Fashion is fixated on moving forwards. In the rush to stack the shelves and racks – and especially the eCommerce frontends and distribution hubs – with new products, the industry is not paying a lot of attention to where previous products go. Beyond the sales ledger of course.
Given the world we live in, this makes sense. Only fashion devotees really get attached to past collections, and the entire fast fashion business has been predicated on the idea that if you don’t like what’s available today, don’t worry, because something new will be dropping next week. That model does not lend itself to looking back.
All of which leads to legacy products getting lost in the churn – left out of sight and out of mind. Until someone wants to revisit them for inspiration, for archival purposes, or to iterate on them when they prove popular.
And that’s where the problems arise: because products designed to get out the door as quickly as possible are rarely left in the right condition to resurrect later. Anyone who’s worked in a creative field will recognise the feeling of being under an urgent deadline, and letting disorganisation get the better of them. Component names, material cataloguing, colour coding, and other organisational elements come second to market pressures. And while every designer – and every creative in other disciplines – has the best intentions to make reusable templates and file meticulously as they go, real life tends to get in the way.
But what happens when another designer wants to use those components again? Or when the commercial team decide to try and make use of excess material left over from a previous season?
In some cases PLM can come to the rescue, assuming the products were designed, developed, sampled, and produced according to a proper hierarchy – and that past seasons and collections are archived in-solution.
More often than not, anyone wishing to revisit a past style will have no real choice other than to reverse-engineer it. And that sounds like more glamorous a process than it actually is; at best it’s a case of using CAD drawings as a baseline, and hoping they were drawn to absolute rather than relative scale; at worst it means finding an unsold SKU in inventory, or a salesperson’s sample that didn’t get lost, and literally pulling it apart.
That latter example sounds extreme, but it’s common practice. The team at SO REAL recently spent some time with a historic footwear business, which boasts centuries of heritage and an unimpeachable brand reputation, and discovered that its archive of tens of thousands of pairs of shoes is kept under lock and key and has never been digitised. Twice a year, experienced and emerging designers in their team are allowed access to the vault. And in order to reinterpret or evoke a heritage style, they need to physically separate soles, uppers, stitches and more to figure out how they were put together.
As you might imagine, this destructive process is not sustainable. And while a fast fashion retailer might not mind unpicking past styles to recreate parts of them, brands and retailers that do not trade in volume and speed are sitting on a crisis of history – and one that reaches much further back than just the last capsule collection. The libraries that house their heritage are falling out of lockstep with the digital systems they use to build new products, and tribal knowledge, along with the opportunity to create new experiences, is escaping through the gaps.
This knowledge leak is a problem in any sector, but especially so in couture and high-end design, where individual works of art – like the various Alexander McQueen garments housed in museums like London’s Victoria & Albert – represent culminations of entire careers’ worth of craft, and incorporate techniques and materials that are studied for decades, or even centuries, afterwards.
The goal, then, is clear: these huge back catalogues of products and the statement pieces which are rightly revered by in-house designers and the fashion world at large, should be digitised. But until recently there has been little hope of doing that digitising in a way that will actually allow those historic products to serve a real purpose.
Take 3D. Brands and retailers have, over the last few years, approached 3D from two angles: retroactively scanning in physical products to generate 3D assets, and creating new products in 3D from the outset. But both have their limitations.
3D scanning, for instance, is only a surface-level process. A physical garment – usually a near-final prototype – is captured using a camera array or another imaging process, with the results being stitched together to create a 3D mesh (the actual geometry of the garment as it was posed) with the appearance of materials created through a process called photogrammetry.
That’s a loaded word, but the essence of it is that retroactive 3D scanning of physical products is, in and of itself, a kind of reverse engineering – an attempt to extract 3D information from static photos. And as that description suggests, the possibilities these kinds of scans offer for interaction and inspection are essentially non-existent; retroactive 3D scans are more complex to animate (without capturing a model’s live performance), they cannot capture interior detail or construction techniques, and they cannot be broken down into components.
3D scans serve a real purpose for product visualisation and to populate eCommerce storefronts. And this month’s 3D Balenciaga lookbook is a prime example of the results that photogrammetry can deliver: perfect for turning physical prototypes into 3D objects for informing aesthetic judgements or buying decisions, but not suitable for patternmaking decisions or process-accurate reverse engineering.
Far more useful for future creation cycles are products that were created in 3D to begin with, during design and development, and where the 3D asset travels with an associated 2D pattern up to the point of manufacturing. Unlike 3D scans of pre-existing products, these assets are typically pattern-accurate, can be broken down into their constituent parts digitally, and are accompanied by digital material specifications, sizing and grading information, artwork, and much more.
Coupled with the fact that these assets can also be pushed from 3D design tools and into offline rendering tools and staging platforms for virtual photography, they are clearly the more flexible and fully-featured of the two routes to generating a 3D model of a product that also has a physical counterpart.
But native 3D assets have a major limitation: there just aren’t that many of them. Even the earliest adopters of 3D for design, range planning, fitting and other development processes have realistically only been working to high-fidelity archival standards for a few years, which limits how far back into their archives these brands can do before needing to switch to physical means. Those brands have also, in many cases, only come as far as designing a subset of their entire range – sometimes a single product category – in 3D, while other areas of the business still rely on 2D design, iterative sampling, and commercial photography. When an organisation does make that full leap to creating in 3D, it’s still considered headline news.
And looking more broadly, 3D adoption is still not commonplace in fashion retail. The pandemic looks to have increased the pace at which brands and retailers are switching to 3D workflows, but the vision for an entire industry that creates, archives, and re-uses from digital-native assets is still years away from being realised.
Which leads us to what seems like an inescapable conclusion: fashion is going to have to keep cutting up old styles to unlock their secrets and try to safeguard them for a new generation.
The team at Swiss technology disruptor SO REAL have recently begun to translate their work to fashion. After creating archive-quality digital twins that can be interacted with and non-destructively disassembled for the heritage and media industries, the company is applying the same principles – and the same fusion of new technologies – to fashion’s ongoing digital transformation.
Their approach blends the best of both worlds: the uncompromising physical accuracy of scanning physical products, and the level of quality, interior detail, and construction information of a native 3D asset.
SO REAL are, to The Interline’s knowledge, the first company to repurpose medical X-Ray computer tomography (CT) scanning for the creation of digital twins of physical products. The immediate results created by the scanner are accurate to both the inside and the outside of the physical object, and are then stitched together using machine learning, which refines a complex dataset – that includes everything from minute stitching detail to negative volume – and outputs a comprehensive 3D model with complete interior and exterior detail, segmented into discrete components, and textured in physically-based materials.
And just as the machine learning model can assemble the CT scan results into a complete product, it can also do the reverse: separate every component part, however small, without affecting the physical item. Anyone who has shopped for a car, smartphone, laptop or another piece of consumer electronics is probably familiar with seeing products “exploded” – where a computer-generated whole breaks apart into its constituent components, allowing consumers to see how they work, to appreciate details and innovations that are invisible from the outside, and to understand the impact of configurations and customisations where those are available.
Those explosions, which can be found in both pre-rendered video and in real-time, interactive experiences, are common in industries where computer-aided design and drafting have been the standard for decades, but prior to the adoption of 3D design tools – which is still at an early stage – they were difficult to realise in fashion. And as we’ve established, there has been no way of reaching back.
CT scanning makes these experiences straightforward to reproduce, and repeatable to create. Any product, no matter its age or provenance, can become an interactive, mixed reality asset, allowing not just designers to explore how they are made, but enabling consumers to immerse themselves in the legacy and identity of the brands they care about.
One key difference you might expect between a scan of an existing product – even if it’s done the way SO REAL proposes – and a product designed in 3D from the outset would be that the scan might be an accurate visual representation of the product, but it might not tell you much about the craft that went into it. On the contrary: CT scanning can reveal detailed characteristics of materials and workmanship that can be automatically tagged to generate extensive data and metadata for each component, or even to trace the provenance of a material or the history of a method – with the latter tied in to universally-agreed labour standards.
By digitising an archive like the aforementioned footwear brand’s, using SO REAL’s technology, a delicate heritage can be both preserved digitally for designers to experiment with, and turned into new, immersive experiences for consumers – on both flat screens and in virtual and augmented reality.
In place of a static physical product that only emerges from its vault once or twice a year, and is used for very narrow purposes, creating a complete library of digital twins opens the door to applications all the way up and downstream. From generating a sort of 3D PLM, where parts, materials, and labour can be quick, easily, and non-destructively combined to create new styles, to creating entirely new end consumer experiences that bring loyal fans and new prospects closer than ever to the styles and collections they love.
“The heritage behind some of the world’s most renowned brands is so rich. To imagine that certain houses take apart seams to understand the measurements of a vintage dress is just mind boggling. The dress is no longer the same after such an intervention. With our technology we are able to ensure the integrity and authenticity of the object without a single damaged seam”, added Erinrose Sullivan, Head of Cultural Heritage and Collections at SO REAL, when she spoke to The Interline earlier this month.
As the fashion industry embarks in earnest on its journey towards digital product creation, it’s going to be important to consider not just what it means to create assets for today, but the implications of building a digital future that’s fit to be archived for generations to come. And together The Interline and SO REAL are going to explore those implications in 2021.
About our partner:SO REAL DIGITAL TWINS AG have re-invented precision model technology for XR. Using patented scanning and conversion technology, we’ve automated the production of digital twinned 3D objects, delivering physical objects for a digital world. Showcase ready, cinematic-quality, engine supported 3D assets for use in fashion, games, films, and all XR.
Erinrose Sullivan is SO REAL’s Cultural and Heritage specialist which is very apt considering her passion for the topic, not to mention her expertise; the savoir-faire needed to be an external expert for global discussion and funding panels on the subject. She regularly speaks on the importance of support for art and technology in Switzerland and beyond.