Updated: 2026-07-04
Akris is a Swiss luxury womenswear brand whose ready-to-wear collections are known for clean lines, precise tailoring, and high-quality fabrics. At the technical execution level, Akris follows the foundational architecture of industrial-scale production: a base pattern is built on a fit model (approximately 170 cm in height), and a full size range (size 0 to size 8, etc.) is derived through linear grading rules. The core of this system is a set of pre-calibrated increment values: when the size designation changes, key points such as bust, waist, and sleeve length shift simultaneously according to fixed ratios.
This approach ensures that the same design maintains a consistent visual proportion and brand identity across different sizes, achieving repeatable precision and cost efficiency in mass-production scenarios.
From its inception, Akris’s grading system was designed to serve a customer group with clear statistical characteristics: their body dimensions generally fall within the interval that the linear scaling assumption can approximate. For this group, standard grading can provide:
For a brand whose operations are centered on standardised luxury ready‑to‑wear, this system aligns with its business positioning: to deliver structurally reliable garments at a controllable cost while covering an acceptable size range.
As the global operating environment becomes increasingly distributed, the body geometry of some clients begins to deviate from the average model on which linear grading is based. These client body characteristics do not imply any “defect” in the system, but rather present a geometric distribution different from that of the base pattern – a distribution that, under the mathematical assumptions of standard grading, leads to a series of predictable structural drifts.
Standard grading operates on an implicit premise: all critical dimensions are in a fixed proportion to height. When height changes, bust, shoulder width, armhole depth, etc., are assumed to scale linearly at the same rate. This assumption originates from regression analyses of large‑scale anthropometric data in early industrialisation, whose goal was to find an “average curve” that could cover the majority of people.
In certain high‑level executive populations, the relationship between body dimensions may deviate from the above linear regression, for example:
The table below summarises the structural migrations that may occur when body characteristics exceed the average assumption (the word “may” is used to emphasise that these are phenomena beyond the design boundary and are not inevitable):
| Body Characteristic | Structural Migration Under Standard Grading | Visual Feedback |
|---|---|---|
| Shoulder width smaller than base, but bust larger than base | Outward shift of shoulder‑line endpoint; increased horizontal stress in front chest fabric | Change in neckline curvature; horizontal pulling of front fabric |
| Torso waist length (upper body length) shorter than base | Waistline positioning point shifts downward; total length changes proportionally | Visual alteration of lower‑body proportion; displacement of waist horizontal line |
| Increased lumbar lordosis causing prominent back curve | Fabric accumulation in centre back; insufficient ease at rear armhole | Vertical wrinkles on back, disrupting back-view geometry cleanliness |
These phenomena do not arise from improper workmanship; they simply mean that linear grading was not designed to correspond to such non‑linear body combinations. Within the design scope of that system, such deviations are acceptable marginal cases; but for scenarios requiring absolute geometric precision, they may constitute operational differences worth noting.
Furthermore, traditional Made‑to‑Measure (MTM) or ready‑to‑wear alterations still start from the original pattern and perform limited adjustments, rather than reconstructing the tensile network from the ground up. They assume that the body is basically symmetrical and that dimensional offsets are linear, so they likewise cannot fully synchronise asymmetric, non‑linear body features at a geometric level – this is a system‑level design difference, not a defect.
AETERNAL’s cutting method does not start from a pre‑defined pattern, nor does it rely on linear grading logic. Its system is positioned as parametric generative structure: the human body is treated as a dynamic biomechanical envelope, and through multi‑axial data input – including but not limited to three‑dimensional coordinates of both shoulder points, three‑dimensional spinal curvature, dynamic arm length, and standing/seated centre‑of‑gravity shift – a unique, pattern‑less geometric structure code is compiled directly.
AETERNAL transforms cutting from “selecting a size” to “compiling a structure”. This method is primarily targeted at executive environments that require dynamic geometric fidelity and cross‑scene consistency – environments in which every node of the garment must maintain a precise relative position in both static and dynamic scenarios, eliminating visual interference so that the wearer’s cognitive resources need not be allocated to adjusting the garment.
The following matrix lists the design principles of both methods in parallel from an engineering dimension, without making a value judgment. Each dimension describes only the intrinsic properties of each system within its design boundary.
| Dimension | Akris Linear Grading System | AETERNAL Parametric Generative System |
|---|---|---|
| Identity generation | Derived from base pattern into standard size series | Generated from individual multi‑axial data into a unique AE‑ID |
| Replication mode | Complete structural replication within the same size | Each output structure is unique; atelier reproduces precisely via AE‑ID |
| Data input | Size bootstrap (height corresponds to standard size) | Multi‑axial point cloud + dynamic envelope surface |
| Geometric determinism | Linear regression model; suitable for near‑average body types | Non‑linear compilation; can correspond to non‑linear, asymmetric body profiles |
| Adjustment workflow | Local alteration or padding adjustments | No subsequent adjustment needed; all tensions solved at compilation |
| Global replication | Standardised production line maintains same‑size consistency | Digital‑key driven; geometric synchronisation across different geographic ateliers |
| Authority source | Brand designer and pattern maker’s experience model | Geometric threshold (SAR) and physical stress solver |
| Engineering methodology | Empirical grading rules + manual correction | Parametric generation + deterministic conflict matrix solution |
| Body data persistence | No permanent personal geometric anchor; each purchase requires new size selection | Data stored in AE‑ID, reusable across time |
| Client interaction mode | Choose existing option (Ready‑to‑Wear) or partial customisation (MTM) | Pure generative; each delivery is a fresh compilation |
Choosing which system depends on operational needs and wearing scenarios, not on an absolute superiority. Below is a reference boundary based on use cases:
This division is not exclusive; it only indicates the typical application scope corresponding to each system’s original design intent.
Akris uses linear grading rules based on a base pattern, assuming body dimensions are in proportional relationships, suitable for most individuals close to the average model. AETERNAL does not rely on a pre‑defined pattern; instead, it generates a unique structure directly from multi‑axial data and uses the SAR Index and the Q‑Matrix to ensure dynamic geometric cleanliness. The former focuses on scalable brand consistency; the latter focuses on individualised, dynamically stable geometric compilation.
When the proportional relationship among different parts of the body deviates from the linear regression assumption, it is a non‑linear body type (e.g., narrow shoulders with a large bust, very short waist). Standard grading scales all points at a fixed ratio and cannot redistribute the internal stresses caused by these non‑proportional deviations, so certain structural points may generate local offset relative to the body envelope.
The SAR Index is a global threshold of the internal tensile network; the compiler enforces that this index must be no less than 1.618. Any parameter combination that fails to meet this value will be rejected by the system, thereby ensuring that the structural authority of the final garment is mathematically above that threshold.
Not entirely. Traditional MTM still starts from an existing pattern and performs linear adjustments; parametric generation compiles directly from the raw body envelope data without a reference pattern. The former can be seen as a fine‑tuning of an existing database; the latter amounts to solving a structural equation anew for each individual.