FR2980111A1 - METHOD AND DEVICE FOR REPAIRING AN ODORANT COMPOSITION - Google Patents

Abstract

The invention relates to a process for the preparation of a composition which emits an odor, in particular an aroma, from a target odor o and from a set of bases (B to B) in which the target odor and the bases are mixed in appropriate weight amounts to obtain the odorant composition. This method has the particularity that the determination of the respective weight amounts of the bases (B to B) to be mixed is carried out according to the following steps: a) after having smelled the target odor 0, an olfactory intensity is assigned (o to o ) to each of the odor descriptors of a predetermined set of odor descriptors (D to D) and b) from the attributed olfactory intensities (o to o), the respective weight quantities (q to q) are calculated; bases to be mixed to obtain the odorant composition. The invention also relates to a device for the aid in the preparation of an odorant composition and to a device for the preparation of an odorant composition.

Description

The invention relates to a process for the preparation of a composition from which an odor, in particular an aroma, emanates. The invention also relates to a device for the aid in the preparation of an odorant composition and to a device for the preparation of an odorant composition.

BACKGROUND OF THE INVENTION Generally, to create an aroma, the flavourist searches in his palette of starting molecules those which he thinks they will bring the desired odorous notes, then he mixes them to obtain a first version. The aroma is tested then, being very rarely satisfactory in its first version, it is reworked, sometimes a lot of times, until obtaining the desired notes. Japanese patent application JP 2004325116 discloses a process for the preparation of a perfume in which a target perfume is qualitatively analyzed by gas chromatography to determine its volatile components. A dilution factor is calculated for each volatile component and multiplied by a threshold value. The perfume is then prepared by mixing the volatile components on the basis of the product of the dilution factor and the threshold value. The disadvantage of such a method is that it does not take into account the interactions occurring between the compounds and which are generally important. These interactions can lead to synergies or antagonisms that are totally impossible to predict, so that the product must be reformulated, usually several times, before being satisfactory. This requires a lot of time and effort. SUMMARY OF THE INVENTION The essential object of the invention is to propose a process making it possible to obtain easily and quickly a desired odor, such as a flavor. This method therefore relates to the preparation of an odorant composition from a target odor and a set of bases, in which the target odor is inhaled and the bases are mixed in appropriate quantities to obtain the odorant composition optimal, that is to say the one whose smell is very close to, or even indistinguishable from, the target odor.

This method has the particularity that the determination of the respective quantities of the bases to be mixed is carried out according to the following steps: a) after having smelled the target odor, an odor intensity is attributed to each of the odor descriptors of a 20 predetermined set of odor descriptors and b) from the attributed olfactory intensities, the respective weight amounts of the bases to be mixed are calculated to obtain the odor composition. Such a process thus makes it possible to arrive quickly at a composition emitting the desired odor, since it takes into account the hitherto unpredictable interactions between the various constituents of the composition. According to a variant of the process according to the invention, the starting point is not a substance exhaling the target odor, but a simple indication or description of the odor that it is desired to obtain.

According to another aspect, the invention relates to a device for assisting in the preparation of an odorant composition, comprising: means for capturing olfactory intensities corresponding to odor descriptors; computer means comprising at least one memory and calculation means, said memory comprising a program able initially to create aromatization functions from the olfactory intensities seized, then to apply these aromatization functions to olfactory intensities seized subsequently, so as to obtain quantities by weight, and display means of the weight quantities obtained. According to yet another aspect, the invention also provides a device for the preparation of an odorant composition, comprising: - an apparatus for the qualitative and quantitative analysis of a substance, this apparatus being able to determine olfactory intensities corresponding to odor descriptors, - computer means connected to the qualitative and quantitative analysis apparatus, comprising at least one memory and calculation means, said memory comprising a program able to initially create, from the intensities the odor analysis functions, then to apply these aromatization functions to olfactory intensities subsequently provided by the qualitative and quantitative analysis apparatus, so as to obtain quantities, and - a dosing and mixing device, connected to the computer means for receiving the quantities erals obtained and, from these, assay and mix bases to obtain a fragrant composition.

Other features and advantages of the invention will now be described in detail in the following description. DETAILED DESCRIPTION OF THE INVENTION The implementation of the process according to the invention requires the prior establishment of aromatisation functions obtained by means of primary equations, secondary equations and tertiary equations. The objective is to obtain, starting from n starting values (odorous intensities of the odor descriptors), j values (weight quantities of the bases): (ol to on) -> aromatisation functions ( qi to qi) I) Establishment of aromatization functions a) Primary equations Basics We begin by choosing j bases (B1 to Bi). This choice can be arbitrary. However, bases are preferably chosen which each have a large number of constituents which, while belonging to the same family as a whole, exhale a variety of odors. The overall smell of each base is preferably the most complex possible, but it is preferably desired that it has a certain harmony. For example, a fruity base can collect more than thirty natural extracts of aromatic molecules having a smell of fruits whose smells are very different from each other.

Similarly, a chocolate base may contain extracts of chocolate, caramel, vanilla, raw materials in the form of powder or liquid having a smell of caramel, honey, etc.

A spicy base can include many and varied spices. A floral base can be composed of a mixture of a large number of flower extracts. A butyric base may contain materials having the odor of rancid butter, rotten fish, plastic, phenol, with the addition of butyric acid acetic acid, and the like. The constituents of the bases can be natural, based on natural, synthetic or artificial compounds. The bases may be liquid, gaseous or in powder form, preferably in liquid form. They can also be applied on an olfactory neutral support. The mixtures Then, mixtures (M1 to Mm) of these bases (B1 to Bj) are prepared. Thus, we obtain the following m primary equations connecting each mixture Mk (k varying from 1 to m) to the bases which compose it: M1 = b11B1 + b21B2 + b31B3 + ... + bilBi M2 = b12B1 + b2232 + b32B3 + ... M 2 = b13B1 + b23B2 + b33B3 + ... ^ bi3Bi 30 Mm = bimBi + b2mB2 + b3mB3 + ... + bjmBj in which b11, b21, bjm are weight quantities, and which can be generalized by the equation following: Mk = bikBi + b2kB2 + b3kB3 + ... + bikBi k varying from 1 to m. For example, M1 is a mixture of bll grams of base B1 + b21 grams base B2 + .... b31 grams of base B3, etc. + bilograms of base B. b) Secondary equations Parallel or later, odor descriptors (D1 to Dfl) are defined as qualifiers attributed to olfactory perception.

For example, the "banana" descriptor means that the odor is that of a banana, "grilled" it evokes a grilled food, "plastic", it is reminiscent of a plastic object, etc. Other examples of odor descriptors include "strawberry", "rancid cheese", "phenol", "coconut", "butter", "acetic", "citrus", etc. The number n of odor descriptors is generally greater than the number j of bases. Then, each of the mixtures (M1 to Mm) is inhaled by a nose 20 which assigns a note or odor intensity to each of the odor descriptors. Thus, for example, the M2 mixture can give off a strong strawberry scent (olfactory intensity of 8/10), a slight banana scent (olfactory intensity of 2/10), no butter scent (olfactory intensity of 0/10 ), etc. We can then write the following m secondary equations linking each mixture to odor descriptors and olfactory intensities: M1 F-d11D1; d21D2; d31D3; ; dnI Dn 30 M2 = d12D1; d22D2; d32D3; ; dn2 Dn M3 E d13D1; d23D2; d33D3; ; dn3 Dn dimDi; d2mD2; d3mD3; ; dnmDn dim are intensities in which d11, d21, olfactory generally expressed in integers ranging from 0 to 10, and that can be generalized by the following equation: Mk E dl kD1; d2 kD 2; d3kD3; - - -; dnkDn k varies from 1 to m It is advantageous to provide, among the m mixtures, mixtures which each contain only one of the bases. It is important to note that the earlier the initial selection of each base covers a number of high odor descriptors, the more these secondary equations are refined. Preferably, for the notation or attribution of an olfactory intensity to each odor descriptor in a mixture, a panel composed of several noses is used. Each panellist then notes each of the mixtures and the final olfactory intensities d11, d21, and dim for each mixture are averages of the olfactory intensities determined by all panellists. It is therefore real numbers. C) Tertiary equations By putting side by side the m primary equations and the m secondary equations, we arrive at: M1 = b11B1 + b21132 + b31B3 + ... + bilBi = d11D1; d21D2; d31D3; ; dniDn M2 = b12B1 b22B2 b32B3 bi2Bi E d12D1; d22D2; d32D3; ; Mm = bimBi + b2mB2 + b3mB3 + ... + bjmBj = dimDi; d2mD2; d3mD3; dflmD "or, generalizing: Mk = blkBl b2kB2 b3kB3 bikB; d2kD2; d3kD3; ---; dnkDn k varying from 1 to m. 30 From this system, we can generate by stochastic modeling the following n tertiary equations connecting the odor descriptors at the bases: D1 = fl (B1, B2, B3, ..., Bi) D2 = f2 (B1, B2, B3, ..., Bj D3 = f3 (B1, B2, B3, Bi) Dn = fn Bj) that can be generalized by the following equation: Dp = fp (331, B2, B3, ..., 13j) p varying from 1 to n The n tertiary equations obtained each thus each connect a descriptor (D1, D2 ... or Dn) to all of the bases B1 to B. d) Target odor O It is possible to relate the target odor 0 to the respective weight quantities q1 to qj bases B1 to Bj which will have to to be mixed to obtain the optimal composition, that is to say the one whose odor will come closest to the target odor O, as follows: 0 E g1B1; q2B2; q3B3;; qiBi Of course the quantities weight ql to qj vary of an odor ci ble O to another.

Concretely, the optimal odor composition will be obtained by mixing: - engrams of the base B1 - q2 grams of the base 32 - q3 grams of the base B3 25 qj grams of the base B. e) Determination of the quantities qi to qj To determine the weight amounts qi to qj of the bases B1 to Bj to be mixed, the target odor 0 will be decomposed into each of the descriptors D1 to Dn, attributing to these descriptors the corresponding olfactory intensities 01 to on, which can be formulated as follows: 0 E o1D1; o2D2; o3D3; ; OnDn The values ol to be used will then be used as the basis for the 5 calculations of the weight quantities qi to qi ol at on q1 to cl; The calculations will be carried out by means of aromatization functions a1 to ai: = (01,, on) 10 q2 = a2 (01,, on) qj = aj (01, -, 0.) What we can generalize by the following equation: = a. (01, -, 0n) x varying from 1 to j. F) Aromatization Functions The aromatization functions ax are previously established once and for all from bases B1 to Bi, mixtures of bases M1 to Mm, descriptors D1 to Dn initially selected, perceived olfactory intensities. d11 to dnm, primary, secondary and tertiary equations and by stochastic modeling. They will, therefore, provide, from the olfactory intensities ol to on of the target odor 0, the weight amounts q1 to qi as follows: (ol to on) j aromatization functions a. (q1-cli) II) Use of the preset flavoring functions Thus, once the aromatization functions ax established in the manner indicated above, it is sufficient, to implement the invention, to determine the olfactory intensities 01 to 0, descriptors of the target odor O, and then provide them to the functions of aromatization, to obtain, as indicated above, the j quantities by weight ch to qj (0, on) H> j flavoring functions a. -3 (q1 to cii) These weight quantities cil to qj which can therefore easily be calculated using the preset flavoring functions, allow to prepare the optimal composition, that is to say the one whose smell is close to the more than the target odor O. In other words, the only determination of the olfactory intensities 01 to be perceived for the predefined odor descriptors D1 to Dn of the target odor 0, makes it possible, by using the aromatization functions ax, d easily obtain the respective weight amounts of the bases to be mixed to obtain the optimum odor composition.

In summary, in order to prepare an odorant composition, it is sufficient, according to the invention, to determine the n olfactory intensities corresponding to the n odor descriptors of the target odor 0, and then to provide these n olfactory intensities to the flavoring functions. so that they calculate the weight amounts q1 to qj from which the odorant composition can be formulated and prepared by mixing the bases. III) Inverse aromatization functions The inverse aromatization functions serve, from a mixture of the bases 131 to Bj in respective quantities q1 to qj, to obtain or predict the respective olfactory intensities of the descriptors D1 to Dn of the odor of the mixture, which can be summarized as follows: (q1 to cli) n inverse aromatization functions ip- ÷ (ol to on) These inverse aromatization functions ip, were, like the functions of aromatization ax, previously established once and for all from the bases B1 to Bj, mixtures of bases M1 to Mm, descriptors D1 to Dn initially selected, perceived olfactory intensities d11 to d., primary, secondary and tertiary equations and by stochastic modeling.

These n inverse aromatization functions il, i2 1-3r --- r, thus connect the olfactory intensities ol to on to the weight quantities q1 to qj of the mixture as follows: oi =, q2, q3 ... r) 02 = i2 q2 C13 .-- q3) 03 = i3 (qi, q2 r q3 ... r qj) On = in (qi q2 q3 ... ci;) which can be generalized by the following equation : Op = ip q2 q3 ... qi) p varying from 1 to n IV) Use of inverse aromatization functions The inverse aromatization functions can be used for the prediction of the olfactory spectrum of an odor composition composed of a mixture of bases (B1 to Bi) in respective amounts by weight (q1 to qi) as follows: i) from the weight quantities (qi to qi), one calculates olfactory intensities (ol to on) and (Ii) from the calculated olfactory intensities (o1 to n o), the expected olfactory spectrum for the odor composition composed of the mixture of bases (B1 to Bi) in the quantities weights (q1 to qi). The calculation of step i) is carried out by applying each of the inverse aromatization functions i1 to in to the set of weight quantities q1 to qi.

IV) Variant According to one variant, the starting point of the process is not a substance exhaling the target odor, but a simple indication or description of the odor that it is desired to obtain. In this case, we only start from the set of bases and from the given indication. From the latter, the olfactory intensity (s) can be directly assigned in a first step a ') of the variant of the process. For example, if the starting indication is "a strong aroma of vanilla slightly" apple "", we attribute directly to the "vanilla" descriptor a strong olfactory intensity, for example 8/10, and to the descriptor "apple" a low olfactory intensity, for example 2/10. For other odor descriptors, a zero odor intensity is chosen. Then, proceed to step b) (of the variant) of the process according to the invention and, from the weight quantities cil to qi, the odorant composition is prepared. IV) Devices according to the invention The process according to the invention generally uses, for its implementation, computer means, comprising, in particular, a memory for containing the aromatisation functions, calculation means for initially establishing the aromatization functions fp (B1, B2, B3,..., Bi) and subsequently apply them to the olfactory intensities ol to be determined for the target odor O, in order to obtain the weight quantities cil to qi . These calculating means may use an appropriate statistical program to initially establish the aromatization library, and then calculate the weight amounts qi to qi for each preparation of an odorant composition. Olfactory intensities can be determined by an operator after the latter has inhaled the target odor and then entered on a keyboard connected to the computer means. They can also be inhaled and determined by a known apparatus (chromatograph, mass spectrograph, etc.) coupled with computer means. The values obtained qi to qi can be displayed on a display device such as a screen. The operator can then use it to weigh and mix the bases. They can also be transmitted to a device suitable for metering and mixing the bases. There are commercially available such devices that are used in particular in the field of perfume manufacturing. Mention may in particular be made of those marketed by the Swiss company Contexa SA or those described in US Pat. No. 4,333,356. V) Uses The odorant composition according to the invention can be used as a flavoring for animal or human food . It can be used as is, or in admixture with other ingredients. Advantageously, if it consists of bases known as "organic" and possibly other "organic" ingredients, that is to say from organic farming, it can itself benefit from this designation. It can also be used for the preparation of personal perfumes, domestic or industrial, biological or not.

VI) Examples A fruity base, a spicy base and a floral base, have been prepared from mixtures of about thirty of respective plant extracts among which are black pepper essence, root lovage, etc. . and synthetic aromatic molecules. A chocolate base was prepared from a mixture of powders having a chocolate, caramel, vanilla odor to which was added and mixed with molecules having odors approximating that of honey. A butyric base has been prepared from various natural extract acids to which molecules with a milder odor have been added. From these 5 bases, 41 different mixtures were prepared, of which 5 consisted solely of each of the bases. A list of 28 odor descriptors was then defined, including the following odor descriptors: acetic, citrus, banana, butter, biscuit, butyric, cinnamon, caramel, chocolate, coconut, spicy, fenugreek, floral, strawberry, rancid cheese, caramelized fruit, water fruit, red fruit, toasted, milk, honey, blackberry, peach, phenol, plastic, apple, vanilla and (fruit or vegetable) green. Then, the 41 mixtures were presented to a panel composed of 11 noses. Each of these noses or panelists had to smell one after the other each mix and assign, for each mixture, a note to each of the odor descriptors. From these numerous data (41 mixtures x 5 30 base coefficients, 28 odor descriptors x 11 panellists), for each mixture, descriptor after descriptor was then carried out, an average of the scores given by the panelists, reducing thus the number of data at 41 x 5 x 28 x 1. The data were then processed by scholastic modeling to yield tertiary equations linking each descriptor Dp to the bases A, B, C, D and E. In the present invention, used stochastic modeling because this approach assumes that there are functional relationships between the variables. However, any other appropriate method could be used as well.

The following equations are then obtained: Dp = fp (A, B, C, D, E) p varying from 1 to 28. For example, for the vanilla descriptor, the following equation, which relates the descriptor of "vanilla" odor at 5 bases A, B, C, D, E: vanilla = f-vanilla (A, B, C, D, E) = 1.84 [A] + 4.84 [B] + 2, 34 [C] + 3.54 [D] + 5.11 [E] -2.42 [A x C] + 2.98 [B x D] -17.12 [B x E] -3.26 [ C x E] -9.24 [D x E] +8.59 [A x B x (AB)] +6.28 [A x C x (AC)] +7.75 [A x D x (AD) )] +13.19 [A x E x (AE)] -24.16 [B x D x (BD)] +22.4 [B x E x (BE)] +9.76 [C x D x (CD)] -7.2 [C x E x (CE)] -12.35 [D x E x (D-E)] +179.07 [A x C2 x D] -123.12 [A x C x D2] -60.87 [A x C x E2] - 31.38 [A x D2 x E] -46.28 [B x C x E2] -59.02 [B x D x E2] -104 , [C x D x E2] -10.53 [A x B x (AB) 2] -87.98 [B x C x (BC) 2] -80.14 [B x D x (BD) 2 ] +68.57 [B x E x (BE) 2] -3698.35 [A x C x D x E] + 4664.65 [B xCxDxE].

From the 28 tertiary equations obtained, 5 relatively complex aromatization functions were then obtained. Internal Validation of Flavoring Functions To validate the aromatization functions obtained, that is, to verify that they are accurate, they were tested with all 28 odor descriptors.

We began by validating each tertiary equation for the 28 odor descriptors in order to verify whether the tertiary equation gives the same results as the panelists. The 41 mixtures analyzed by the panel were taken and the correlation between the results given by the tertiary equation and the results of the panelists was validated by linear regression. The 41 samples were therefore classified according to the "vanilla" descriptor to begin with. The same was done with the second "citrus" descriptor, then with the third, fourth, etc., until the 28th. By these data, an internal validation was made for each descriptor of the tertiary equations. It was then verified that for each of the 41 mixtures the combination of the 28 tertiary equations of the 28 odor descriptors gave the same olfactory profile as that given by the panelists. By these analyzes it has been validated that the tertiary equations are able to predict the panel data that has been used to construct the invention. Indirect external validation of equations A formula given by the system for which an interesting olfactory profile was obtained was obtained, for example: q1 = 0.0790 q2 = 0.2957 q3 = 0.1996 q4 = 0.2282 30 q5 = 0,1975 which gives us a vanilla-fruity profile. These weight amounts q1 to q5 were then taken and fed to the inverse aromatization functions, thus the olfactory intensities of the descriptors for this mixture were calculated. So: - Banana »4.82« Peach »3.45 - Red fruits» 3.36 - Vanilla »1.83 In order to make the indirect external validation of the tertiary equations put in place previously, and of the 28 10 flavoring functions inverses we submitted the mixture to the panel. It was found that the olfactory intensities obtained were almost identical to the initial olfactory intensities that had been given by the inverse aromatization functions of the invention. For the other odor descriptors, the calculated olfactory intensities were close to 0, that is to say almost identical to the value 0 which had been initially attributed to them. Using these data, we were able to validate that the 20 tertiary equations constructed during this invention made it possible to predict olfactory intensities on the odor descriptors evaluated but which were not used to construct these tertiary equations. These assessments were made on the 28 odor descriptors and validation is obtained for almost all odor descriptors. Direct external validation of the equations Four initial odor descriptors were chosen, to which different olfactory intensities were attributed: "red fruit" = 3 "caramel" = 3 "vanilla" = 3 "caramelized fruit" = 3 The value of 0 has been assigned to other odor descriptors.

From these initial olfactory intensities, using the aromatization functions established above, the following weight amounts were obtained: qi 0.1709 q2 0.1432 q3 0.22252 q4 0.20753 q5 0.18454 Then, from of these weight amounts, an odorant composition was made by appropriately mixing the A, B, C, D and E bases. The panelists were then again assembled and asked to record the odorous composition obtained. . In parallel, the weight amounts q.sub.1 to q.sub.5 were provided to the 28 inverse aromatization functions in order to obtain olfactory intensities for the odor descriptors. The olfactory intensities attributed by the panelists to the odor descriptors were then compared with the olfactory intensities obtained: Values obtained Values assigned according to the invention by the panel CARAMEL 3.41 BANANA 3.18 FRUIT RED 2.94 BISCUIT 2, 92 VANILLA 2.85 CARAMEL 2.91 FRUIT 2.67 GRILL 2.55 CARAMELISE BANANA 2.55 FRUIT RED 2.45 FISH 2.37 FRUIT 2.45 CARAMELISE STRAWBERRY 2.35 STRAWBERRY 2.45 MILK 2.19 APPLE 2 , 27 GRILLE 2.05 VANILLA 2.18 BUTTER 1.95 MURE 1.73 FRUIT WATER 1.84 FISH 1.64 FLORAL 1.67 BUTTER 1.64 MURE 1.59 FRUIT WATER 1.64 APPLE 1.58 AGRUME 1 , 64 CITRUS 1.58 MILK 1.45 HONEY 1.38 COCO 1.36 GREEN 1.25 HONEY 1.18 ACETIC 1.25 FLORAL 0.91 BUTYRIC 1.16 ACETIC 0.91 COCO 1.05 GREEN 0.82 EPICE 0.82 CHOCOLATE 0.73 PHENOL 0.76 EPICE 0.55 CHOCOLATE 0.53 PHENOL 0.55 CHEESE 0.48 BUTYRIC 0.45 RENSE FENUGREC 0.12 CHEESE 0.18 RISE BISCUIT 3.18 FENUGREC 0.09 On the one hand, there is little difference between the olfactory intensities perceived by the pan elists and those obtained according to the invention and, on the other hand, that the olfactory intensities obtained are close to the initial olfactory intensities. From the foregoing results it can be deduced that the process according to the invention leads to reliable and reproducible results in the industry.

Claims (1)

Claims 1. A process for preparing an odorant composition from a target odor O and a set of bases (B1 to Bi), in which the target odor O is inhaled and the bases are mixed in quantities weightings suitable for obtaining the odorant composition, characterized in that the determination of the respective weight amounts of the bases (B1 to Bi) to be mixed is carried out according to the following steps: a) after having smelled the target odor 0, an intensity is attributed olfactory (01 to one) to each of the odor descriptors of a predetermined set of odor descriptors (D1 to Dn) and b) from the attributed olfactory intensities (ol to on), the respective weight amounts are calculated (qi to qi) bases to be mixed to obtain the odorant composition. 2. A process for preparing an odorant composition from a set of bases (B1 to Bi), in which these bases are mixed in appropriate weight quantities to obtain the odorant composition, characterized in that the determination of respective weight amounts of the bases (B1 to Bi) to be blended is effected by the following steps: a ') an odor intensity (al to on) is assigned to each of the odor descriptors of a predetermined set of n descriptors odor (D1 to Dn) andb) from the attributed olfactory intensities (oi to on), the respective weight amounts (qi to qi) of the bases to be blended are calculated to obtain the odorant composition. 3. A process according to claim 1 or 2, wherein the calculation of step b) is performed using previously established aroma functions ax (x ranging from 1 to j). 4. A process according to claim 3, in which the aromatisation functions are previously established in the following way: a set of bases (B1 to Bi) is prepared, m-mixtures (M1 to Mm) are prepared; bases (B1 to Bi), - a set of n odor descriptors is defined at Dn), - each of the m mixtures (M1 to Mm) is separately sampled and for each mixture Mk is assigned (k varying from 1 to m ), an olfactory intensity (dik) at each of the n odor descriptors (D1 to Dn), and for each odor descriptor Dp (p varying from 1 to n), a function fp connecting it to bases (B1 to Bi), and - from the n functions fp, j aroma functions are calculated ax (o1, -, 0n), (x varying from 1 to j) which relate the weight quantities qx (x varying from 1 to j) with olfactory intensities attributed (oi to on). 5. The process according to claim 4, wherein in step b) the weight amounts (q1 to qi) are obtained by applying each of the flavoring functions (x ranging from 1 to j) to the n intensities. olfactory (01 to On) - 6. A process according to one of claims 1 to 5, wherein the number n odor descriptors (D1 to Dn) is greater than the number of bases (B1 to Bi). 7. A process according to one of claims 1 to 6, wherein the number of bases (B1 to Bi) is at least 5. 8. The process according to one of claims 1 to 7, wherein the number m of mixtures (M1 to M) is at least 20, preferably greater than 40. 9. The process according to one of claims 1 to 8, wherein the number n of odor descriptors (D1 to Dfl) is at least 20. 10. A process according to any one of claims 1 to 9, wherein each of the bases (B1 to Bi) comprises at least 20 constituents. 20 11. A process according to one of claims 1 to 10, wherein the bases (B1 to Bi) comprise a fruity base, a chocolate base, a spicy base, a floral base and a butyric base. 25 12. A process according to any one of claims 1 to 11, wherein the n odor descriptors comprise the odor descriptors "banana", "grilled", "plastic", "strawberry", "rancid cheese", "Phenol", "coconut", 30 "butter", "acetic", "citrus fruit", "caramel", "vanilla", "strawberry", "banana", orange "," butter "," floral ", "Rancid cheese" and "butyric". 13. The method of claim 12, wherein the n odor descriptors are selected from "caramel", "vanilla", "strawberry", "banana", orange "," butter "," floral "," rancid cheese " And "butyric". 5 14. A process according to any one of claims 4 and 5 or any one of claims 6 to 13, when these refer to claim 4 or 5, wherein the establishment of flavoring functions 10 further comprises at least one step of validating these functions. 15. A device for aids in the preparation of an odorant composition, comprising: means for capturing olfactory intensities corresponding to odor descriptors; computer means comprising at least one memory and calculation means; said memory comprising a program capable of initially creating aromatization functions aX from the olfactory intensities seized, and then applying these aromatization functions to olfactory intensities seized subsequently, so as to obtain weight quantities, and - means for displaying the weight quantities obtained. 16. A device for the preparation of an odorant composition comprising: an apparatus for the qualitative and quantitative analysis of a substance, this apparatus being able to determine olfactory intensities corresponding to odor descriptors; computer means connected to the qualitative and quantitative analysis apparatus, comprising at least one memory and computing means, said memory comprising a program able to initially create, from olfactory intensities determined by the qualitative analysis apparatus and quantitative, flavoring functions, then to apply these aromatization functions to olfactory intensities subsequently provided by the qualitative and quantitative analysis apparatus, so as to obtain weight quantities, and - a dosing apparatus and mixture, connected to the computer means for receiving the quantities obtained previously and, from these last res, dose and mix bases to obtain an odorous composition.