Does Adobe Need to Start Again and Build Lightroom from Scratch?

Does Adobe Need to Start Again and Build Lightroom from Scratch?

Lightroom totally dominates the realm of digital asset management (DAM) — a solution for everything, it fits the mold of most photographic workflows, however the bitter pill to swallow can be the treacle-like performance and that monthly subscription. Is it time for Adobe to start again?

Digital asset management is something we all do as photographers — whether its as simple as copying image JPEGs straight off your SD card and dumping them in to a "Pictures" folder, or ingesting the raw files pre-tagged into date named folders that are cloud synced for anywhere access. The care you take will depend upon what you want to achieve and who you are delivering the images to.

Of course it's not always been this way. The product that eventually became Lightroom — Shadowland — started out in 1999 with the first public beta in 2006 before version 1.0 arrived in 2007. Rolled into that first product was Pixmantec's Rawshooter which enabled Lightroom to de-mosaic raw files, in addition to more traditional JPEGs and TIFFs. In one stroke it nestled up to Photoshop, leaving the heavy lifting for layer based editing, whilst specifically targeting the photographic workflow of ingesting images, raw conversion, and non-destructive edits.

The paradigm was to offer the digital equivalent of the darkroom, allowing you to develop your images to a final product. Perhaps this is where Lightroom really shines and highlights Adobe's true focus: the creative process and ultimately the output. Lightroom was intended to deliver print and digital media which is why the book, slideshow, print, and web modules have such prominence. The crowning glory is arguably the image catalog which is incredibly important for maintaining a coherent digital archive and — again using the analog metaphor — as part of the creative culling process. It's laid out as a contact sheet for a reason, to allow you to tag the photos you want to develop. Of course, the catalog remains the big buy-in: if you have your library of 100,000 photos then that is a big disincentive to move to another product.

This process has served film photographers for decades and fits the mold of the digital realm. Lightroom has hit the sweet spot when it comes to workflow and with each iteration adds increasingly sophisticated levels of photo development. All of this is non-destructive leaving the original digital negative untouched. So what's the problem with this?

So What Is the Problem?

The dramatic expansion in digital photography from the 1988 Fuji DS-1P through to the rise of the burgeoning DSLR market fitted the niche that Lightroom was intended to fill. The DS-1P could shoot up to ten 0.4MP images on its 2MB memory card, whilst Nikon's D1 shot 2.5MP images on to a 2GB CF card. Early on images were predominantly JPEGs, but DSLRs made the raw format more commonplace. More problematic was the vendor specific file so if you shot with different brands you needed a range of manufacturer software products to import them. Photojournalists exemplified the problem where multiple photographers, potentially with different cameras, needed to go from digital file to broadsheet.

What has changed since the birth of digital asset management is our shift as photographers to shooting more images than ever, using higher resolution sensors that create larger files. This "wealth of the visual" is creating a data headache that affects all aspects of the photographic workflow, foremost amongst these is the size of the data archive being created. When shooting film there was always an upfront cost associated with image processing: you paid for the film, the development, and then the printing. There was a cost at every stage, before you carefully indexed and filed your negatives. Digital was heralded as an almost "no cost" solution; you already had a computer and just dumped those tiny JPEGs in to a spare directory. However with cameras such as Fuji's GFX100 creating 100MB+ size files you need large media cards, an ultra-fast connection to your PC, manifold storage, and a large back up solution. If you are a wedding photographer, shooting 2000 images for a single event is common, which creates a significant data processing headache. That's 200GB of data for one wedding which needs to be ingested, culled, processed, delivered, and backed up. That entire processing chain costs a considerable amount to set up.

The problem still remains asset management, but added to this task is one of data management. It's not so much that Lightroom can't manage your photos — it can — but rather how quickly can it do it.

Rapid Asset Management

As a result of the much greater number of larger image files, we are now seeing pressure on the software that manages those photographic assets; when files were small there wasn't an imperative to seek high performance processing, but this has become an obvious bottleneck. This is even more important in time critical photography such as sports and news, where you can be required to upload your imagery literally seconds after having captured it. There is an acute need for Rapid Asset Management in these domains, but all areas of photography would benefit from being able to rapidly cull and catalog imagery. I would then separate processing in to two areas: those that require simple batch driven edits and those that require more refined manual processing. The former benefits significantly from being integrated in to the culling process, whilst the latter can more readily be performed externally (for example, in Photoshop). Tethered capture is perhaps a special case.

Rapid Asset Management is relatively new as Lightroom has largely sat on its own, with bespoke products (such as BatchPhoto and PhotoMechanic) targeting the rapid processing of images outside of an image catalog. Competition has come in the form of image processing on its own (e.g. Photoshop, Affinity Photo) or photo based ingestion and de-mosaicing products. These tend to operate in the same vein as Lightroom, focused upon photo editing and broad global edits, rather than the layer based model of Photoshop. This has changed over the years with adjustment brushes, control points, and more recently adjustment layers. That said, integrated cataloguing has been late coming to many products (e.g. Luminar, PhotoLab, CaptureOne), yet this can be one of the most important tools for a photographer that can lock you in to a product.

I know one of the tasks I dread after a wedding is ingestion and culling. It can be a soul destroying monotonous task but in terms of delivering the final product, everything builds from this. It is critical to cull out the images you don't want, tag the keepers, and flag anything else that is worth returning to. I then also want to color code them depending which media stream they will be delivered on. Lightroom is satisfactory for this process, but the import routine isn't flexible enough to let me cull, tag, and keyword at the same. I find myself either importing everything and then culling/tagging, or culling only then returning to tagging later on. This either wastes time importing all the images I don't want or repeating everything twice and then struggling with Lightroom's more pedestrian pace. Added to this is the "Adobe Subscription Tax" when I would rather own the software outright. However other products are starting to gain traction in the market with ACDSee PhotoStudio, DxO PhotoLab, Skylum Luminar, and CaptureOne — to name a few — all offering alternatives.

It's been 20 years since Lightroom first hit the drawing board and whilst it has huge market share, photography has changed in that period. Not least the sheer volume of imagery and the need for software performance to be top and center. Lightroom isn't known for it's speedy interface. However it runs deeper: I want to speed up my workflow and adding more developing tools to the interface isn't top of the list. I want to see the greatest efforts put in to culling and tagging at the earliest opportunity, along with lightning fast performance. Does Adobe need to start again with Lightroom?

Lead image composite courtesy of hamiltonjch via Pixabay, used under Creative Commons.

Log in or register to post comments


Nicholas Tinsley's picture

They should start over and write a new version of Lightroom, built for the modern age of cloud storage. It can start off as an MVP and grow to eventually be good enough for most photographers. The rewrite would also be useful because it could be written in such a way that it could work on new operating systems and form factors like tablets and smartphones. Since re-writing 20-year-old software takes a long time, Adobe can keep maintaining the classic Lightroom for a while so as to not break power users' workflows while they continue to iterate on the new ground-up rewrite.

… oh wait

Leopold Bloom's picture

This is already available. It is called Lightroom (the non classic variety). A piece of software I tested and tossed immediately. The interface is unusable on the PC. Such a waste of space that keeps me scrolling around and losing much time.

Howard Shubs's picture

You missed the "...oh wait" at the end, yes?

Duck Man's picture

It's also how they're tossing out rule-breakers. Lightroom CC is still a loophole for those who proclaim to be "financially challenged" or anti-corporation.

derek j's picture

i think until they start seeing a significant number of subscriptions being canceled, which probably wont happen until there is also a viable photoshop alternative as the "industry standard" its unlikely to be done, as their investors likely are more about a few shiny features that keep the subscriptions coming than a costly overhaul that will make some people happy, but not lead to increased revenue. just like we all probably think to ourselves, cool update, but can you just make it run faster and eat up less of my memory? but again, we didnt add anything new this time, but decreased memory usage doesnt make for a good quarterly investor update.

Jim Bolen's picture

The software is perfectly fine. It just needs to be faster. Even with a beefy computer, it still bogs down quite a bit.

ignacy matuszewski's picture

I can agree. Tethering feature is a nightmare though. Crappy, unstable and with no functionality.

Alexander Petrenko's picture

They've started. It is called Adobe CC (not Classic).

PS: Doesn't seem to help too much.

Jesse Hildebrand's picture

While I would love quicker import/culling... the main thing that's keeping me from jumping from Bridge to LR is it's just not built for modern workflows involving a local NAS, cloud storage, multiple editing stations using the same catalogue etc. It's still built around the Jurassic concept of a photographer doing everything on one editing PC at home. Everytime I get up the nerve to try and switch, I read through the hoops and hacks to get it working and park it for another year.

Rayann Elzein's picture

To be honest, lately I have had as many head scratching moments with DaVinci Resolve than with Lightroom... Something's just not playing nice on my computer/driver/software version combination and renders may or may not work as intended depending on the mood of the day...

Les Sucettes's picture

The bitter pill is that Adobe milks the traditional design / photography fields to cross finance new areas such as 3D types of software - where the bulk of the money flows towards.

There’s really not any big difference between Photoshop 2020 and Photoshop 20 years ago. And actually it shouldn’t change all that much.

So really what should we be paying for? Well, really just the keeping the software up to date for the latest Operating System, and security, speed improvments that sort of stuff and okay maybe the odd feature that gets added like Sky replacement (which I personally couldn’t give a damn about because if you know what you’re doing it always was easy.

But instead we (and Graphics Designers) are paying for the development of software that has no benefit to us at all - I couldn’t give a damn about 3D, VR, AR etc... but I pay for all the great new developments in this domain.

So Adobe can f right off. I don’t use LR, I use C1 which is a better RAW converter. And I share a Photoshop license with another Designer

Derek Novaes's picture

Did any of you attend the Adobe Max conference this week?! Some of the things that Lr can do now is pretty amazing. You can see all of the sessions still, they will be up for another year. Just search for Lightroom and you will see three great videos by Ben Warde. Minimally check out the 3rd one.

Ryan Cooper's picture

The challenge with Adobe software is that while, yes, they are continually adding really cool and impressive features, Adobe leaves structural issues to simply languish. Ps and Lr have had major performance and UI issues for years that have been completely ignored. As a customer, I'd be thrilled if Ps/Lr received zero new features for like a year but at the end, I got to update to a re-architected piece of software that is blazing fast.

In 2011 I regularly edited 2-4gb composites using 36mp source files on an iMac. Today I work more with simple 24mp retouches on a modern high-end Windows rig. My workflow is slower now than it was then. When editing my powerful graphics card sits idle often while ONE of my eight cores is maxed to 99%. Ps and Lr have some very serious performance issues based around core architecture designed for the old days. Adobe need sto go back to the drawing board and write their application to leverage modern GPUs and multithreading.

It is telling that I can edit 4k video, trivially, in DaVinci or run Shadow of the Tomb Raider at 4k Ultra RTX at 30+ fps, but Ps stutters like crazy as I try to use things like the clone stamp on a flat image. I've left Premiere in my rear view mirror and am thrilled about it but as of writing this comment there are no true alternatives to Ps. (Affinity isn't there yet)

Tim van der Leeuw's picture

I suspect that one of the reasons why Adobe cannot easily "fix" this is the plugin architecture and desire to maintain compatibility with existing plugins.
It will be very hard for them to move to a multi-core architecture without breaking plugins.

Regarding usage of GPU, there might be similar complications integrating a GPU based processing pipeline with plugins that are not written to take advantage of the GPU and would break that pipeline; I'm speculating a bit here.

Ryan Cooper's picture

I'm not talking about plugin performance, just general laggy-ness using core features.

But yeah, eventually they will have to take the plunge and accept that maybe plugin developers will have to update their plugins to a new system.

Tim van der Leeuw's picture

I understand that you're not talking about plugin-performance, but what I strongly suspect is that the plugin architecture holds back the performance of LR / Ps itself until Adobe either makes major changes to the way plugins are invoked, or until they come up with a new plugin API and architecture. (Which I think they're doing?)

Duck Man's picture

Yes, I remember a few years back, video editor programs were also reluctant to use GPU acceleration. GPU should in theory lead to more speedups but in my experience it also leads to some odd hiccups which don't occur with GPU acceleration disabled.

Rayann Elzein's picture

I honestly don't care about what LR can do now. I have updated version when it came out when I read they promised improved performance. I see no difference whatsoever to my workflow. I am currently only able to use my laptop, but with a 10th Gen 6-cores i7 CPU, a 1650 ti GPU and 32 GB of RAM, I think I have the right to expect a smooth workflow when I don't even apply any radial/gradient or any kind of local adjustment, no?

Tim van der Leeuw's picture

It's always easy to just say "rebuild from scratch!" but in reality., that's generally the worst thing you can do.

I don't use LR so I can't judge for what Adobe should improve, but based on comments I'd say it's speed of importing and handling the catalog of images.

From what I've seen of imaging software, there's nothing really that can match LR for features and if you're really going to rebuild any piece of software "from scratch", you will regress in features, if not peformance.

However Adobe won't really change anything in LR if they don't feel the pain of people abandoning their SW for any of the alternatives and as it stands, that's still a bridge too far for many photographers.

Ryan Cooper's picture

Why is it the worst? From scratch means they can build the application on modern, fast architecture while leaving the monolith behind. An optimized turd is still a turd.

From a business point of view, "re-wrote in X" isn't nearly as exciting as "new AI feature" but a reckoning eventually comes if you don't stay modern.

Adobe is experiencing this with DaVinci right now. Premiere users are dropping like flies. The only reason it isn't happening with Ps (yet) is that the "DaVinci of Photo Editing" doesn't currently exist. It will eventually. If Adobe waits until that competitor shows up to start solving their issues it will be years before they can begin to turn back the tide.

Tim van der Leeuw's picture

Rebuilding from scratch means losing a lot of time, standing still, while your competition is pulling ahead of you.

Of course competition is already catching up, and that's good.

But rebuilding from scratch, throwing away all the old code, means that for a few years current LR will remain absolutely stagnant while they put together a new product that may or may not be faster but for sure has fewer features.

No matter how bad the current code-base might be, it's almost never totally "beyond hope". Almost always improvements can be made in the current code base, hotspots identified and eliminated.

I haven't seen the code for LR of course and I assume that neither have you, but in my own experience as software developer you can better rewrite components of your code, piece by piece.

Ryan Cooper's picture

It absolutely takes time, but that time has to be invested one way or another. Foundational architecture is important and Adobe software is running on architecture that is a decade old. Currently Photoshop has no real competition and Lr's main competition (C1) is radically more expensive. Now is the time BEFORE a real competitor shows up.

There is a reason a web app architected using the latest version of React will outperform say one built 10 years ago on jQuery.

There is also a reason a game like World of Warcraft that is north of 15 years old consumes an insane amount of resources relative to its graphics, in spite of relentless refactoring to improve performance by a company that dwarfs Adobe in size.

Adobe won't keep or lose me based on whether they put some random new ML-based algorithm into the filters menu but they WILL lose me if a competitor comes out who's app runs radically smoother and faster.

I haven't seen the Adobe codebase, you are right, but I've been in software development since before it was created. (assuming their claim that CC was a fresh re-write in ~2012 when they claimed it was was true. (I have my doubts of even that though).

Everything was different back then. Sure you can go back and refactor certain parts of the application to perform better but that can't fix underlying architectural patterns that the ENTIRE application is built around.

I'm not saying Adobe should re-write photoshop every year, but re-writing your flagship onto modern architecture once every decade or so isn't an unreasonable ask. Especially when your users have been complaining about poor performance since the day the existing one first launched. I also think it is safe to assume, at this point, that if fixing the performance issues was a relatively easy lift, it would be done already.

Tim van der Leeuw's picture

Well it seems we both have a long pedigree of software development. ;)

As such you have pehaps at some point ran across Joel Spolsky:

Jim Hofman's picture

The latest update killed my Fuji Tethering Plugin. Paying $79 to tether with a Fuji was ridiculous to begin with, but then roll-out an update that kills it? Adobe customer support told me to roll-back to a previous version, but now the plugin doesn't reinstall. What a mess this program has turned into. It's great when it works, but lately that's a rarity.

Giovanni Stoto's picture

Just improve performance and make additional modules an option to save space.

By the way, I use an external software for culling: Fast Raw Viewer. I import from SD to my RAID hard disk, with FRV I do a first culling to delete very bad photos (easy: while culling hit space and when finished choose delete selected), then a second culling for selecting the very very few images I want to import into LR catalog (same story: while culling hit space and then choose "move to folder...). But I'm using a folder based cataloguing system, for others may not work.

Warwick Cairns's picture

I used to have the standalone version and was happy with it but it no longer works with my new Mac. I resent paying a monthly subscription. As soon as I find a satisfactory one-time purchase alternative I’m cancelling.

More comments