The Vietnamese Coffee Flavour Wheel
A sensory vocabulary for Vietnamese green coffee, written separately for the two species, Robusta and Arabica. Each descriptor carries a sensory definition, a physical reference sample for calibrating the senses, and the growing, processing or roasting factors that govern it.
The purpose is a shared vocabulary between producer, roaster and buyer, so that one term means the same thing at every step.
Recognition practice
Five kinds of exercise, all drawn from the vocabulary in the Reference tab: place a descriptor in the right group, identify a descriptor from its definition, name the growing or roasting factor that strengthens it, and identify a descriptor from its reference sample.
This trains naming, not guessing what is in a given cup. What is recorded is your accuracy against the vocabulary, so you can see which groups need more work and which pairs you confuse. It concludes nothing about any particular coffee.
Change the species and level in the Reference tab; the questions follow whatever is selected there.
Session result
The data stays in your browser and is sent nowhere. The export is a JSON file holding every answer, for when the system moves to a server.
Cupping form
Built on the SCA Coffee Value Assessment (CVA), Standard 104-2024: one descriptive section scored by intensity, one affective section scored by impression of quality, and a total drawn from the affective section alone. The vocabulary comes straight from the wheel in the Reference tab, so the naming matches the document.
1 · Sample recordtyped and selected
2 · Cups, uniformity and defectsnumbers — deducted from the score
Each non-uniform cup costs 2 points, each defective cup 4. A defective cup has a clear fault in it, not merely a slight difference from the rest.
3 · Descriptive intensityscale 0–15 · not part of the score
This section records howstrong or weakeach attribute is, not whether it is good. Acidity at 12/15 says nothing about quality.
4 · Descriptors identifiedselect — multiple
Select the descriptors you actually recognise in the sample. The list follows the species chosen in section 1.
5 · Affective assessmentscale 1–9 · this is the score
Quality scale: 1 extremely low, 5 average, 9 extremely high. The eight sections below are the entire basis of the total score.
6 · Notestyped
Calibration
Two different things go by the name calibration. One is resetting your own senses against physical reference material. The other is getting a panel to speak one language, by having everyone grade the same sample and then seeing who sits where. Both are below.
This tool issues no certificate and imitates no certification scheme.
A · Panel deviation tabletyped · calculated
Enter each grader's eight affective scores forthe same sample. The page computes each grader's CVA score, takes the panel mean as the reference point, and shows who deviates on which attribute. The panel mean is a point of consensus, not a correct answer.
B · Calibration against physical samplesselect — progress recorded
Prepare the reference material named on the right, smell or taste it, then mark whether you recognised it. This is the only way two people come to mean the same thing by the word “caramel”. The list follows the species and level selected in the Reference tab.