4 New Blender Features That Will Blow You Away
Andrew Price gives a quick overview of the exciting new features that appeared in Blender last week.
Almost overnight, 4 new blender features were added to trunk.
Most of us have probably heard rumors of these features for a while, but now they are finally available to play with.
To celebrate the release of these features, I created this quick summary explaining what they are and how to start using them.
The new Cycles render engine is fantastic, but Andrew is wrong in saying that it will/should replace Blender Internal. Its the same as saying that sculpting tools should replace polygon modelling tools. Blender internal is a fantastic biased render engine! Cycles, being an unbiased engine can achieve better lighting and the use of OpenCL and other accelerations is great but it will never be as versatile as Blender Internal. Blender users need both render engines as much as we need both Sculpting tools and Vertex Polygon modelling tools. They’re fit for different purposes.
Hey Jonathan,
I agree that the internal renderer has some pros. I was only repeating what Thomas Dinges (one of the developers) said to me on Twitter. The plan apparently is to make Cycles the default rendering engine. I’m sure that the internal renderer will still be available for use, but the focus will be on Cycles.
At least that’s my take on it.
Andrew
Andrew is likely correct though.
quote from Brecht:
“There may of course be developers that want to continue developing Blender Internal, and the intention is not to drop it immediately.”
http://code.blender.org/index.php/2011/04/modernizing-shading-and-rendering/
quote from thomas dinges :
“The goal of Cycles is to probably replace the Internal Render Engine in 1-2 years and have a fast and modern Render/Shading system which can be used for production rendering and for animation as well.”
http://dingto.org/?p=157.
I also “feel” that Blender Internal and Cycles are two different things, but for some reason, the developers agree that Blender Internal will be eventually dropped.
I guess its just so early in development that I can’t see Cycles getting to the point Blender internal is right now in terms of versatility/integration to make it useful enough for production use for a long long time. My original hopes were to see Blender internal get the lighting solutions similar to Vray or Renderman with features like Irradiance maps, light caches etc for proper light-bounce rendering. To see Cycles come in with promises of ‘in 1-2 years’ till its fully featured has stopped Blender Internal reaching its potential. And restricting Blender to only using an unbiased renderer, in my view, is short sighted and not in the best interests of many of its users.
I look forward to Cycles’ further development and hope the devs keep on surprising us.
Yes, an unbiased (or biased ray-tracing heavy) engine for default-renderer seems pretty weird. There’s a long startup-period in any project where I want quick, sharp and predictable test renders. Modeling, sculpting, animating, messing around with modifiers and features… a huge part of my time spent with Blender, I just want clear quick snapshots of the material at hand.
It’s nice to see nicely shaded auto-pleasant results early in the process, but on the other hand… if it doesn’t stand up for shoddy renders, I don’t deserve do feel cozy with what I’ve produced.
I also think the situation with two powerful rendering engines (+hopefully a trunked Freestyle down the road +some pretty nice GLSL viewport renders) makes for a great showcase of the power of a totally integrated compositing tool.
The most obvious recent example is this: http://vimeo.com/30236062 (Cycles + SSS from internal)
“Thinking outside the box” is a weird term with Blender… it makes creative “cheating” so natural I don’t even know where the box is!
I totally agree with you encn. In fact Cheating is half the fun of working with blender. In fact please remember that Brecht’s SSS in the internal render is in and of itself a cheat! it’s a post process trick and not a true SSS model.
I personally think that they are looking at relying upon external renderers more and more often, hence the perspective that the blender internal of today is on its way out. I personally hope this is the case because then we can have more flexibility eventually with rendering (freestyle would be great but what about better renderman integration!?). With Noodles for mixing results and making more complex materials think about the flexibility and ability to cheet if we could use Cycles, Mitsuba’s fabric rendering as well as Freestyle?!? What about the renderers that are out there that support micro poly subdivision natively? Blender heads wouldn’t need to bother working on that monster of a project!
Of course that’s the potential I see, it will take a long time for this but until then we still have the blender internal to help us… too bad you can’t mix Blender internal and Cycles for different aspects of a material outside of the compositor. I would love to bake some of those renders…
I agree, it would be a shame to loose Blenders ‘classic’ internal renderer, and I hope it doesnt happen in the near future, and I expect it wont. Cycles is great, yes, but its not the answer to every render you could ever want to do.
And yes, theres been a lot of exciting new additions to Blender lately :)
I’m with Jonathan. While Cycles is excellent, the old renderer has some distinct advantages in specific applications. For example: I make planetarium shows and information graphics. A photorealistic raytraced solution is sometimes not wanted!
My hope is that, instead of dropping the biased renderer, they will implement a CUDA/OpenCL render option for it. Further, I hope that the materials will eventually be unified, so scenes can be rendered both ways without having to change material settings.
I contribute to the Blender Foundation every month. Hopefully that gives my opinion some weight. :)
Cycles is great, but I hope it will be possible to have Cycles with a different algorihtm to allow it to be biased and do Global Illumination – this would be especially important for animations. Other 3D apps have this already and for a long time. I’m no programmer, so maybe I’m just dreaming!?
Hmmm… I once turned the diffuse path to zero and then Cycles changed to “direct lighting” mode… Or not. Who knows.. :)
No much to said, everything has been already told ! If you are an artist wanting to create a full photo realistic scene, may be you will love Cycles ( Actually I heard in the last Blender MP3 Podcast someone saying: ” After using Cycles I almost feel like no touching Blender Internal ever again”>>>>Did he really mean it or this guy has never made an Animation in his life?),
but if you need to render a 30 secs Cartoon Commercial, probably BI is the way to go for speed. 750 Cartoon frames will render lot lot lot lot faster in BI than using Cycles, so there are users for every application, is like Jonathan Said about Multiresolution and Polygon Modeling. If one day Cycles became faster than BI using CPU, ( NO GPU ) and much faster using GPU, then that day probably BI could be deprecated, but NO at this early stage..PLEASE , it would make Blender unusable for many!
Sorry I forgot: It is like Ray Shadows and Buffer Shadows Right?
It is like toggling the Ray Tracing Render button !
Speed vs quality, some times we use Ray Shadows, some times we use Buffer Shadows from Spot lights.
My personal hope is that the internal renderer and cycles will remain concurrent in development – it’s better to have a number of mature options, particularly if one is trying to get a distinct look and feel for a project. Blender will be in a very good place once cycles can output passes and be fully integrated with the compositor (and vitally to be able to see local transformation axes in the live render view !). Perhaps it is then that one can properly judge the relative value of the internal renderer
Just to balance up the Cycles-Tomato frenzy a bit:
I really encourage people to dive in and play with Dynamic Paint instead of waiting for tutorials. It’s stupid fun, and the effects are right there in the viewport! It’s such a different new set of tools, it really has to be experienced to be integrated into future creative thinking.
Andrew mentions the texture map-exports, but a lot of the power lies in Dynamic Paint for Vertex Groups (esp. with the new Vgroup modifiers). Any Modifier that uses Vertex groups can basically be used in a huge number of new ways.
Just a couple of examples: dynamically erasing parts of a mesh with Mask, controlled by a particle system. Let’s see… yeah, particle hair length can be dynamically painted in real-time, no problem. And shrink back into the mesh if Fade is active.
Heck, a character model could probably walk through a Brush object and be transferred to another rig.
Dynamic wetmap/texture creation is cool, but this immediate cross-functionality with other parts of Blender is something else: As far as I know, it’s unique to Blender! And unless other software packages ape Blender’s central Vgroup-system, that’s not going to change.
So for me, Miika is the hero of this batch of goodies!
I like Cycles for certain things. Nevertheless, the current Blender internal render engine gives much, much faster results for many purposes, especially where absolute photo realisism or physically correct lighting are not of primary importance. Dropping it from Blender would be a big mistake.
i hope cycles will play in opencl 1.0
Damn those donuts look tasty.
pls andrew do not post food when i am hungry :(
Cycles is really a very welcomed feature, but not everybody needs a photo realistic renderer for everything, the internal is fast and useful for motion graphics and personalized look and feel animation (and really fast).
Up to this time my personal favorite one is Dynamic Paint , you can do crazy things with it, the Motion Tracker will shatter VFX grounds, and the ocean simulator is simple to use yet looks really nice and very useful, everything an artist looks for.
As always the devs are storming us with excellent and massive new features…
Have you ever wondered how the Blender 3.0 will be? maybe “Make me a Blockbuster movie” button? :P
while cycles looks really cool, i haven’t touched it personally because it doesn’t seem like a good solution for an animator like myself, correct me if i’m wrong but it seems like cycles is really better for stills that would normally take forever in the blender internal. physical renderers seem to better for stills rather animation in general
now the dynamic paint thing has me really exited after messing with it on some of the test builds there’s something super fun about having physical paint splashing everywhere and making everything look like it accidentally stepped into paintball match.
I tried d/l the trunk 2.6 version like in the vid, but the motion tracking is horrible. I get an error of 20,000. I can only imagine it’s this build, because other versions, the vid has tracked almost perfectly.
I’ve found the tracking features to be far less stable, in terms of handling image caches, tracking, and resolving tracking data.
The older separate tomato branches are what I would stick with for now until the official release of 2.61, of course testing them will be good as it’ll help Sergey and the other developers iron out issues, but it’s definitely not ready for production use at the moment.
Hi Daniel,
I’m the BDFL for libmv, the vision library used for 3D reconstruction and tracking in Blender. I’m interested in any problems you have with the tracker in trunk. In my experience trunk works much better than the old tomato branches. If you have footage that worked before but doesn’t now please send it to me!
Keir
What’s a BDFL?
Benevolent dictator for life.
;-)
We need a render called a Unicycle, where we have a unified biased and unbiased renderers, a unified node and materials systems. I agree that the internal blender will end, since Brecht said it’s already too hard to extend and improve. When something is hard to improve, it will decay over time. And when it decays, it simply dies.
Sometimes i get easier results with biased renders than unbiased ones.
I believe that we should have a Vray like engine. With an option for unbiased mode (RT?). The perfect solution should be to improve Yafaray, the Blender team and Yafaray coders could work together. Or maybe Cycles support biased renders with all professional features: light cache, irradiance maps, advanced passes and a lot of different features.
As much as I like the overall approach, yet cycles is a “no go” for a real production enviroment. The biggest issue IMHO is the simple fact that it can not use the “old” materials from blender internal. This is a really, really tough cookie to swallow.
I just can not re-create all objects I’ve already created for an ongoing production just because of that. Does that mean I’m cutted of from further developments? That would be kinda stupid, not?
I have a querstion for the camera tracker.
Is there a way to measure/determine the necessary parameters of my camera? I did my footage with my smartphone, tracked it, but the resulting camera movement is very bad. I could not find the parameters for the camera of my phone online.
Hi Adam,
There isn’t currently automatic calibration, so you’ll have to guess the parameters. Here’s what you can try if you don’t know the parameters for your camera: find the parameters for a similar model phone and put them in for the camera. Then set “refine intrinsics” to “focal length, k1, k2”, which is on the left side. Pick a good pair of keyframes (the defaults of 1 and 30 are rarely good; pick frames that have good perspective and lots of shared tracks), then run reconstruction. The result should be good if your guess is reasonable.
I am working on an uncalibrated pipeline, but it won’t make it into Blender 2.61.
Keir
The internal Blender renderer should stay concurrent with the new Cycles engine. The Blender internal renderer is used by a lot of people and is well known and people may have many scenes already setup for it that they would have to change a lot for the new renderer.
And again not everybody wants to have photorealistic renderings all the time, how many times do we Blender users have to say that? There are also other 3D programs out there that have more than one renderer integrated.
3D Studio Max for example already have its original renderer (which have evolved considerably over the years) plus the Mental Ray renderer which is also very powerful and they are both very well integrated with the program.
I would like Blender to have more than one choice. I think that the internal renderer should be kept and further evolved.
Everything has its merits and usability. Cycles-tomato Tracking-DynamicPaint-OceanSimulator have different uses.
So far Dynamic Paint seems to me to be the most useful implementation for Animation. I guess the name “Dynamic Paint” does not seem to be attractive to grab new customers. It sounds like: Why I am going to paint Dynamically if I can hardly paint statically?
The Answer is that it is not just Painting, it generates Dynamic Displacements ! ( Cavity, groves, sand Foot Prints in real time etc ), Dynamic Waves ( Hydroplane landing on water ), heavy rocks landing on earth, mesh deformation self regenerated to its original position ( Exorcist movie kind of ) The possibilities are truly endless and it is very straight forward to use. This is something that I am going to use from now on.
Thanks Miika Hämäläinen for you outstanding work :), also thanks to all developers involved in other Branches.