I wished I still could be a Flame advocate

I would say comparing Nuke and Flame is apples and oranges for the most part.

Being that they are both fruit (in the sense that both flame and Nuke can comp), in the Venn diagram where they overlap, they share a space where their respective implementations of tools required to fulfill those functions operate—but everything behind those tools is quite different architecturally and it’s that difference that the discerning eye knows how to exploit.

I think that a viewpoint of “they’re both compositing systems therefore largely interchangeable” drastically misses the point while short changing both system’s relative strengths and weakness contra the other.

In the end it wasn’t simply marketing that drove the changes, but architectural differences which favored one platform for a certain deployment and another for its citrusy counterpart. That’s the long way of saying each system ended up kinda where it was intended from a design perspective compounded by the influence of the folks buying it, deploying it in the manner they saw fit and voting for features with dollars.

If I misunderstood your intended meaning, I’m apologizing in advance.

3 Likes

Hi All,

I see a few people mention that Autodesk should give our software away free to educational institutions and that may help in the effort to build the user base.

Perhaps you are aware of this or not, but Autodesk already offers all the software free to educational institutions as well as students signed up with them. Every Autodesk product including flame is FREE to education. You can see it by going to this link - https://www.autodesk.com/education/edu-software/

I’ve worked along side multiple educational institutions and a lot of it boils down to the resources they must implement a new product and teach in their classes. So finding the right educators, designing the courses and teaching Flame the way it’s meant to be used long term and not in Nuke or Adobe styles. That is the bigger challenge in my opinion.

11 Likes

AD has come a long way in their educational offerings, especially when you look at where they’ve come from.

Given how passionate our user base is, I’m sure a lot of Flame artist like myself would love to talk to students about our profession and how we use Flame to achieve our goals in it. As a student at NYU, one of my fondest memories was of two lighting directors taking our class to the studio of the soap opera they worked for in NYC for a few hours. Being able to have this unique behind the scenes experience with two people who were in the trenches every day (and night) doing incredibly challenging work in extremely difficult conditions obviously left quite an impression on me.

Not sure if the resources exist to do so, but it would be great to have Autodesk facilitate some of these experiences that pair willing artists with local schools to talk not just about Flame, but their trade as well.

Full disclosure: I also read t your entire email in the Grant Kay voice.

3 Likes

As I am in the process of installing Rocky Linux and getting my system ready to try Flame I have to say that I am really surprised that Flame has such a nice community and that logik and autodesk learning channel is here and that autodesk is as I understand supporting these efforts. This is a good sign for me to go ahead and invest my time in setting up linux box :slight_smile:

As my drives are NTFS is the support in Linux and kernel version used in Rocky 8.5 taking full advantage of speed from those NTFS raids? Or should I use native linux format for Flame stones?

Yo @filmtools I’ll relocate this question to a new thread to get you the visibility you need.

I’ve always done xfs or ext4 for Linux drives. What kind of raid ya got?

1 Like

I have 3 raids all raid 0 :slight_smile: Most are custom built by me LSI cards and spinning enterprise drives around 1000MB/sec on those and then I have few 6000-7000 MB/sec drives pcie4.

I have read that write speed on NTFS volumes in Linux ntfs-g fuse is slow on kernels past some number I cant recall. Read is fast but write suffers. Will have to test this under Rocky Linux to find out.

Just made a new topic here : Rocky Linux NTFS volumes or not … so maybe we can del this

I am the happy recipient (and ongoing communication) of one of Grant’s visits to our University as well as one of his colleagues Stuy Holloway (who has since left Autodesk, sadly in a cull of creative finishing publicity and orientation staff) and they were great at communicating the advantages of Flame to the students who attended. Grant did a great demo of a twitchy nose wine advertisement on Smoke on Mac (pre connect-fx) and Stuy did a great demo of the Framestore Eastenders promo and students were blown away by the rain removal and time freeze through recursive ops trickery and DoF for the frozen droplets.

I teach sessions on Autodesk Flame at our place and I always find that students love what comes out of the lefthand viewport in a 2-up view but are intimidated by the left-hand schematics and are often a little resistant to the journey and revert to AfterEffects. I was speaking to a colleague in animation the other day who used to do sessions a few years ago on Fusion and was told by someone in the that students are used to the AfterEffects way and so we should, as it were, cave in.

This is a speedy diagnosis, and not a rigorous one, but the Microsoft-Adobe paradigm that has Word and Powerpoint, kind of looking like After Effects and Photoshop has students thinking that this is the only paradigm. They do loosen up to the nodes when you take them through it, but like someone with Stockholm syndrome the layers and menus become really baked-in.

Again, a speedy diagnosis, you can always de-program the Stockhold syndrome, but you have to invest time and invest various rhetorical strategies to try and budge and nudge the paradigm and this take marketing and it take persistence. This is my old drum, but from around 2010-2014 it looked like Autodesk was gathering a head of steam and building up a stock of clear demonstrations and drawing off a lot of voices in the biz, but this has obviously petered out whilst others are now building up their stock and image bank.

I thought that Autodesk were onto a really nice (beginnings) of a strategy with Smoke 2013 in terms of humanising the imagery in humanising the node (there is nothing naturally intimidating about the node, it is just paradigmatic muscle-memory, as it were, programming) but you have to have persistence of vision and not a miracle view of such deprogramming.

Strange thing is that in 3d the node is obviously pretty sovereign, but for some reason compositing and editorial aspects have a stronger resistance.

I would hate to see Flame humanised and dumbed down at all, but I thought the movements being made on the Manager were a really nice way of double-siding the software and I also kind of think that if something like Python could be used to generate internal software tutorials and could guide you through the software then it would be like having video tutorials and walkthroughs and cookalongs embedded into the software. Parts of the software could light-up, from import through to export, like a C21 version of the Gladiators tutorial, for example, or Fix-it-in-Post? This way of “lighting up” aspects might be like unlocking levels within a game, where you get to see some parts earlier in the “tutorial mode” (a tutorial-sided interface?) and others kind of hidden away during the tutorial sequencing? Whilst I hate the idea of having versions of Flame (like the later Smoke was a smaller version of Flame with some differences), maybe “versioning” Flame and finding a way to guide within the present “perspective” and hide the larger labyrinth might help to, not replace live education, but to help the bits in-between a little more?

What I do love about Flame actually “is” that you are not faced with a splurge of menus but you move through it and it only presents itself to you when you need that given aspect (you could easily not bump into Action, if you want) but, as with a labyrinth, you can easily start to get lost, and so a helping-hand, baked into the software, can give you a quick escape route?

Grant is doing a great job, btw, on those current introductory tutorials and using some footage from one of our student projects!

Just a few speedy-messy thoughts.

Cheers
Tony

3 Likes

Render farms.

1 Like

Burn works.

1 Like

:joy::joy::joy::joy:

with many caveats and lots of baby sitting. It is not a reliable solution. Since getting the insanity ThreadRipper, 3995WX w/A6000, I render locally.

1 Like

@ace_elliott

1 Like

I typed out a response to this way back and never actually published it since I find the topic rather tiring, but since this thread keeps getting bumped and its also germane to the other topic going around about this Who is the person in charge of Flame marketing and promotion right now?, I guess I’ll just come out and say it.

I am drastically missing the point whatever it is. I thought it was that Flame is really only for a very small specific subset of people.

Thats why my initial response was to the idea that Flame is only for 1% of people out there. For me, that just isn’t borne out by my experience. Way more than 1% of the artists I’ve worked with over the years have been Flame artists and even more would have benefited from knowing Flame.

I probably shouldn’t have brought Nuke into it since my original question would have been better worded as

Given that the 1% almost certainly doesn’t include your average compositor(since more than 1% of people in our field are compositors), why is Flame not the right tool for the average compositor?

It just happens that the average compositor is most likely a Nuke artist and so inevitably it becomes a comparison. I reckon I’m probably not the first to compare them.

I agree that Flame and Nuke each have their strengths, but just because each program has its strengths doesn’t mean that they don’t have a significant amount of overlap. The differences between the two programs are exactly why Nuke artists could benefit from learning Flame. The same way that Flame artists benefit greatly from using Nuke. There’s a reason that a recent logik live featured Nuke tools in one of the segments. Conversely, it kills me to see a Nuke artist painting a clean plate up when if they could only chum through all the rushes they would find that little bit at the end of the take where everyone clears frame.

I’m not shortchanging either program, but I’ve spent most of my career working alongside Nuke artists often doing the same work. We can wax poetic about all the differences between the two, but when there are companies consistently hiring Flame and Nuke artists for the same work then I don’t think its missing the point to make a comparison between them.

There is always the right tool for the job, and I’ll admit that Nuke and Flame are often suited to different kinds of work. It is to everyone’s advantage to have that “discerning eye” separating out where each program can play to its strengths. In practice I find that this is an admirable yet often unachievable goal, the circumstances of which are more often dictated by scheduling and artist availability. I said this in the other thread, but it bears repeating. I can’t count the number of times I’ve seen a commercial bid out as Flame hours only to have the producer turn around and outsource half the work to Nuke artists in India because the timeline is too tight to do it all in house. It all comes out the same in the end.

I think most producers don’t care to understand the “architectural differences” between Flame and Nuke. You may not like the viewpoint “they’re both compositing systems therefore largely interchangeable,” but every day there are producers who are just trying to staff their jobs with whichever compositor they can find. And for the most part why shouldn’t they? The jobs that pay the bills are, by and large, relatively simple mundane work that can be done effectively in either program.

I think my experience has probably colored my perception of the issue. In the past I’ve worked at two large companies with teams of 30+ Flame artists. I’ve seen people go years on Flame without touching the timeline, doing conforms, pulling rushes, etc. They were just compositors and some of them were the best I’ve seen. Given that Flame is now cheaper to rent than NukeX, why shouldn’t there be a new generation of artists learning Flame purely as a compositing tool?

At the end of the day, I think Flame is a tool for a lot of different people doing different things. After all it does 95% of the things.

10 Likes

Let me try this another way.

The architectural differences in the compositing Venn overlap–the ones that producers ignore or don’t understand–make up the immense edge nuke has over flame when dealing with, but not limited to 32 bit images, multi-channel exr, data window/roi processing, deep, stereoscopic and now machine learning just to name a few. That list along-side a comparable price (that at one time was less than flame) makes nuke not just the financially responsible choice for a lot of compositing tasks outside of the 30second spot, but in some cases the only choice. And what happens to that calculus when you compound the result with differences in rates between a flame artist and a nuke artist? Add to that all of the benefits of pipelining, nuke ease-of-deployment and integration into the controlled multi-user environment required for long-format shot work, flame’s traditionally complex hardware, infrastructure and general administrative needs? It’s fairly easy to see how, for large deployments of compositors on a huge number of shots, the choice isn’t even close.

You’re viewing the scenario from the artist perspective, which is fair but not really at all where I’m coming from. My point was that the reasons the average compositor is on nuke rather than flame has little to do with artist perspective at all and because of of those reasons learning flame is never an option for most. Now it’s economies of scale–nuke is everywhere. If you were just starting out, learning to comp, you would look at what software does all the big shows and see it’s nuke. You would look for a learning edition that runs on your windows box and see there was one for nuke. Job postings are for nuke, tutorials are for nuke. Everything screams nuke. So you learn nuke.

And in the end I agree with you that flame is great tool. I’m not saying it not, but to say that the compositing differences between flame and nuke are negligible and can be ignored has already proven to be demonstrably false and are very much the reasons why people don’t learn flame, don’t have access to flame or in a lot of cases even know what flame is.

7 Likes

And don’t forget, there used to be a learning version of Flame, but for some bizarre reason ADSK got rid of it.

1 Like

Not true. Educational version…


And 30 day trial.

1 Like

When I first got started 6 years ago, there was a watermarked learning edition that I had on a laptop at home to practice with and learn on when I didn’t have access to a flame. No expiration date for a trial, no affiliation needed with an institution (that I was aware of), just a watermark. I think this is what @paul_round is referring to.

3 Likes

Hi,

The learning version edition went away when Flame switched to the subscription model.

The subscription model allowed Autodesk to offer a 30 days trial period to all users, which was impossible with the previous licensing model.

As a side note, the 30 days trial model is available across Autodesk, while the learning edition was something specific to Flame.

For a consistent user experience, it was decided to remove the learning edition and align Flame with the rest of the Autodesk brand.

I hope it clarifies this topic,

Best,
Yann

2 Likes

All fair points to be sure. Those are all areas that Nuke trounces Flame. I’m not eliding the differences, but I haven’t seen you mention any of the advantages that Flame has over Nuke either. I’m pretty sure I was open about the fact that the softwares have differences and at no point did I say the word negligible. That was your choice of words. What I did say is that the products are competitive with each other, which I think is hard to deny even if you feel that Nuke is winning the competition, and that, in my opinion, the disparity in the number of users between the two programs can’t be accounted for solely by the differences that currently exist. Every place I’ve ever worked at some point had Flame and Nuke artists doing the exact same work.

Do I wish Flame was better with 32-bit stereoscopic deep multi-channel exrs? Of course. Do I come up against that all the time? No. I’m happy to acknowledge I may be a mediocre artist working on b-list projects, but it seems to me that these tools aren’t the bread and butter of visual effects. Can you make money doing full CG environments with years of animation days and decades of rendering days? Sure and Nuke is the right tool for that, but you can make a lot more money removing zits from people’s faces and Flame is the right tool for that. Just happens that everyone has zits and not everyone is a transformer.

Now ML is different. I think Copycat is a game changer and it could be a Flame killer at least for a significant amount of compositing. Makes me glad to know Nuke.

Fortunately we do seem to be in agreement! This was also my point. Foundry has made a number of excellent decisions that were directly aimed at increasing its accessibility. You can call this product design, but its also marketing. The two go hand-in-hand. Flame wasn’t ported to Mac because its better than Linux, it was done to increase accessibility so that Flame could reach a wider market. I just happen to feel that AD has done too little too late. When Nuke surpassed Flame’s market share it wasn’t even close to as robust of a tool as it is now.

Yann’s post is a perfect example of how unimportant Flame is to Autodesk. Someone decided that Flame didn’t look and act like the other softwares in AD’s portfolio so it needed to be changed, user experience be damned. The idea that I need a consistent user experience across AD’s different softwares is a complete joke. I’m a Flame artist not a civil engineer. I could care less whether Flame aligns with the rest of the brand. Because of course if Flame is more similar to AD’s other offerings thats really going to make me want to switch careers and learn Autocad /s. This is the exact approach that let Nuke become the defacto compositing software in our industry.

In the end you’re right, I am coming from the viewpoint of an artist since thats what I am. It seems that we’ve just had different experiences. I’ve spent about half my time working in commercials and the other half longform. Weirdly the longform was mostly being done in Flame and the commercials in Nuke. The longform stuff was done relatively easily with only the loosest of pipelines. On the other hand, I watched a major vfx house struggle for years to make Hiero work, writing loads of bespoke wrap-around python to muscle it to do what they wanted, so that Nuke could operate in the pipeline. Just because Nuke is flexible enough to bend into a pretzel doesn’t make it easy or cheap to do.

Anyway, I could probably go on ad infinitum. This has made me realize I’m just a rabid flame evangelist and my mind is too ossified to change, so I’ll let you have the last word if you’d like.

5 Likes

No need for last words. I’m just here for the conversation :pray:t3:

1 Like