Why Cycles-X and not Eevee?

Instead of putting so many resources into Cycles, which is pretty fast as it is, why not invest in Eevee, is it because if they implemented Ray-Tracing in Eevee and make it faster Blender would directly compete with Unreal for rendering films, does it have to do with the fact that Epic Games, the makers of Unreal, invested 1.2 million into the Blender development? :thinking:


Making Eevee use Vulkan to make it future proof means that no investment is being made into Eevee? Continuously improving and adding features to Eevee may indeed partially be funded by Epic.

There is no reason to move to the lands of conspiracies if you want to support the development of Eevee, especially not if you don’t have a solid foundation for your claims.


Where do you read in my comment that I am saying that there is no investment made into Eevee, look again, I am not saying that at all, and there are a couple of maybes in your reply!

Read the 2021 road map, Vulkan is only listed under projects, and it clearly says that Vulkan will not impact speed, at least not at first, and not a word about ray-tracing!

[https://code.blender.org/2021/04/blender-2021-roadmap](https://Blender 2021 Roadmap)

So, my question, which is relevant to this thread, stands, why put so many resources into Cycles, when clearly making Eevee faster and incorporating real-time ray tracing into Eevee would clearly benefit the vast majority of Blender users?

70% of Blender funding comes from corporations, it is hard to believe that this doesn’t impact development even though Ton said it wouldn’t, and this move of having several programmers working on Cycles for at least 6 months instead of investing in Eevee, which is, at least IMHO, the greatest improvement of Blender 2.8, makes me believe that there is something here.

1 Like

My main point was to address the silly conspiracy claim. There is absolutely no need for that!

There haven’t been huge contributions to Cycles by the Blender Foundation in quite a while as far as I can see. They made some internal changes, such that Brecht could again work on Cycles and this is a result of that. According to the feedback from the community regarding Cycles-X, there are also a lot of users who are keen on getting those improvements!

This is completely unrelated to ray tracing in Eevee. Yes, that would be great, but it has nothing to do with Cycles improvements.

Edit: To also consider the changes in the reply post.

Since Eevee was introduced (and even before that), there has always been at least one developer working on it fulltime. This has not been the case for Cycles in quite a while. Now, they are focusing more strongly on it for some time and you are starting to complain because of that?


There is nothing silly or inspirational about my question, and no, it has nothing to do about complaining!

I am asking why they decided to put several programmers on Cycles, when clearly making Eevee faster and making it support Ray-tracing would serve the vast majority of Blender users.

Real-time rendering is the future, and that was the point of Eevee, so why are they backing down on the promise that Was Eevee?

1 Like

Just to note, Cycles will remain a critical part to many workflows because while Eevee produces great results for animation, it cannot hit the heights of physical realism that Cycles can provide (which includes caustics, random-walk SSS, and completely unbiased GI).

The way I see it, Eevee is for general animation and NPR, Cycles is for film-quality animation and high quality realism. Both have their own strengths.


I totally agree, but since the vast majority of Blender users would use Eevee instead of Cycles, especially if it had ray-tracing, I fail to see the logic here!

1 Like

Now we moved from conspiracy land to speculation lands. I might continue the discussion in another thread, where it is not off-topic.


You’re the one speculating that maybe Epic is investing in Eevee development, not me, and my question is right on topic, like it or not!

And the fact remains, 98% of Blender users would be more than happy with a faster Eevee with support for ray-tracing!

I rendered this Unreal demo at 4K on a GTX 1650, and it took less than a second per frame, it looks great, and no ray-tracing was involved, this could be Eevee, if proper funding was put into it!


Most studios would welcome a real-time, or near real-time render engine with accurate lighting, shadows, transparencies, reflections, refraction, caustics and the likes, it would cut time tremendously in lookdev, one of the most time-consuming part of a production, why do you think more and more studios are using Unreal for lookdev, and some even use it for final render!

Final rendering on the other hand is low in a pipeline priority, as it requires little man power!

So, my question stands, why spend so many resources on making Cycles faster, when making Eevee a real-time ray-tracing render engine makes much more sense and would benefit the vast majority of Blender users much more?

To spend so many resources on Cycles will mostly benefit studios, IMHO, not the average Blender users, again, IMHO!

1 Like

Perhaps the willingness to use Cycles instead will greatly increase once it has path guiding and many-light sampling, in which both are on the roadmap for Cycles X (to note, we’re talking huge speed boosts once those are in).


I didn’t realize that the goal for Blender was “average users” instead of professional studios. I would think that getting Blender taken seriously and used by big name studios producing movies, etc. would be a good thing. Making Cycles faster moves it toward that direction.

And think about future advancements in hardware. With new hardware that has incredible rendering horsepower, a more efficient and capable Cycles may soon become “real time”, making EEVEE look like unecesary hacking.

Reading your statements makes me think that your focus is on fantasy characters, sci-fi, animation, etc., where EEVEE excels. For subject matter where photo realism is the goal, cars, architecture, landscapes, etc., Cycles is the way to go, if I’m not mistaken.


I find this whole premise to be flawed. Are you aware that the foundation was allowed to look at unreal engine’s source code while they were developing eevee? So since epic games was aware eevee was being developed, why didn’t they pull strings in the initial development phase and have the foundation incorporate eevee into blender in such a way that would keep a third party from improving it in exchange for the 1.2 million dollars?

Regardless of what the foundation does with eevee, there’s nothing stopping outside developers from forking it and adding features to it that would make it more of an unreal engine competitor. There’s already an addon from an outside developer that adds SSGI. What stops people like that from adding features directly to eevee? They may even develop a way for eevee to hand off some elements in the scene to cycles and have both renders work to produce the final image in the future, so we can’t even say improvements in cycles won’t impact eevee.

To me this is just another case where someone would rather development focus on just the areas they want. The foundation committing to improve one area doesn’t mean they have anything against the other areas. They can’t improve everything at once and there will always be users who aren’t satisfied with what they do chose to work on.


There’s another point I would like to make. You mentioned earlier that 70% of the money the dev fund receives is from businesses. I’m pretty sure those aren’t businesses with rival render engines. They mostly need eevee as much as everyone else, so you could argue they would have more influence than epic games in this situation. 1.2 mill across 2 years is nothing compared to maintaining the consistent support of both the other involved companies and blender users. I my opinion the only reason amd and nvidia are interesting in improving cycles is to have it function as an overgrown bench-marking tool. They can’t stop tech sites and youtube channels from using it for that after all.

Don’t worry. Eevee is still in it’s infancy. Of course it won’t blow unreal engine out of the water at this stage. In doubt they’d do too much before the vulkan work is done.

1 Like

Could you please elaborate?


i agree, im an architect and i only use Cycles for my work


You keep repeating claims about what the ‘vast majority of Blender users’ would or wouldn’t prefer, and what would ‘clearly benefit’ said ‘vast majority’, over and over again.
I’m curious: How do you know about the needs and wishes of the ‘vast majority’ of Blender users?
Can you point me to your sources for these observations?

greetings, Kologe


That does not really make sense to present things like that.
Any subject can be represented in any graphical style.

Currently, all of these are more accurate and more simple to set-up using Cycles.
So, quickest way to obtain an improvement is to make Cycles faster, period.

For vast majority of users, EEVEE is not capable to deliver same result as Cycles.
And independently of question of speed, it will take years to bring Cycles features missing in EEVEE, into EEVEE.
Cycles has been developed since a decade. And its use is still in phase with new hardware developments.
EEVEE is only 3 years old and does not have same amount of contributors, just because of its age.
And by its nature, presuming that, at some point, “it will be able to deliver same quality as Cycles, in future” is still science fiction.
If improving EEVEE’s indirect lighting had been an option for this year, there would have been a discussion.
But there are so much more basics to settle, first : that is not. Clément’s target, for this year, is Vulkan support.

So, any generalist user has interest to see Cycles improvement happening, this year.

Money does not have the super power to modify history. Money is NOT time.


Doesn’t real time ray tracing require recent high end graphics cards? I highly doubt “98%” of Blender users even have those yet. I don’t expect to have one until at least 2024. Maybe you meant “98% of professionals who make money with Blender”… but even then I think 98 is a laughable stretch.


Hi! Few thoughts about your statement.

As already was said, I’d like to know how did you figure out needs of majority? In my opinion sheepit farm, gallery and Artstation can give good picture of which engine people use mostly for production. Besides this I regularly see complaints about Cycles, where speed not only one problem, also a lot of people asking about Corona and Vray for Blender to switch from Max and other soft, because Cycles for now hardly can compete with other engines.

I remember in official statemen, new rendering team will support Cycles as well as EEVEE, so i suppose after finishing CyclesX project next one will be dedicated to real-time improvements.

And finally, UE4 is completely free for everything except games with 1m+ profit. So what a problem to use one of the best engines? Also release of UE5 probably will change realt-time industry at whole, so maybe it better to wait until release to know how to improve EEVEE?!



Since the beginning of 3D modeling and animation software there has always been the biased directx based openGL based game engine rendering versus the unbiased software based ray-tracing engines. Unbiased high quality brute force raytracing and light transport algorithms has always had its place in rendering and its the only way to get the most visually convincing and accurate rendering… But if the day comes when one of these biased vulkan, openGL or Direct X or AI based rendering engines can fake everything to the point i visually cannot tell the difference, then i would make the switch. My guess is it will take AI and machine learning to close the gap and make it visually indistinguishable. Start looking for video games that go beyond just DLSS, and start to use AI for actual game effects/shaders.

Going a bit off track here but it reminds me our own brain renders the data it gets from the optic nerve not by doing brute force calculations but instead it fakes its all like an AI would, optimizing and culling stuff we dont need to notice, and generating an image that is convincing to us but is not an unbiased representation of the actual physical world.