Taming Bitterness and Off-Notes With Flavour Masking

Flavour masking works on three separate levers, and choosing the wrong one is why some briefs stall. Bitterness is detected by roughly twenty-five to twenty-six human TAS2R receptors, and because just three of them account for about half of the bitter compounds tested, a masking system has to be matched to the specific bitterant rather than bought as a general solution. The second lever is suppression rather than blocking, and the third is removing the off-note at source before a masking system is asked to cover it.

Twenty-Five Receptors, Thousands of Bitter Compounds

Start with the receptor arithmetic, because it explains the matching problem. Meyerhof and colleagues, writing in Chemical Senses 35(2):157-70 in 2010, set out the mismatch: “Humans perceive thousands of compounds as bitter. In sharp contrast, only approximately 25 taste 2 receptors (TAS2R) bitter taste receptors have been identified” They then measured it. The team reports that “we have challenged 25 human taste 2 receptors (hTAS2Rs) with 104 natural or synthetic bitter chemicals in a heterologous expression system”, and that “Thus, 3 hTAS2Rs together were able to detect approximately 50% of the substances used.” The count has since been revised upward: the 2024 BitterDB update in Nucleic Acids Research reports that “human TAS2R2, which until recently was considered a pseudogene, was added as a functional variant for which 8 ligands were recently found” and that “We recalculated the multiple sequence alignment (MSA) for all 26 human TAS2Rs.”

Why One Blocker Will Not Fix a Whole Formula

Two further findings from the same 2010 screen decide how a brief should be scoped. First, “though 63 bitter substances activated only 1-3 receptors, 19 compounds stimulated up to 15 hTAS2Rs.” A system tuned to one receptor will therefore fix some bitterants outright and barely touch others. Second, receptor assignment is still sparse: BitterDB reports that “For ∼700 molecules, at least one associated bitter taste receptor (TAS2R) is reported. The overall number of ligand-TAS2R associations is now close to 1800.” For most bitter molecules there is no published receptor to target, so the route is sensory work in the real base rather than a receptor-level claim.

Steviol Glycosides: the One Complete Mapping

Sweetener bitterness is the one case where the mapping is complete and usable. Hellfritsch and colleagues, in the Journal of Agricultural and Food Chemistry in 2012, report that “A comprehensive screening of 25 human bitter taste receptors revealed that two receptors, hTAS2R4 and hTAS2R14, mediate the bitter off-taste of steviol glycosides.” A 2026 review in Frontiers in Nutrition sets out the practical consequence across the glycosides: “Major SGs, such as STV and Reb A, are commonly associated with pronounced bitterness, licorice-like aftertaste, and prolonged sweetness lingering” while “In contrast, minor SGs, particularly rebaudioside D (Reb D) and rebaudioside M (Reb M), generally show a faster onset of sweetness, lower bitterness, shorter aftertaste, and a more sucrose-like flavor profile.” The same 2012 paper also explains why raising the dose backfires: “For some test substances, e.g., stevioside, we observed a decline in sweet intensity at supra-maximum concentrations. This effect did not arise from allosteric modulation of the hTAS1R2/R3 sweet taste receptor but might be explained by intramolecular cross-modal suppression between the sweet and bitter taste component of steviol glycosides.” Overdosing a stevia system to chase sweetness makes the problem worse, not better, which is why sweetener bitterness is handled inside the wider sugar reduction brief rather than after it.

Blocking Is Real, and Narrow

Blocking at the receptor is real but narrow. Slack and colleagues, in Current Biology 20(12):1104-9 in 2010, describe “a novel bitter receptor antagonist (GIV3727) that inhibits activation of hTAS2R31 (formerly hTAS2R44) by saccharin and acesulfame K”, note that it also inhibited five further receptors, and report that “In human sensory trials, GIV3727 significantly reduced the bitterness associated with the two sulfonamide sweeteners.” The sensory result was reported for those two sweeteners. A different route, from Ming and colleagues in PNAS in 1999, is adenosine monophosphate: “AMP and chemically related compounds inhibited in vitro responses to several bitter compounds (e.g., denatonium, quinine, strychnine, and atropine). AMP also inhibited behavioral and electrophysiological responses of mice to bitter tastants, but not to NaCl, HCl, or sucrose. GMP… inhibited neither.” Selectivity is the point and also the limitation.

Suppression: What Sodium Does to Bitter Amino Acids

Suppression is the lever that most protein and hydrolysate formulas actually respond to. Harmon, Ahmed and Breslin, in the Journal of Agricultural and Food Chemistry in 2024, report that “We found that the essential amino acids comprise the most bitter stimuli, with six of them conveying the most bitterness” and that “We also show that bitterness can be largely suppressed by sodium salts for 5 of the 6 most bitter amino acids.” A formula carrying free amino acids or a hydrolysate will often move further on a sodium adjustment than on an aromatic addition, which is not what a flavour brief usually assumes, and it is the same balancing that sits behind salt reduction and fat reduction.

Alcohol Burn and Astringency Are Not Taste

Two off-notes are not taste at all, and treating them as taste wastes rounds. Alcohol burn is a trigeminal heat response: Trevisani and colleagues, in Nature Neuroscience 5(6):546-51 in 2002, report that “Ethanol potentiated the response of VR1 to capsaicin, protons and heat and lowered the threshold for heat activation of VR1 from approximately 42 degrees C to approximately 34 degrees C.” Thirty-four degrees is below body temperature, which is why a spirit-forward drink reads as heat rather than as flavour. Astringency is polyphenols binding salivary proteins: a 2026 study in Food Chemistry 502:147605 reports that “the binding energies of the polyphenols with MUC were generally lower than those with bPRPs, following the order of binding affinity TA > PC > EGCG > C > GA.” Neither responds to a bitter blocker, because neither runs through a bitter receptor.

Removing the Off-Note at Source

The third lever is to remove the off-note upstream rather than cover it, and in plant protein this is now measured work. A 2026 paper in Current Research in Food Science identifies hexanal as “the primary aldehyde formed from linoleic acid oxidation and a key driver of beany, grassy, and green off-flavors due to its low odor threshold”, and reports that “catechins at 500 ppm effectively reduced the content of 1-pentanol by 3.5 times, hexanal by 6 times, and hexanol by 1.3 times, compared to the control sample.” Across the dose range, “the concentration of hexanal decreased from 6 to 1.2 μg g −1 DW, with catechins concentrations ranging from 250 to 2000 ppm” This is an extraction-stage intervention, not in-product masking: the compound is prevented from forming rather than hidden after the fact. A companion paper on enzymatic hydrolysis reports that after Alcalase treatment “five compounds responsible for the undesirable notes (hexanal, 2-ethyl-1-hexanol, 1,3-di-tert-butylbenzene, undecane, and 2-pentylfuran) showed a significant decrement in odor intensity.”

Why Iron Reads as Metallic

Fortification off-notes have their own mechanism, and it is often nasal rather than oral. Epke, McClure and Lawless, in Food Quality and Preference in 2009, report that “Metallic sensations were decreased by nasal occlusion, which in the absence of any orthonasal metallic smell, implies that the sensations were retronasally perceived volatiles in the nose open condition”, consistent with metal salts generating volatile lipid-oxidation products in the mouth that are perceived retronasally. That matters practically: an aroma-side fix can work on a metallic note where a taste blocker cannot.

The Measured Threshold in Salt Replacement

Salt replacement has one well-measured threshold. A 2022 study in the International Journal of Gastronomy and Food Science, working in bread against a two per cent sodium chloride control, reports that “the proportion equivalent to difference threshold was obtained, defined as the KCl percentage to 75% of correct responses, which corresponded to 0.92% of KCl.” In other words, KCl becomes detectable as bitter or metallic at 0.92 per cent within the same two per cent total salt. The method matters as much as the number: the trial was “carried out according to method E 679-04, ASTM (1997) with ninety bread-usual-consumer volunteers”, using a two-alternative forced choice against a 75 per cent correct-response criterion. The threshold applies to a bread matrix.

No Regulator Has a Masking Category

Now the label, where a common industry word has no regulatory home. Search Codex CAC/GL 66-2008, Regulation (EC) No 1334/2008, FSANZ Schedule 14 and Standards 1.3.1 and 1.2.4, and the Singapore Food Regulations, and the word “masking” appears in none of them. What regulators recognise are flavourings and flavour enhancers. Codex defines the first as products added to food “Flavourings are products that are added to food to impart, modify, or enhance the flavour of food”, and excludes the purely tasting substances: “Flavourings do not include substances that have an exclusively sweet, sour, or salty taste (e.g. sugar, vinegar, and table salt).” The EU turns on the same verbs at Article 3(2)(a): “'flavourings' shall mean products: (i) not intended to be consumed as such, which are added to food in order to impart or modify odour and/or taste”

Flavouring, or Flavour Enhancer and Modifier

The enhancer side is where the word “modifier” actually lives. FSANZ Schedule 14 lists the sub-classes and the definition together: “flavour enhancer, flavour modifier, tenderiser enhances the existing taste or odour of a food” Its flavouring entry is drafted differently, as “intense preparations which are added to foods to impart taste or odour, which are used in small amounts and are not intended to be consumed alone, but do not include herbs, spices and substances which have an exclusively sweet, sour or salt taste” Singapore splits the same way: regulation 22(1) provides that “In these Regulations, “flavouring agent” means any wholesome substance that when added or applied to food is capable of imparting taste or odour, or both, to a food.”, while regulation 23(1) provides that “'flavour enhancer' means any substance which is capable of enhancing or improving the flavour of food, but does not include any sauce, gravy, gravy mix, soup mix, spice or condiment.” So a masking system is described as a flavouring, or as a flavour enhancer or modifier, and the right word is the one matching the function the regulator defines.

What Actually Goes on the Ingredient List

The declaration then follows from that classification. In Australia and New Zealand, section 7(4) of Standard 1.2.4 provides that “If a *flavouring substance is an ingredient, it must be listed in the statement of ingredients by using: (a) the word 'flavouring' or 'flavour'; or (b) a more specific name or description of the flavouring substance.”, and Standard 1.2.4 separately provides that a statement of ingredients need not list an ingredient of a flavouring substance. That exemption has named carve-outs: section 7(5) of Standard 1.2.4 requires that “If any of the following substances are added to a food for sale as a *flavouring substance or as an ingredient of a flavouring substance, the name of the substance must be specifically declared”, listing nine glutamate and ribonucleotide substances, and section 7(6) of the same Standard adds that “If caffeine is added to a food for sale (whether as a *flavouring substance or otherwise), it must be listed in the statement of ingredients as caffeine.” Telling a customer that a masking system all declares as flavouring is wrong the moment it carries any of those.

Two Limits in Every Brief

Two limits belong in every brief. The first concerns evidence: ask any supplier quoting a percentage bitterness reduction for the panel size, the method and the matrix. The second is regulatory and sits in Codex, which permits flavourings only where “The use of flavourings is justified only where they impart or modify flavour to food, provided that such use does not mislead the consumer about the nature or quality of food.” and requires that “Flavourings should be used under conditions of good manufacturing practice, which includes limiting the quantity used in food to the lowest level necessary to accomplish the desired flavouring effect.” A claim that a masking system makes a high-salt or high-sugar product taste like a reduced one invites exactly the objection that clause was written for.

Receptor Antagonism, Measured in Patients

Three 2026 findings add mechanisms. The first is receptor antagonism by a compound that is already an approved flavouring. A 2026 study in Scientific Reports reports that a rinse-and-spit solution containing sodium homoeriodictyol, “an approved flavoring compound and antagonist of bitter taste receptors (TAS2Rs), reduced the perceived bitterness of caffeine in patients receiving carboplatin ( n = 15)”. Fifteen participants, a clinical population, and the authors' own word for the work is preliminary.

Umami Suppression, and a New EU Modifier

The second is suppression by umami rather than blocking. A 2026 review in Sensors records that representative umami compounds including monosodium glutamate and umami peptides “suppress the bitterness of several bitter APIs, suggesting potential applicability beyond compound-specific taste-masking strategies”, operating at receptor level and centrally rather than by covering an aroma. The review covers pharmaceutical applications; confirm the effect in your own matrix. The third is regulatory: on 26 January 2026 the European Union authorised hesperetin dihydrochalcone, FL No 16.137, as a flavouring substance at not more than 10 mg/Kg in each of fourteen food categories, a dihydrochalcone whose function is sweetness and mouthfeel modification. Its close relative neohesperidin dihydrochalcone was given a temporary acceptable daily intake of 0 to 3.8 mg/kg body weight per day by JECFA in June 2026, with measured exposure of 1.8 mg/kg below it and no safety concern flagged at the proposed levels. Neither is a Singapore permission, and the EU listing tells you the substance is authorised there, not that it is available to you here.

How to Brief a Flavour House

A masking brief moves faster when it carries five things. Name the off-note precisely, bitter, metallic, astringent, sulphurous, earthy, or a lingering aftertaste, because the descriptor points to a different lever. Send the real base at the real dose, because the flavourist has to taste the off-note in the matrix. State the format and the processing, since heat, pH and shelf life all change what holds. State the label and certification constraints, natural, allergen-free, halal or kosher. Name the destination markets, so the sweetness and aromatic direction can be tuned to them.

How VKA Approaches Masking

At VKA, masking is a core part of how we develop flavours for difficult bases. We combine taste modulation and bitter-blocking with aromatic masking and sweetness balancing, and where the off-note is best handled physically we use EssenceLock, our encapsulation technology, to keep bitter or metallic ingredients away from the palate until after they are swallowed. For powder products the same work pairs with TasteGuard, our preservation system, so the masked profile holds through storage and transit rather than drifting before it reaches your line. This work runs across the categories where off-notes bite hardest, including plant-based products, nutraceuticals and supplements and the wider set of food and beverage solutions we develop for. If you have a base with an off-note you cannot remove, talk to a flavourist directly and bring a sample of the real thing.