At this point in the fashion industry’s adoption of 3D and digital product creation, everyone knows the value of replacing a physical sample with a virtual one. It’s quicker. It’s cheaper. It’s smarter. And yes, it’s more sustainable: less wasted material, reduced carbon impact from freight, fewer returns, and so on.
This particular use case has very quickly become synonymous with the idea of 3D in general, but seeing it as the starting point from which 3D can start adding value is short-sighted. As powerful as digital tools are for improving the prototyping and sampling process for pre-planned product lines, their potential begins even earlier.
Consider this scenario: your designers are creating a new range of high-performance outdoor apparel, or sportswear, by either iterating on an existing technical material, or collaborating with a series of suppliers to create a new one. Perhaps it needs to be waterproof and breathable, moisture-wicking and light, or thermal and flexible. If it’s a new footwear category you’re introducing, maybe both a new upper and a new sole are needed.
The process of technical material innovation is traditionally a manual one. Designers and product developers draw up a list of the characteristics they need from a new material, and then determine whether it can be sourced from a supplier, or whether it needs to be created. In the first case, the initial port of call would be a trade show – which is an avenue that COVID has now shut down for the foreseeable future. As a fallback, the design and development team could request material swatches to test, accompanied by information about their performance profile.
In the second case, where the brand chooses to work with either an in-house R&D team or a supply chain partner to create a new material, or improve on one or more facets of an existing one, the process is still often a manual one. And crucially it’s a manual process that requires far more than the 4-5 physical samples it typically takes to bring a pre-defined style from concept to approval.
To develop a new material, multiple rounds of testing are required. And those tests are often performed separately – whether they’re for durability, thermal profile, ventilation, waterproofing or any other essential performance or aesthetic characteristic. Some of these, by their very nature, have to remain physical – breathability has a subjective element after all – but for the purposes of assessing many of these performance characteristics at the same time as the more common criteria of drape and other material behaviour, 3D simulation can not only suffice but actually enhance the process.
With a physical material, for example, any shortfall in performance means creating a new iteration on fabric – whether the problem is with its underlying substrate, treatment, or other composition characteristic. And with synthetics being heavily oil-based, new material R&D is not just a time and labour-intensive process, but an environmentally damaging one.
With a digital alternative, any characteristic can be tweaked in-simulation, with the material draped over an avatar that can be animated to place the right stresses on the garment. Where defects or shortfalls in performance are identified, any component of the recipe can be tweaked and simulated again, without needing to create another round of material samples for approval. And it goes without saying that the proven benefits of virtual sampling then kick in when the design and development team begin to plan complete product ranges around the final, approved material – supplemented by field tested of the final production samples.
Combined, the environmental benefits of all but eliminating physical samples, and of simulating and refining new materials through all-digital R&D can be profound. And the same can be said for the quality of the products themselves: it is simply easier to design a high-performing shoe, for example, when we can simulate its components, identify pressure points, and stress-test its constituent materials virtually, through a process of digital then physical refinement.
There is, though, a critical element that’s missing from this picture: a digital channel of communication between the brand and the supplier. As I wrote in my previous article, existing digital material platforms are heavily weighted towards a scanning-as-a-service model that prioritises the brand experience and exclusive partnerships. This experience is usually omni-directional; the designer or technical developer browses a catalogue of pre-scanned materials, selects the one they want to bring into their 3D solution, and experiments with it before making it part of their bill of materials – or not, as the case may be.
When we are talking about specific products that have reached the sample phase, then communication does happen to some degree between the brand and the material supplier, since the supplier will eventually receive an order. But in the case of R&D and other early-stage development and simulation, there is no pathway for the kind of communication, iteration, and collaboration needed between the two parties.
This is also emblematic of a broader problem: that material suppliers who open their catalogues to customers through digital materials platforms are not then being given insights into how their content is being used when it does not make it into a product’s BOM. While this does not directly affect the brand in the short term (they simply chose not to move forward with a particular material for a particular style), in the longer term it will lead to reduced diversity in digital and physical materials, since suppliers are not given the opportunity to iterate and improve.
This will also become a more significant problem for both creative and sustainability reasons as the need intensifies to find alternative materials to traditional synthetics, and as the pandemic forces smaller production runs where having access to a more diverse range of suppliers – whether they are providing off the shelf materials or collaborating on new ones that can push the envelope on performance without sacrificing the planet.
And it’s here that discovery will become a vital element of any digital materials workflow. As the range of materials available to designers expands, making it simple and intuitive for them to find the right ones for their needs will be a key factor – whether they are choosing on the basis of aesthetics or performance.
In both cases, machine learning has a big role to play, which is why we’ve invested in making AI a cornerstone of the Frontier platform. For discoverability, color, pattern and weave recognition can be managed via a computer vision model, which speeds the process of digital material creation by automating data input that would otherwise have been a task for the material supplier. And at the same time, a natural language processing engine can also help to translate technical terms that suppliers are accustomed to dealing in (“indigo twill”) to broader categories, helping to both distill technical and performance information down into a more usable form, and to power a recommendation engine similar to the ones used by YouTube and Netflix to better cater to designers’ tastes.
And when it comes to product quality and performance, machine learning can also shoulder a big burden: objective criteria such as breathability, weight, rapid-drying and so on can be automatically categorised and recommended for designers who input those criteria. Using the Frontier platform as a channel for collaboration, the designer can then iterate and improve on those criteria as needed – improving a base material in simulation alone before creating a brand new product.
At Frontier, we take a different approach to digital materials – equipping brands with access to the content they need to make informed choices and innovate entirely new styles and categories in 3D, and also democratising access to our platform for suppliers, to open a channel for collaboration, co-creation, discovery and innovation. Because a sustainable fashion industry will not be built on a restrictive approach to 3D and digital materials any more than it will on manual, carbon-intensive R&D processes.
About the sponsor: Frontier.cool is a new kind of digital textile collaboration platform that leverages machine learning to build the world’s most powerful textile and fabric image exchange for the apparel and fashion industries. We run on AWS cloud and can be integrated seamlessly into existing PLM, ERP, and 3D design software, or your fabric library.