still the most reliable method, and the only one that also teaches you something.
- Loop the sample.
- Play single notes on a keyboard until one feels like home — the note the loop wants to resolve to. That is your tonic.
- Play the major third above it, then the minor third. Whichever fits gives you the mode.
If step 2 leaves you with two candidates a third apart, you have found the relative major/minor pair — A minor and C major, say, which contain identical notes. Listen for which one the loop resolves to rather than which one contains its notes.
Time: thirty seconds once you are used to it. It never confuses a tonic with its relative the way an automatic answer can.
Method 2: match against a reference
If you cannot find the tonic by ear, work backwards. Load the sample, load a chromatic instrument beside it, and play an ascending minor scale from various roots against the loop. When the scale stops clashing, you have the key.
Slower, but it works when the loop is dense or heavily processed and picking a single tonic by ear is hard.
Method 3: your DAW might already know
Some DAWs suggest a key for imported audio. Where it exists it is free and instant, and it is a perfectly good first guess.
The caveats are the same as for any automatic answer, plus one specific to DAWs: key detection there is a side feature rather than the product, so it tends to assume standard tuning and rarely tells you how sure it was.
Method 4: an analyser
Drop the file in, read the answer. Fastest option and the only one that scales past a handful of files. DropWave does it in about the time it takes to let go of the mouse, and gives you the Camelot position alongside the key so you can immediately see what else it will sit with.
The rest of this article is the part nobody covers: how to know when an automatic answer is wrong.
The four ways it goes wrong
1. The sample is detuned
The big one, and far more common than people expect.
Sample packs are routinely recorded, resampled or bounced off concert pitch. Anything pitch-shifted is off the grid by definition. A tool that assumes A = 440 smears that energy across two neighbouring pitch classes and lands one key over — confidently.
The check: does your tool report a tuning offset? If it does and the offset is large, you have learned something real about the file. DropWave measures the tuning before deciding anything and prints the deviation in cents. If yours does not report one, pitch a known file by 50 cents and see whether the answer survives.
2. Relative major and minor
A minor and C major contain exactly the same notes. So do D minor and F major, E minor and G major, and every other pair a minor third apart.
Nothing in the note content separates them — only which note the music resolves to — and eight bars of a loop may not resolve anywhere at all.
The check: if the named alternative is the relative key, both answers are defensible and you decide by ear. In Camelot terms these are the same number with a different letter — 8A and 8B — which is exactly why the wheel treats them as compatible.
3. The sample has no key
Drum loops, hi-hats, risers, noise sweeps, filtered white noise, one-chord pads. There is no tonic. An analyser will still print something, because noise correlates with something.
The check: look at the confidence. A tool saying "unsure" on a hi-hat loop is behaving correctly. One printing "F♯ major" in confident type is telling you its confidence display means nothing.
4. The sample modulates
If the loop changes key halfway, there are two right answers and you will be handed one.
The check: chop it and analyse the sections separately.
A practical routine
For a folder of unlabelled samples:
- Analyse everything. Note the key, the Camelot code and the confidence.
- Trust the confident readings and move on. That is most of the folder.
- Check the uncertain ones by ear — tonic versus relative, thirty seconds each.
- Ignore readings on unpitched material. A drum loop's "key" is not information.
- Rename with the key and Camelot code —
vox_loop_Am_8A.wav. The filename is the most portable metadata there is: it survives every tool, every transfer and every backup.
And if the material is a channel in a session rather than a file on disk, you can read the key without bouncing it at all.