AUTODESK – Smoke/Flame
If you don’t care about Smoke or Flame, probably best to skip ahead to the next page, because I might ramble on quite a bit on this topic…
Smoke is near and dear to my heart, as it was the first real finishing VFX tool I learned after 10 years in editorial. In those 10 years, I was often the client on the couch in Flame finish sessions for TVCs, and I was constantly awed by the power and flexibility of the systems, and the creative thinking employed by the artists in such sessions.
When Smoke on Mac was released ($15k software only), I stepped up my finish game from Final Cut Studio round tripping, to an all-in-one finish system – doing conform, color, VFX, graphics, deliverables – all in Smoke. It was a huge improvement from my previous FCS toolset, but it still wasn’t quite a Flame. But they shared some of the same toolsets, and I even collaborated with Flame artists on finish jobs, as the file formats and most FX work were completely compatible with each other.
A few years ago, “Flame Premium” was introduced as the flagship product – bundling what used to be 3 separate tools – Flame (for visual effects), Smoke (timeline based editorial, conform, versioning and deliverables, with much of Flame’s VFX toolset), and Lustre for Color Grading. This is a fairly complex Linux based ecosystem that can all work together very seamlessly with render farms, SANs, file sharing, assistant workstations, etc – but requires a fair amount of engineering to support the “Autodesk Systems” facility.
The Smoke on Mac was almost exactly the same as the Smoke on Linux – same UI, same under the hood architecture, etc. – in fact, it was basically a virtualization of the Linux system, running on a Mac – but with one key VFX toolset omitted (nodal compositing). Nothing about it looked or acted like a Mac program. There was no “project file” saved to your desktop, no media files visible from the Finder level – it was a very complicated Linux system under the hood, where your project files are invisible databases, and there were other complicated “virtual computers” within your own system that handled functions like caching media and exporting files.
This shared the architecture of the Flame ecosystem – but the nature of the complicated system meant that things could easily break. And unless you were very skilled in Linux admin tools, you could not fix things yourself, so you better hope you had a support contract.
Recently, they overhauled both Smoke and Flame in 2013, brought almost all of the visual FX tools into Smoke (including the nodal compositing), and also brought the Smoke timeline into Flame – so Smoke 2013 had approximately 80% of the Flame tools, and Flame had everything Smoke had and more.
They marketed Smoke on Mac as an all-in-one editing and visual FX tool – and aimed at the FCP/Adobe crowd with a dramtatically reduced price of $3500, while Flame (still sold only as a turnkey Linux system) was still well over $100k.
They did not make a big dent in the Desktop Editor market (for reasons I’ll hypothesize about later), but they inadvertently created a low cost competitor (Smoke) to their own Flagship product (Flame).
They also found a market to sell Smoke to Flame artists, who rarely own their own systems – they usually use the systems owned by a facility – but now they could own Smoke on their Mac at home, or even on a laptop, and with full Flame compatibility, they could take work home for the weekend, do their own freelance gigs, or even plug in their system to the facility and instantly have another workstation available for large jobs.
But guess what happened? Flame sales, which had slowed considerably since their hay-days, took a nosedive.
(NOTE – I do not have any sales figures from Autodesk, this is all based on my readings of the Flame & Smoke community from message boards and personal discussions.)
The new version of Smoke was approximately 80% of a Flame for 10% of the cost – and most of the tools that were missing from Smoke could be supplemented with relatively cheap 3rd party products, like Mocha and SynthEyes – or for more complicated FX, Nuke. I know for a fact that some facilities were considering buying a Flame, but opted for a Smoke, for purely financial reasons. In this era of always being expected to do more for less money – who could blame them, right?
So – that brings us up to speed on the recent history of Smoke/Flame. Here’s what’s up now…
The 2013 overhaul was… a bit rough to say the least. (Let’s not go down that rabbit hole here). But we’re over that hump, and the 2015 versions of both Smoke and Flame are fantastic products. They have both been re-engineered under the hood for realtime GPU performance, 4k output, the entire color & image processing pipeline is now 16bit float, and their feature sets are closer than ever (Smoke finally got a 3d camera tracker, for example.)
(There are still plenty of features that are Flame-only to differentiate the products – projectors, UV Maps, Normals Mapping, etc – and Smoke is missing Flame’s 3D masking and shape creation tools, to name a few….)
Flame is even now optimized for dual GPU cards, effectively giving Flame its’ own built in render node, by utilizing the 2nd GPU as an internal “burn node” that doesn’t even require a network transfer. Very cool!
I’ve been on the beta test program for Smoke, and I can say that this version is far and away better than the 2013 version, and probably what Autodesk wished they could have shipped after the initial overhaul – but market forces won’t allow a product to go 3 years between upgrades, so they were forced to put out 2013 as a “half baked” product. Don’t get me wrong, I’ve been using Smoke 2013 (Ext 1) for well over 6 months, and I’m mostly happy with it, but it did feel years behind the curve when it came to performance/rendering, etc. (I’ve had to say to my clients, “Pardon me while I render those cross dissolves before playback.” And don’t even try to finish 4k in the 2013 versions – painful!)
BUT – and here’s the big BUT… This is all about to change.
Autodesk has re-thought their product line-up, and will make significant changes to both products in the future. For existing customers, the 2015 versions will function as they did before – but moving forward, things will be different.
Flame is now the Flame Family, consisting of 4 products – Flame, Lustre, Flame Assist (on Mac), and Flare. Flame Assist 2015 is basically, no exactly, Smoke on Mac with a Flame skin on it (blue buttons to match the Flame UI). Flare (Linux only, for now) is the visual FX component of Flame – no timeline/conform/export tools – just the nodal compositing – almost like Nuke for Flame. These systems will continue to operate as they have before as part of the Flame ecosystem – all connected, compatible formats with access to the render nodes, shared projects, etc…
Smoke (now only on Mac), moving forward, will become a “Desktop product”. What does that mean? For one, they are following Adobe’s lead, and it will be a subscription-only model for new licenses, priced at $1750 per year, or $200/month. It also means it will no longer share the under-the-hood architecture of the Flame/Linux products. This has big ramifications for some users. While Autodesk haven’t said exactly what that means, they have said that project files will no longer be compatible with Flame systems, 3rd party plugins (Sapphire Sparks) will no longer be functional, 2013 Smoke archives will not be readable by the next version, and there will no longer be the ability to tap into the other portions of the Flame Systems, like render nodes.
In effect, they are going to sever the compatibility link between Smoke and Flame, and they while they may still share the same interface and many of the same tools, they will live in separate worlds – Smoke on Mac as a “desktop app”, and Flame as the Facility Systems Product.
At first glance, it may seem like a ridiculous move – why break that link that enables them to work together? But as they have said, “desktop” users have different needs than “facility” users, and they feel they can better address each target market by making a product geared specifically for each market.
What does that mean it practical terms? (Again, this is just reading the tea-leaves on my part – I have no insider information). I think it means they might radically simplify the Smoke product to act more like a Mac application that new users will find easier to adopt. I hope that does not mean dumbing down the product as far as its’ capabilities – but I would expect the “under-the-hood” architecture to drastically change. Gone will be the invisible databases that make up your “project” – to be replaced by a single “project.smoke” file that can be saved to your desktop or USB drive. Gone will be the complicated process known as Backburner that exported your files (and was also prone to breaking.)
In short, I expect it to act much more like a traditional editing or FX app that runs on a Mac. (again – just reading tea leaves.)
As a current Smoke-Mac user, I find this quite frustrating, as I sometimes use and enjoy the compatibility with the Flame ecosystem – render nodes, compatible formats, etc.
But if I were to look at things from ADSK’s perspective, I think this moves makes a lot of sense.
1) It restores their commitment and protects their high-end facility product – Flame.
2) It removes some of the barriers to attract new users of Smoke, and gives Smoke a 2nd bite at the apple, so to speak.
In my opinion, Smoke 2013 faced 3 hurdles in attracting a new market.
1) Cost. $3500 is not much compared to Flame, but compared to Adobe CC, it’s a lot to plunk down for a user base accustomed to $1000 for Final Cut Studio, or $50 / month for Adobe.
2) Setup, Admin, workflow. All of those Linux under-the-hood architectures made it very difficult for users to adopt Smoke in the way they were used to working. You could not just do a project on a portable drive, and move it from your office to your laptop, for example. Smoke 2013 is not nimble, the way editors and After Effects artists are used to working – passing files and media around between apps and systems. Reliability was also a factor, where one of those background services would break, and all of a sudden, you can’t export files, and it requires hours of tech support and terminal commands to right the ship.
3) Complexity of the program itself. Smoke is deep – especially compared to NLE suites. You really have to commit yourself to learning it to make the most out of it. But this deep power is also what makes it attractive, so I hope they do not remove any of the functionality in order to simplify the product for new users. In the initial beta builds of the 2013 versions – there were many tools that were removed, and eventually put back in after strong pushback from the Smoke user community- but might those tools get lost if they don’t fit the new vision of the program? Who knows?
My hope is that the new Smoke addresses Issues #1 and #2, and leaves #3 in place as the final hurdle, and they respect the users enough that they will learn the tools, if the other hurdles are removed from the equation. Otherwise, they will have Smoke-Lite – and it’s just another NLE – and years behind the competitors. It doesn’t make strategic sense to remove the powerful FX tools, so I’m pretty confident they will stay in the next versions. (But I was also the guy that thought Final Cut Pro 8 would be great – so don’t listen to me!)
But – where will Autodesk focus their development? I’m guessing that they will try to improve the NLE portion – multi-cam support, dual screen layouts, more camera-native format support, etc. And the VFX side of the program will get minimal development in the near future. It might also get opened up for more 3rd party support via OFX? (Have I said I’m just speculating here?)
So each of those paths seem to make sense from a corporate strategic perspective – but guess who gets screwed? Existing Smoke-Mac users that are operating in a facility, and freelance Flame artists using Smoke as a supplement to their Flame systems.
The soon to be released 2015 versions FOR EXISTING LICENSE HOLDERS will be fine, and will continue to support the Flame-Smoke compatibility, but there are many people out there who spent $3500 for that compatibility that will likely be left out in the cold for future versions.
For facilities that own a Flame and a Smoke, Autodesk will upgrade their Smoke license to a Flame Assist (Mac), and they will also have the option to purchase additional Flame Assist and Flare licenses (for $9k each) – but a Smoke owner cannot upgrade to a Flame Assist. So their choices are to stay on 2015 forever, follow the “Desktop Smoke” future (and lose all compatibility with the Flame ecosystem), or step up to a Flame system (which is, of course, what Autodesk wants in the first place.).
More about these issues here:
Whew – did I mention I had a lot to say about Autodesk? Let’s wrap this up with something exciting…
You must be logged in to post a comment.