CoffeariumBeta
  1. Plant
  2. Farm
  3. Process
  4. Trade
  5. Roast
  6. Brew
  7. Cup
Roasting · Quality

Roast degree and colour

How far a coffee has been roasted, and how it is measured – mostly as colour, by reflected light on ground coffee – with the scales in use, why they do not convert, and what colour does and does not tell.

Across the chain: Science · Equipment

Source-checkedby Coffearium verification pass· Last source-checked 12 Oct 2026· Prepared with AI assistance· 15 sources (11 tier A)· 2 contested· Updated 20 Oct 2026
Reading depth
Contents
  • Roast degree says how far a roast has gone. Colour is its most widely used measure and, according to a roasting textbook chapter, the best indicator, but no universally accepted protocol exists.1,6
  • Colour is measured as reflected light, usually on ground coffee prepared in a standard way, as CIE L* lightness or on an instrument scale such as Agtron; lower numbers mean darker.1,3,5,6,9
  • The SCA defines its cupping roast as ground-coffee L* 26–29, about Agtron/SCA tile #65, or 63 on the Agtron “Gourmet” scale; other meters give other numbers for the same roast.3,4
  • Whole beans read darker than the same beans ground, probably because the outside of a bean roasts ahead of its core; the gap was largest in the lightest roast of one study.5
  • Names such as “light”, “medium” and “dark” are not standardised. In one study, the same colour was called medium, and also light for commodity and dark for specialty roasting.8

What roast degree means#

Roast degree describes how far a roast has gone. One study, for example, records it as the final bean colour, so that different time–temperature paths can be compared.8 It can be defined through bean weight loss, colour or selected chemical indicators.8 Sensory character is also used.6 Researchers note that a concise, universally accepted protocol for evaluating it does not exist,6 and a chemistry chapter says the same of its definition.11

In practice, roast degree is mostly judged by colour. Beans change from greenish grey-blue to yellow, orange, brown, dark brown and almost black.1 Colour development is closely tied to flavour development, which makes bean colour the best single indicator of roast degree.1 Colour is also the most widely used parameter in day-to-day practice, measured as the light reflected by ground beans or, still, by eye.6

Two coffees roasted to the same colour can still differ. Colour fixes where the roast ended, but not the time–temperature path that led there.8 One study therefore recommends reading colour together with development time (see Colour as a sufficient indicator).8

How colour is measured#

Roast-colour instruments shine light onto the coffee and measure the light reflected back.6 Two families of measurement are in use.

  • Colorimetry in CIE L*a*b*. This international colour space describes a colour by lightness (L*) and two colour coordinates, a* and b*, from which chroma and hue are derived.13 For roasted coffee, L* falls steadily as the roast gets darker, and, with hue, it is the most useful of the parameters for following roast colour.5 The SCA names CIELAB colorimetry as one way to control roast level.3
  • Roast meters with their own scales. The SCA also names infrared spectrometry, with the Agtron device as its example.3 In one study, a dedicated meter (Neuhaus Neotec Colortest II) measured the light reflected by 25–30 g of ground coffee at 900 nm, in the near infrared.6 Dedicated roast meters, many of them infrared, usually measure more accurately than other methods, but their scales are arbitrary.4

On the scales used in the studies cited here, a lower number means a darker roast: lower L*,5 lower Agtron value,2,9 lower Colortest II reading.6

Colour tiles. The SCA also refers to the Agtron/SCA Roast Color Classification System, a set of numbered reference discs (tiles).3,4 A Colombian study describes the Agtron/SCA chart as a roast-development pattern based on the caramelisation of sugar, and used it to compare samples by eye during roasting.7

Scales and why they do not convert#

The SCA’s cupping protocol gives one target roast and states it in several ways. It defines “cupping level roast” as ground-coffee L* 26–29, approximately tile #65 on the Agtron/SCA classification, or a reading of 63 on the Agtron “Gourmet” scale or its equivalent on other meters.3 For the same target, the SCA lists 48 on the Agtron “Commercial” scale, 96 on the Probat Colorette 3b and 62 on the Colortrack.4 Many other meters have adopted the Agtron “Gourmet” scale and should use the same target.4 The L* value is defined on ground coffee; for the meter targets the SCA does not state the sample form, but asks for grind level and coffee temperature to be controlled.3,4

These numbers are targets for one roast, not a conversion formula. The SCA notes that roast-meter scales are arbitrary, and that even meters using the same scale do not always correlate completely across brands.4 The SCA itself is researching how its roast-level parameters correlate, and expects more accurate guidelines.4

L* does not map onto meter numbers in a fixed way either. The SCA pairs L* 26–29 with tile #65;3 a Colombian study, roasting to 85, 55 and 35 on the Agtron/SCA scale, states that these correspond to L* 22.0, 18.5 and below 18.5, citing a Colombian roast-degree standard (ICONTEC 2442) and without saying whether the L* refers to whole beans or ground coffee.7 L* itself depends on the sample: in one study, the same Arabica roasts read up to 4.6 L* units apart as whole beans and as ground coffee.5 Relations between colour and other indices also depend on the coffee. In one study, chemical aroma indices tracked colour with r² of 0.88–0.89 in Arabica but 0.96–0.99 in Robusta and a 50:50 blend.6

Light, medium, dark: names are not standards#

Baristas often use names – “city roast”, “espresso roast”, “French roast” – for degrees of roast, while industrial operators and scientists prefer to measure.1 The examples below show how loosely the names are tied to numbers.

30405060708090100110120130Agtron number (lower = darker)Tile numbers and meter readings share this axis only approximately.Called ‘dark’357 · 45.79 · 768 Called ‘medium’557 768 Called ‘light’768 · 857 · 116.69 SCA cupping roast (meter)633 SCA bitter reference, darkroastabout 3510 Range in eight sensorystudies46–992 Very light specialtyroastsup to 1202
Data table
PointAgtron number (lower = darker)Source
Called ‘dark’35357
Called ‘dark’45.745.79
Called ‘dark’76768
Called ‘medium’55557
Called ‘medium’76768
Called ‘light’76768
Called ‘light’85857
Called ‘light’116.6116.69
SCA cupping roast (meter)63633
SCA bitter reference, dark roastabout 353510
Range in eight sensory studies46–9946–992
Very light specialty roastsup to 1201202
What published studies and SCA documents called dark, medium and light, against Agtron-scale numbers; one value (Agtron 76) carries all three names. Tile numbers and meter readings are close but not identical scales, so the axis is approximate and is not a conversion.2,3,7,8,9,10
  • One number, three names. A development-time study roasted to Agtron 76 ± 1 and called it a medium roast. The authors add that this degree counts as a light roast in commodity roasting and as a dark roast in specialty roasting.8
  • Specialty runs lighter. A multi-study paper covered Agtron “Gourmet” 46–99 (mean 79) and notes that very light roasts, often up to Agtron 120, are commonplace in specialty coffee; its range would be among the lightest in the commodity industry and among the darkest in specialty.2
  • Same name, different numbers. In a defect study, the “light” roast measured 116.6 and the “dark” 45.7, around a normal roast of 74.4.9 A Colombian study, matching samples by eye to the Agtron/SCA chart, called 85 light, 55 medium and 35 dark.7 The SCA uses dark-roasted coffee of about Agtron #35 as its reference for bitter taste.10

Whole beans and ground coffee#

A roast does not colour a bean evenly: a temperature gradient across the bean leaves the outside and the inside at different colours.5 Very fast roasting can over-roast the zone near the surface while the core stays under-roasted.1 Extremely fast roasts are traditionally treated as defects for this reason – a more pronounced gradient of roast degree from the surface to the centre.8 The SCA describes the same sign in sample roasts that are too fast: the surface looks much darker than the inside.4

A · Whole beanB · Cut openC · Ground12345
  1. Bean surface. A whole-bean reading sees only the outside, which roasts ahead of the core.1,5,8
  2. Outer layers. Darker; very fast roasts can over-roast this zone.1,4
  3. Core. Lighter while a gradient remains from surface to centre.5,8
  4. Ground sample. Grinding mixes outer and inner tissue. Meters are usually read on ground coffee prepared in a standard way.1,6
  5. Roast meter. Shines light onto the sample and measures the light reflected back.3,6
Why whole-bean and ground readings differ: a meter on a whole bean sees only its darker surface; ground, outer and inner tissue are mixed. Schematic, not to scale.5

A meter on whole beans therefore sees mostly the outside, while ground coffee mixes outside and inside. In a study of Arabica and Robusta at three roast intensities, whole beans had lower L* (were darker) than the same beans ground. The gap shrank as the roast got darker, which the authors suggest may come from the temperature gradient across the bean.5

3540455055L* (lower = darker)Filled dot: whole beans. Open dot: ground. One colorimeter, illuminant D65.T1 · 10.5% weight losswhole 45.2 · ground 49.85 T2 · 13.4% weight losswhole 40.0 · ground 41.95 T3 · 19.4% weight losswhole 37.1 · ground 37.45
Data table
PointL* (lower = darker)Source
T1 · 10.5% weight losswhole 45.245.25
T1 · 10.5% weight lossground 49.849.85
T2 · 13.4% weight losswhole 40.0405
T2 · 13.4% weight lossground 41.941.95
T3 · 19.4% weight losswhole 37.137.15
T3 · 19.4% weight lossground 37.437.45
Lightness of Arabica beans, whole and ground, at three roast intensities (T1–T3, labelled by their weight loss), measured with one colorimeter under illuminant D65.5

A coffee-review publication that reports a whole-bean and a ground reading for each coffee makes the same observation: the difference narrows in darker roasts and widens in lighter ones.5,15 Standards measure ground coffee: the SCA cupping roast is defined by the lightness of ground coffee.3

Reference values#

The values below come from different instruments and preparations. They are not interchangeable and are not targets except where a standard says so.

Parameter Value reported Source
SCA cupping roast, colorimetry Ground coffee L* 26–29, approximately Agtron/SCA tile #65 SCA 102-20243
SCA cupping roast, roast meters (targets, not a conversion; sample form not stated) Agtron “Gourmet” 63 (also for meters that adopted this scale); Agtron “Commercial” 48; Probat Colorette 3b 96; Colortrack 62 SCA 102-2024; CVA document3,4
Light, medium, dark in a Colombian study (100 Colombian coffees, sample roaster) Agtron/SCA scale (matched by eye) 85, 55, 35; L* 22.0, 18.5, below 18.5 as given by a Colombian standard, sample form not stated Gómez et al.7
Light, medium, dark on a Colortest II (900 nm, ground; the instrument’s own scale) 57–62, 46–53, 35–42 Liberto et al.6
Defect study (one Arabica) Light 116.6; normal 74.4; dark 45.7 (Agtron; instrument and sample form not stated) Giacalone et al.9
Development-time study (washed Colombian Arabica) Agtron 76 ± 1, “medium”; Javalytics meter, finely ground, three replicates Alstrup et al.8
Eight sensory studies (washed Arabica) Agtron “Gourmet” 46–99, mean 79 (Javalytics meter, three readings per sample); specialty roasts often up to 120 Münchow et al.2
SCA bitter taste reference Dark-roasted coffee, about Agtron #35 SCA 103-202410
Whole beans vs ground, Arabica, L* (D65) T1 45.2 vs 49.8; T2 40.0 vs 41.9; T3 37.1 vs 37.4 Bicho et al.5
Weight loss at T1, T2, T3 (Arabica; Robusta) 10.5, 13.4, 19.4%; 10.1, 12.5, 16.7% (200–240 °C, 5–12 min) Bicho et al.5
Correlation with flavour attributes (mean absolute r) Study 1: colour 0.40, development time 0.21; Study 2: colour 0.46, development time 0.07 (n.s.) Münchow et al.2
Chemical indices vs colour (r²) Arabica 0.88–0.89; 50:50 blend 0.99; Robusta 0.96–0.99 Liberto et al.6
Chlorogenic acids after roasting Nearly 50% left at 230 °C for 12 min; almost trace at 250 °C for 21 min Moon et al.12

Measuring well#

A colour reading depends on how the sample is prepared. The SCA asks for grind level and coffee temperature to be controlled when roast meters are used, and warns that samples with a lot of chaff show a larger deviation.3,4 For the most reliable results, beans are ground and prepared in a standard way before measuring.1

  • Grind. Research protocols state the grind, for example “finely ground” coffee measured in triplicate.8
  • Replicates. Studies average several readings per sample, often three.2,8
  • Instrument and settings. One study measured under two illuminants (D65 and C) and found that L* and hue followed similar patterns under both.5 Sources using different instruments and preparations give different L* values for roasts described the same way (see Contested).3,7
  • Visual checks. Judging colour by eye is still used.6 During roasting, one research group used a colour reference tool stated to be accurate to ±1 Agtron value, and a Colombian study compared samples with the Agtron/SCA chart.2,7

No study opened for this entry measured how readings change with the time since roasting, so no rule is given here.

The colour coordinates in roasting#

During roasting, L* falls with roast intensity, in whole beans and in ground coffee.5 The other coordinates behave less simply. In whole beans, a* (red) rose in the first roast stage and then fell, and b* (yellow) fell throughout. After grinding, only L* and hue fell steadily; a* in the two darker roasts stayed well above its value in green coffee, unlike in whole beans.5

Species can matter. At the same roast intensity, whole-bean colour parameters of Arabica and Robusta did not differ significantly in one study; ground green coffees of the two species did differ in several parameters.5

Colour, roast loss and chemistry#

Colour is not the only index of roast degree.

  • Roast loss. It rises steadily during roasting and is higher in dark roasts. As long as green-coffee quality and moisture stay constant, it can indicate roast degree, but batches roasted to the same colour can differ in roast loss when the green moisture varies.1
  • Organic roast loss. Calculated on dry matter, it takes the varying water loss into account. It correlates well with the lightness from colour measurement.1 Dry-matter loss is regarded as a reliable complement to colour, including in the plant.6 The SCA allows dry-matter loss or bean volume increase for cupping roasts once their correlation with colour or spectrometry has been established.3
  • Chemical markers. Roast degree can be followed through changes in chemical composition.6,8 Chlorogenic acids fall as roasting intensifies: in one study, to nearly half at 230 °C for 12 min and to almost trace levels at 250 °C for 21 min.12 Sucrose and free amino acids react early; arabinose side chains and chlorogenic acids are degraded later.11 Earlier studies, summarised in a development-time paper, found sucrose degraded by up to 99%, depending on the roast profile.8
  • Aroma fingerprints. A headspace mass-spectrometry method separated roast degrees in 80 samples of different varieties, origins and blends; ratios of diagnostic ions tracked colour, and the authors propose them for on-line monitoring.6

Colour as a sufficient indicator#

Colour is a strong predictor of flavour. A 2020 paper pooled eight descriptive sensory studies of washed Arabica. In the two that varied both, colour predicted flavour better than roasting time.2 The authors note that colour measurement is often left out of quality control in specialty roasteries, which could leave consumer-relevant batch-to-batch variation unchecked.2

Colour is not the whole story. At one fixed colour (Agtron 76 ± 1), changing only the development time shifted both the chemistry and the sensory profile; the authors conclude that colour alone is not a sufficient indicator, and recommend logging colour and development time together.8 A chemistry chapter likewise holds that roast loss and colour cannot describe the aroma profile on their own.11 The SCA adds that roasting problems such as too-fast or baked roasts have no simple measurement across roasting technologies; cuppers learn to detect them by bean structure, colour gradient and flavour.4

Decaffeinated coffee#

Decaffeination changes the beans’ structure and composition, so roasting must be adapted. A suitable target colour can differ from that of regular beans, and mass loss is sometimes the better control parameter.14 Decaffeinated green beans often look dull, yellowish or even brownish before roasting.14

Open questions#

  • Conversions. No published, general conversion links L*, Agtron and other meter scales; the SCA is researching the correlations.4
  • Natural and other processes. No source opened for this entry tested whether natural or other processed coffees read differently at the same roast degree. The main sensory evidence linking colour and flavour comes from washed Arabica.2
  • Time after roasting. How a reading changes in the hours and days after roasting was not addressed in the sources opened.
  • Light roasts. The multi-study evidence stops near Agtron 99, while specialty roasts often go lighter.2

Key studies#

Münchow et al., 2020 – Colour versus time. Pools eight descriptive sensory studies of washed Arabica, roasted on drum and convection roasters to Agtron 46–99. In the two studies that varied both, roast colour predicted flavour better than roasting time.

Beverages 6(2), 29 · doi:10.3390/beverages6020029

Bicho et al., 2012 – Colour parameters in roasting. Measures L*, a*, b*, chroma and hue of Arabica and Robusta, green and at three roast intensities, whole and ground. Lightness fell steadily in both forms and whole beans read darker than ground, with the gap shrinking as the roast darkened.

Ciência e Tecnologia de Alimentos 32(3), 436–442 · doi:10.1590/S0101-20612012005000068

Alstrup et al., 2020 – One colour, four roasts. Roasts one Colombian coffee to the same colour with development times of 90–390 s. Sensory panels, NMR and GC–MS still separate the roasts, so colour alone does not describe the coffee; the chosen colour is called medium, and light or dark depending on the market.

Beverages 6(4), 70 · doi:10.3390/beverages6040070

Liberto et al., 2013 – Roast degree from aroma. Analyses 80 coffees of different varieties, origins and blends, measured for colour by near-infrared reflectance on ground coffee. Mass-spectral fingerprints of the headspace separated roast degrees, and ion ratios correlated with colour more closely in Robusta and blends than in Arabica.

Journal of Agricultural and Food Chemistry 61(8), 1652–1660 · doi:10.1021/jf303067q

Contested
  • L* of a “medium” roast: the SCA defines its cupping roast as ground-coffee L* 26–29, about Agtron/SCA tile #653, while a Colombian study set its medium roast at 55 on the Agtron/SCA scale and gave it L* 18.5, citing a Colombian standard7; the literature values for “medium” that the SCA lists also vary by author.4 The sources use different instruments and sample preparation, so the numbers are not comparable, and no conversion is given here.
  • Colour as a sufficient indicator. In two sensory studies that varied both, colour predicted flavour better than roasting time2; at one fixed colour, changing only the development time still shifted chemistry and flavour, and the authors conclude that colour alone is not a sufficient indicator8; a chemistry chapter calls roast loss and colour insufficient to characterise roasted-coffee quality.11
Share or cite this entry

Change log

  • – The statement that particle size affects colour readings, which rested on a 2025 conference e-poster, was removed. Reason: Conference posters are not accepted as sources (editorial ruling, 20 Oct 2026).
  • – Colour is now "the best indicator" of roast degree according to a roasting textbook chapter (was "the best single indicator"). Reason: Cross-check during verification of the new browning-phase entry; the chapter (Schenker & Rothgeb 2017, p. 247) says "the best indicator".

Sources

  1. BSchenker, S.; Rothgeb, T. (2017). The Roast – Creating the Beans' Signature. In: Folmer, B. (ed.), The Craft and Science of Coffee, Academic Press, ch. 11, pp. 245–271. doi:10.1016/B978-0-12-803520-7.00011-6 · other copies
  2. AMünchow, M.; Alstrup, J.; Steen, I.; Giacalone, D. (2020). Roasting Conditions and Coffee Flavor: A Multi-Study Empirical Investigation. Beverages 6(2), 29. doi:10.3390/beverages6020029 · other copies
  3. ASpecialty Coffee Association (2024). SCA Standard 102-2024 Coffee Value Assessment: Sample Preparation and Tasting Mechanics. Specialty Coffee Association.
  4. ASpecialty Coffee Association (writing team Fernández-Alduenda, M.; Ionescu, K. E.; Giuliano, P.; Rugolo, J.; Delrue, R.) (2024). A System to Assess Coffee Value: Understanding the Specialty Coffee Association’s Coffee Value Assessment (June 2024). Specialty Coffee Association.
  5. ABicho, N. C.; Leitão, A. E.; Ramalho, J. C.; Lidon, F. C. (2012). Use of colour parameters for roasted coffee assessment. Ciência e Tecnologia de Alimentos 32(3), 436–442. doi:10.1590/S0101-20612012005000068 · other copies
  6. ALiberto, E.; Ruosi, M. R.; Cordero, C.; Rubiolo, P.; Bicchi, C.; Sgorbini, B. (2013). Non-separative headspace solid phase microextraction–mass spectrometry profile as a marker to monitor coffee roasting degree. Journal of Agricultural and Food Chemistry 61(8), 1652–1660. doi:10.1021/jf303067q · other copies
  7. AGómez Parra, C. R.; Gallego Agudelo, C. P.; Echeverri Giraldo, L. F.; Pabón Usaquén, J.; Ortiz, A.; Osorio Pérez, V. (2023). Determinación de compuestos químicos del café tostado por Espectroscopia de Infrarrojo Cercano (NIRS). Revista Cenicafé 74(1), e74104. doi:10.38141/10778/74104 · other copies
  8. AAlstrup, J.; Petersen, M. A.; Larsen, F. H.; Münchow, M. (2020). The Effect of Roast Development Time Modulations on the Sensory Profile and Chemical Composition of the Coffee Brew as Measured by NMR and DHS-GC–MS. Beverages 6(4), 70. doi:10.3390/beverages6040070 · other copies
  9. AGiacalone, D.; Degn, T. K.; Yang, N.; Liu, C.; Fisk, I.; Münchow, M. (2019). Common roasting defects in coffee: Aroma composition, sensory characterization and consumer perception. Food Quality and Preference 71, 463–474. doi:10.1016/j.foodqual.2018.03.009 · other copies
  10. ASpecialty Coffee Association (2024). SCA Standard 103-2024 Coffee Value Assessment: Descriptive Assessment. Specialty Coffee Association.
  11. BPoisson, L.; Blank, I.; Dunkel, A.; Hofmann, T. (2017). The Chemistry of Roasting – Decoding Flavor Formation. In: Folmer, B. (ed.), The Craft and Science of Coffee, Academic Press, ch. 12, pp. 273–309. doi:10.1016/B978-0-12-803520-7.00012-8 · other copies
  12. AMoon, J.-K.; Yoo, H. S.; Shibamoto, T. (2009). Role of Roasting Conditions in the Level of Chlorogenic Acid Content in Coffee Beans: Correlation with Coffee Acidity. Journal of Agricultural and Food Chemistry 57(12), 5365–5369. doi:10.1021/jf900012b · other copies
  13. AInternational Organization for Standardization; International Commission on Illumination (2019). ISO/CIE 11664-4:2019 Colorimetry – Part 4: CIE 1,976 L*a*b* colour space. ISO, Geneva.
  14. BPietsch, A. (2017). Decaffeination – Process and Quality. In: Folmer, B. (ed.), The Craft and Science of Coffee, Academic Press, ch. 10, pp. 225–243. doi:10.1016/B978-0-12-803520-7.00010-4 · other copies
  15. CStachura, T. (2010). Agtron Numerology. Coffee Review.

Tier A · authoritative · B · trade and expert · C · general. Supporting quotes for every source are in the data version of this entry. If a publisher’s site refuses access, try the doi link or “other copies”.

Suggest a change

On Roast degree and colour. Tell us what you would change and why. A source that supports it helps most.