Absolutely love the splash image for this version.

Because I have no patience, I spend some time on the developer.blender.org portal and code.blender.org blog to learn about what's coming ahead of time and I can report some highlights for those curious about what 2.91 and 2.92 will be like:

Some of the highlights:

Compositing
The plan seems to be adjust how compositing works to make it much more responsive and faster. Compositing effects will be reworked to be resolution independent so that the composite preview can be rendered at a lower resolution in the viewport.

Vulkan
Blender will slowly gain more Vulkan support over time, eventually this could lead to performance increases.

Eevee
Depth of Field is getting a refactor, to improve the quality and performance. Highly likely other improvements will come but we'll just have to wait and see what those will be. 'Feature Film Requirements' is the priority.

Animation
Performance optimisations for animation playback began in 2.90 and will continue in 2.91 and 2.92.

Asset Management
A code sprint has already started for implementing an asset manager in Blender that will make it easy to select something from a scene and add it to an asset library, then when editing another Blender scene, quickly and easily add assets from your libraries, the first version of asset management will make it into 2.91.

Grease Pencil
It's coming, SVG Import and Export, as well as Image Trace for turning pixel based images into 3D viewport grease pencil objects. SVG Import and Image Trace will likely land in 2.91, SVG Export in 2.92.

Mesh Editing
Performance optimisations started in 2.90, and will continue in 2.91 and 2.92. There's more tools being added and improvements made for game & environment related modelling tasks. Dedicated retopology tools seem likely at some point, probably 2.91/2.92.

Booleans
A new way to do booleans is coming in 2.91, the new algorithm should be much more accurate than 2.90, which will make hard surface modelling fans happy.

Node Particles
Already part of Blender as an experimental feature, Node based particle systems will probably become a proper feature in 2.91, and be refined in 2.92. Node based hair is planned for 2.91 as well, probably only initial work at first, node based Volumes planned for 2.92.

Baking
A rework of how baking is done in Blender has been on the drawing board for a while, it's likely coming in 2.92 assuming they keep on track with it. The plan seems to be to allow setting up multiple baking passes for collections of objects, and to allow a node based compositing setup for choosing what baking values go into what channels of what output textures. Sounds big.

Sculpting
The ability to paint vertex colours in sculpt mode is coming, 2.91 will probably make it officially a feature and 2.92 will refine it, plus there are plans for improving the performance of sculpting.

UI
A new thing called "property search" is coming in 2.91 to make it easier for new users to find something in the properties panel, and in 2.92 the UIs related to reports and warnings will given a facelift. Probably lots of other small changes.

VFX & Video
Loads of stuff coming, storyboarding tools and more. The video editor will get a major upgrade to make it much more enjoyable to use.

VR Milestone 2
Milestone 2 involves actual editing in VR, sculpting, modelling, object layout, etc.. This should be very interesting, it probably will happen in 2.91/2.92.

There's loads more stuff I've seen on the developer portal that seem likely to arrive in one of the next couple of versions but aren't certain yet and might get pushed back, so wait and see.

Blender 2.91 will be coming around November 24, 2020, and 2.92 will be likely at Alpha stage by new years eve or early Jan. So that list is roughly what they're aiming to have done by the end of the year.

It's insane the speed the Blender devs work at.

View this comment - View article - View full comments

Lovely!

View this comment - View article - View full comments

Love the potential of Ubisoft's Mixer addon.

Especially the fact that they are designing the server portion of the system to allow for technically any application to be used in conjunction with any other application. Not only would it open the door to potentially wonderful collaborative real time experiences with other Blender users, but also users of other software, and maybe even a way of effectively editing a scene in two applications at the same time? And combine the functionality of both?

Lots potential, will watch what Ubisoft does with that addon with great interest.

It's also just nice to see Ubisoft getting in on the fun and making some open source goodness.

Now if only they could open source UPlay..

View this comment - View article - View full comments

Quoting: setzer22Great news! But let's hope there's some room in those 120k to work on bugfixes and stabilizing existing features. Also some of those needed performance improvements for sculpt mode and undo!

That's actually mostly what the Blender devs have been doing for the past couple of months.

I watch the developer.blender.org and code.blender.org sites pretty closely, watch what features they're working on, read each developer's weekly report of what they've been doing, and I can report that 2.90 has a stack of bug fixes in it, undo received a big performance boost in 2.83 and is going to have a few more improvements 2.90, and sculpt mode has a number of improvements as well. Animation playback in 2.90 has also had a few performance optimisations too, and a few UI optimisations, such as the redrawing of the outliner.

So more or less exactly everything you listed will be in 2.90, which will be out in a few days.

Some more improvements related to what you've mentioned are already in the master for Blender but won't be pushed out until Blender 2.91 (November), so it's only a matter of time until they arrive.

Good times ahead.

Personally I'm hoping they'll have some time to look at baking and texture painting, those are the areas I'd love to see some attention.

View this comment - View article - View full comments

Blender is doing so well lately that it's starting to lead rather than follow. A great example is Blender's viewport, with Autodesk quickly following Blender's 2.80 update with similar functionality being added to their software. Or Blender's cloth brush, shortly after it was shown off, Pixologic immediately began work on creating something similar to add to ZBrush.

Ubisoft and Tangent Labs came good on open sourcing some of the plugins for Blender they've been working on internally. Ubisoft open sourced an addon for multiple users to collaborate on a Blender scene together, Tangent Labs just open sourced a plugin to make Blender's Cycles rendering engine compatible with Hydra, and showed a demo of the Cycles rendering engine working in Houdini's viewport.

Now with this milestone reached, Blender Foundation Dev Fund is up to $121,931 USD/month and soon will have 20 full time developers working on it. Which means it's only going to improve even faster than before.

The industry is waking up and realising how much they have to gain from supporting Blender. It's going to put Autodesk and other commercial players in the industry on their toes and force them to cut prices while ramping up their development efforts just to keep up with Blender.

This is a great example of open source simply working for everyone.
Well, everyone except Autodesk. :wink:

View this comment - View article - View full comments

Nothing but Fall Guys.

It's so weirdly addictive.

Aside from that, do Godot and Blender count? I mean, they're on Steam. :tongue:

View this comment - View article - View full comments

Quoting: CreakAs a developer, I understand the urge to own and control your own code but, to make a poor analogy, developing your own game editor to make your game is a bit like developing your own Photoshop to create your textures.

No no, that is an EXCELLENT analogy.

I say that as someone who once well victim to it myself.

These days for modern 3D games you need things like inverse kinematics for positioning hands on weapons, tools for generating navigation meshes from level layouts, lightmap bakers and global illumination solutions, soft body simulation for character hair or cloth to move in the wind, ability to stream in level content in separate threads, and of course if you were writing an engine from scratch today you would be using Vulkan..

.. the idea of making a game engine for a game in 2020 is madness unless all you're making is a simple 2D game. Even then I can't help but ask 'Why?'. What exactly do you think your 2D engine is going to do that some other game engine that has been worked on by more people for longer won't do?

View this comment - View article - View full comments

Does this put us any closer to a Proton update? Proton 5.14 for example?

View this comment - View article - View full comments

Oh that is damn cool. Awesome! And that puts Blender sooo close to reaching it's "20 full time developers" target too!

Blender is unstoppable! :grin:

View this comment - View article - View full comments

Thanks Valve! :grin:

View this comment - View article - View full comments