I just discovered Natron and was wondering how it compares to the new compositing and VFX features in Blender 2.80.
Natron is a powerful Digital Compositor that can handle of your 2D/2.5D needs. Its robust OIIO file formats and OpenFX architecture is what make Natron the most flexible open source compositor for the visual effects community. Its interface and functionally are the same across all platforms such as MacOS, Linux and Windows. Natron has a powerful keying, roto/rotopaint, 2D tracking tools that are staple for all current film production project that requires visual effects.
Natron had a lot of hype behind it as it promised to bring commercial-grade compositing for the FOSS world.
Unfortunately, the development was powered mainly by a couple of developers with little to no funding to keep it going. The consequence that resulted is the development rate going almost to zero (ie. barely enough to not qualify it for the FOSS graveyard as even the bugfixes stopped).
Too bad really, they had a lot of support on this forum back when the features were streaming in.
It’s quite usable as it is , blender compositor and natron as roughly the same big features . Blender is more stable and maintained .
But, Natron as cool stuff too over blender :
it has a cache system so you can easily hit play and see your work in motion.
all effects like masks, gradients ect have viewer widgets so you can ajust them in the viewer instead of changing numerical values in nodes.
it’s openFX compatible, so you can use commercial plugins like RealSmart motion blur, frishluft lenscare ect…
it as many comunity nodegroups available, all the Gmic filters, and a shadertoy node that allow to create your own effect using shader coding.
you can get similar effects/nodes than blender, but in Natron you got much more options, mainly because parameters arent in the node itself but in a dedicated toolbar.
I use it when I want to push my comp a little further than with blender, I haven’t used it as a main tool on a big project but I use it regularly on commercial projects to make a few shots.
It’s quite usable and complete if you forgot some stuff like a 3D viewport, and some effect that you can replace
with commercial plugins.
It really depends on the type of work you’re doing if it’s worth using it instead of blender, with blender it’s all integrated so you don’t have to export all the passes, so for adding quickly some filters over a render it’s quite handy !
Ok, it’s hard to tell you exactly without seeing the elements.
One strategy could be to animate them in blender 3D viewport and use the compositor to merge them with the video. If you don’t need to animate opacity or using different layer mode , this can be a good solution.
With Natron it should work quite well too, if what you want to do is quite simple and have never worked with Nuke (or I suppose Natron) you may find it a bit difficult at first to even make something simple.
That doesn’t mean the software is overly complicated, just that it’s made for maximum control over what you do, with many option ect… Where blender tends to stay minimalist for the sake of simplicity.
All that said , it’s not that complicated and once you get the logic it apply all over the software. So if you need to get the result as quickly as possible maybe with blender it will be faster, if you have some time to learn some new stuff try Natron, it’s great and a really dedicated compositing application.
I would expect features like this to be brought into products like Blender, giving Blender better compositing facilities and so forth. And, here’s why: when you’re doing heavy-duty compositing, you are probably also working with a 3D world. You have “3D awareness” in the form of object-IDs, Z-distance and so on. But you really don’t want to have to go outside of your 3D software environment to get it.
Check out my channel, Indie Rebel. I was a HUGE advocate for Natron and did quite a few tutorials and trainings showing how to use Natron and Blender in conjunction with each other. youtube.com/indierebel
I’m sad that development has stalled, but at the same time I can understand why it happened. For now, Natron still is a very powerful (and I’d say better) compositor than Blender.
With all the new found support for blender, from sources such as Epic Games, Ubi-Soft, etc; it would seem they, (the blender foundation), could easily reach out to these developers and seek their services in development of blender’s compositor, or help them find similar funding grants to keep their project going.
I don’t have a lot of experience with these types of compositors specifically, but based on the descriptions and features listed; I would be excited to see it’s continued development.
Have you seen the closing event of BCON19? The present part of the community voted for focus on the compositor. So we can expect some movement in this area.
Btw. The fund doesn’t really work that way. Sure there will be some kind of communication between developers on either side but the work has to done by ‘our’ people.
Edit: Sry, you can find the voting in the Developers AMA video.
That’s great! i was simply suggesting that perhaps there could be some more movement towards this getting support one way or another. I have been working my way through the Blender Conference videos, little by little. The one by the fellow from ‘KISKA’ explaining their workflow for high fidelity visualization of KTM motorcycles; got my google searching going. After hearing they used ‘Nuke’, I then searched for open source alternatives. I know Blender does have a compositor; though it seems to currently lack the performance needed to be used in production. (Based on the presenter’s comments.)
I sincerely hope it, (blender’s compositor), will become similar in it’s performance; the way EEVEE has become a game changer to the industry standard overall workflow for VFX, Visualization, etc.
Tbh. I personally don’t see the urgent need for focusing on the compositor when a viable gratis alternative is available with DaVinci Resolve and it’s built in Fusion.
Of course I’d rather like to see it in our beloved Blender and implemented in open way. But at least the work could be done. More headache makes me the situation about PBR texture painting. There we still have no solution and with Substance products moved to Adobe it only got worse.
Don’t get me wrong on this, I want everybody to have what they need, I am just questioning priority.
That’s fair. I’m not going to say its more important than PBR, or Substance painter like workflows; though addressing some of the baseline performance issues before trying to add more features or functionality to the compositor could be a good place to start.
I will have to check out DaVinci Resolve. I hadn’t heard of it until now.
My background is in industrial design, product design as well as exhibits & architecture.
I have been superbly amazed at the improvements in blender over the last 5 years or so.
I always kept my eye on it, (I think 2.49 was my first version I tried out), but never really put much stock in actually using it; until I saw CYCLES. Then EEVEE comes along and knocks it outta the park!
VR is the other area that gets me really excited. I think EEVEE has a leg up on creating optimized realistic interactive content. I hope further development goes to that area as well.
I am absolutely with you on the performance point. That has to improve in every aspect of Blender. But again I point at that AMA. Devs are aware of the problems and it’s interesting to hear their actual views on the things that have to be fixed in order to improve the situation.
It seems like there should be two seperate voting panels;
What do the users want or need? (user votes; likely as you already described from the conference.)
What is required (by the developers) to best reach those goals. (Voting should be mainly developers; but might also include a minority vote by large donors to the foundation, etc. Perhaps as a response, ‘tier two vote’ to developer presentations of how issues could be solved or features implemented. Breaking down the presented options into what is the comparative value of the functionality to users; and what speed of implementation is required.)
I feel like many things fall through the cracks; because there is no simple evaluation process in place to help conclude a direction to take.
That’s pretty much what is already happening. It appears to be a bit opaque at times since the community as a whole is distributed over different platforms and this special one isn’t what we can call heavily frequented by devs/ Blender officials. It’s gotten a lot better since I joined the B-tribe a few years back but can be really confusing when either just starting out or just observing casually.
Perhaps what would be helpful for ‘casual observers’ like myself, (though I might qualify slightly above casual), is to create a standard for what is the criteria for a feature to be officially reviewed. Or what is a path towards that end. Is it 50 up-votes on right-click select? is it less; more? And could there be a more transparent way to reflect the general hierarchy of priorities of say the current top 50-100 tasks for developers. I realize blenderdeveloper.org exists somewhat for this purpose; though its is kind of a mess for a non-developer to get an overview of where things stand and perhaps areas where user input or testing is needed week to week.
I would suggest that the lack of this contributes to poor responses to users with good intentions. It is as much of a website communication issue as anything else.
Blender compositor is quite good for some tasks, and the fact that it’s integrated (you don’t need to save some renders before doing compositing) is quite a time saver in many cases.
However , blender is a 3D application and as many things to improve on this side. Also Compositor depends on the node interface that is shared with materials and upcoming particles nodes.
To be really efficient UI-wise , maybe it would be better to have settings of nodes in the sidebar with more parameters .
It’s very unlikely that blender can become a true alternative to Nuke, as the VSE can unlikely become a true alternative to Premiere. These software have many developpers to work on a single dedicated application. Blender as many developpers too but most of them are split between 3D related work.
However , I found VSE to be very efficient when working on 3D animation for many reason , one of which that it’s the only NLE that can be fully scriptable , and as a pipeline tools it’s awesome.
Same goes with the compositor, it can save a lot of time and disk space, but on the flip side it has it’s limitations.
I think these tools will be improved with time, but for now it must be used as it is, and dedicated application like Nuke, Fusion are here to stay. It’s a wrong assumption that these modules could eventually become as efficient as there commercial counterparts (that would means commercials apps are doing things really wrong).
But maybe at some point someone out of nowhere will come up with the energy to code a realtime compositor working on the GPU like Eevee. Or add a cache system to compositor, or add the many widgets needed to interact directly on the viewer instead of tweaking values. Not to mention some nodes or filters that needs to be added.
This is what hapened with grease-pencil and 2D animation throughout the past years, so it can work. That’s one of the true power of open-source, someone adding some unexpected features. But I doubt that it will be a BlenderFoundation initiative even if they can help in the process.
