When Wed, Mar 4th 2020 until Fri, Mar 6th 2020

Looking forward to presenting vvvv at this years Ircam Forum. With Ircam being the birthplace of both Max and PD, which quite inspired vvvv initially, it will be a great honor to hopefully give some inspiration back.

Here is what it is generally about:

My talk is on Friday, March 6th, 11h45 - 12h15 and and here is the full list of this years speakers.

Hope to see some of you there!

Help!

One of the most requested features for VL ever since was the support for help-patches. Obviously they had to come back, but it took us a while to figure out how exactly. One of the brainstorming sessions at LINK was especially helpful to sort out the details. Thanks everyone for your input...

Help Patches

So for the user it is just as before: Press F1 on a node that you need help for. Now what happens is:

• If a help-patch is found, it is opened and a bubble indicates the node you were looking for
• If no help-patch was found, a bubble indicates that no help-patch was found

Help patch was found...

No help patch available...

In both cases, you can now optionally opt-in for more help: Click the bubble to open the help browser showing the nodes Node Info.

Node Info

Shows some info about the node plus 3 lists of how-to patches this node is used in:

• most relevant (the dedicated help-patch)
• still relevant
• list of other patches using this node (automatically generated)

This should increase the chance of finding good info about how a node can be used in different scenarios.

When "follow mode" is active, you can click around in the patch and always get the node info to the last selected node.

Help Flags

Creators of libraries who want to provide help patches now use help flags to indicate relevance for a node. Read all about using help flags in the graybook.

Available for testing in latest previews!

Fantastic!

is probably not exactly the word you'd use to describe vvvv's 2019 numbers. If you can't handle bumpy roads, please set your Damper to ~2s before reading on... Not sure what this is all about? Check out last year's numbers of 2018 to make yourself familiar with this vvvvery tradition. Then pull yourself together:

Webwise

While we still don't have any new numbers from vvvv.org itself we do have what the forum tracks. It shows us different statistics regarding user engagement over time that are generally, how can I put this... not improving recently. The most uplifting I found was this:

According to the description of the graph, this indicates your stickiness in % and 30% is to be aimed at. Even though overall fewer things are happening on the forum you keep sticking around. Who could blame you, after all, it is rather cozy around here. Here is another one that doesn't look all too bad:

Ignore the red ones (crawlers) but see the rather constant number of logged-in users (dark blue) and the curious spikes at the end of December from anonymous users (light blue). Are these already the harbingers of numbers to expect when we finally will have an official 1.0 of vvvv gamma out? (hums) Dreams are my reality...

Not really numbers, but still interesting, here is a listing of top forum search terms in order of your priority:

Downloads

* x86 and x64 combined

What can we say? Only 1 release was really lame. But was it? Ok, yes, it was, but the way I'd like to whitewash these numbers a bit is that they are only reflecting vvvv beta, when obviously most of our work (to some of your's concern) went into VL and vvvv gamma. And inconveniently we don't have any numbers for vvvv gamma downloads at all. But then again those wouldn't be high numbers also because as you know we're still keeping a rather low profile with vvvv gamma, mostly advertising it to you as long as it is still in preview. So anyway, nothing to be particularly proud of here. Moving on...

Licenses

Phew, not as bad as the declining number of downloads would have suggested. Certainly not an upwards trend but still the 4th best result so far. So let's see what this number is composed of:

Mkay, so according to the number of individual commercial users we're basically thrown back to 2012. Quite a bummer but to be expected when you put most of your resources into R&D for the shiny new product that just still isn't there yet to make its own money. We were certainly hoping to keep a few more of you with our advances in VL but we know that before making any new friends, VL needs to grow out of vvvv beta and get a 3d engine...

That was it? Far from! This year we held by far the largest number of workshops for onboarding new users. These included a series of 8 free introductory workshops and 8 full-day workshops on various topics in Berlin. Further, we toured Kiel, Linz, Timișoara and Moscow. Also, we held 10 meetups and 3 patching circles in Berlin, one meetup in Hamburg and one in Moscow. Finally, we even held a dedicated Teaching Patching mini-conference to better understand the needs for vvvv in education.

All these activities will see a continuation in 2020 and culminate in NODE20 in October. Until then, the road to NODE is filled with some precious gemstones:

• The first official release of vvvv gamma that you can finally own for money, already available for preview
• Initial release of 3d engine for vvvv gamma, still work-in-progress
• Polished libraries to support your basic needs
• More workshops and learning opportunities

Is 2020 finally the year of the future of visual programming or is the world not ready for VL yet? Stay tuned, spread the word and become part of a truly wonderful new numbers blog post next year! You know what to do...

We wish you what!
Yours,
vvvv.

Happy new year, evvvveryone! Welcome back and we hope you had some fine holidays. More than one year has passed since the last report. An update is well overdue. This post sums up last year’s (2019) progress on our upcoming 3d engine for vvvv gamma.

Xenko’s codebase is huge and even after one year, we are still in explorer mode. But help is near! As you might have read, Xenko’s lead developer Virgile Bello aka xen2 is now with us in Berlin. While his main focus is maintaining Xenko and building a community around it, he also helps us to achieve our goals faster by answering questions and fixing bugs that we encounter.

As it is still early, for each technology we work on, we try to focus on creating only the basic nodes required to get a working prototype. Because if we have to change something fundamentally, every node has to be adapted. Nevertheless, there are tons of nodes already.

Besides many little things, here are the main topics we worked on:

Scene Graph

The scene graph is the fundamental data structure of the 3d engine. It contains the entities and scenes of your application. Most engines use a tree view to visualize and work with it. In vvvv, we have “Group” nodes to build the parent-child relations of entities.

For convenience, every group and entity got a transformation and name input. The name becomes handy for finding and accessing certain entities later programmatically or just for getting an overview. It also shows up in the tooltip.

Scene Graph Explorer

The scene graph can become large and complex quite quickly. Especially when you are a brave patcher and you use sub-patches to stay organized. Sometimes you want an overview of the whole tree structure and see basic properties (at runtime, of course!). Luckily there was an older project for Xenko that traverses the entity tree and visualizes its components and properties.

We updated it to the latest Xenko version and imported it as a debug helper.

Materials

If you look at the options and properties of a material in Xenko it’s quite complex and some features require expert knowledge. To make things quick and easy, we made a few simple nodes to cover the basics.

We’ll also add texture-based versions of them and for more complex setups, you can always design the material in Game Studio and import it.

Runtime Asset Loading

For almost every project you have to import content. Be it images, videos, models or just text files. This sounds trivial at first, but as you come closer to deployment, it gets more tricky to have everything at the right place. Even more so, with the new executable export in vvvv gamma.

Game engines usually solve it like this: Resource files like models, textures, etc. get imported via the editor and compiled to runtime assets. These asset files get placed next to the executable on compilation. This works well for static assets, but for rapid prototyping and changes at runtime, we need something more flexible. So we took the Xenko asset builder and adapted it to work at runtime. This allows for importing and building assets while the application is running.

File Assets

Load single resource files directly from disk, aka FileTexture. Good thing is, this works together with Xenko’s asset database and caches already loaded files. So everything is pooled by default.

Load Project

That’s a nice one, you can create a Xenko project in Game Studio, work on Assets, Scenes, Prefabs, etc and import everything at once with the LoadProject node. It even updates on save!

Modify Imported Entities

Since we work directly with Xenko’s data types, imported entities are in no way different from patched ones. You can access and modify them in the same way:

Shaders

vvvv is known for treating shaders first class. The Xenko team seemed to have a similar attitude. They created their own shader language called XKSL. It is a superset of HLSL, so you can paste shader code from vvvv into an xksl file.

However, there is much more. They added quite unique features that make writing shaders easy, fast and clean. If that wasn’t enough, you can compose shaders with other shaders at runtime and avoid an explosion of shader permutations.

TextureFX

The well known TextureFX node-set in vvvv beta makes working with textures easy and fun. Of course, we want to have the same experience in vvvv gamma. But, with preview in the tooltip! There you go:

ShaderFX + FieldTrip

Visual shader programming with nodes? Yes, please! The composition feature of XKSL makes this easier than expected. Turns out, the Xenko team was working on something similar which made it relatively easy to build a node-set around it.

To test these nodes, we had the pleasure to have Kyle McLean in our office for some days and worked with him on porting FieldTrip nodes to VL.Xenko.

FieldTrip is a modular shader library Kyle McLean created for vvvv beta. It includes raymarching and general tools to work with scalar fields and vector fields on the GPU.

We ended up with a working prototype, as you can see in this video from the vvvv berlin meetup:

StreamOut Instancer

Since Xenko has no native instancing for their material shaders yet, we tried to use stream out aka transform feedback to transform a mesh multiple times and build one big vertex buffer that works with the Xenko material pipeline. The node works nicely and takes a Mesh, a material, and the instance transformations. You can easily draw a few thousand instances with it. The header image was mostly done with it.

Profiler

The bigger your project gets, the more important it is to know how well your shaders perform. Thanks to Xenko’s profiler infrastructure we can give every shader a profiler key and see it listed in the GPU page:

Next Up

There are still some topics left in order to make VL.Xenko as convenient as we like it to be.

Multiple Windows

One major feature that vvvv beta does exceptionally well is working with multiple windows. Games are usually focused on one output window. In order to get this to work, we have to find a solid way to instantiate multiple windows in Xenko and render content into them. This requires some coding behind the scenes. We are planning to get this done by early 2020.

One-Click Installation

At the moment one still needs Visual Studio to set up a VL.Xenko project. This should, of course, be just a one-click package installation as we have it with other VL libraries. Initial tests went well and we are confident that this will work out just fine.

Rename to Stride

Because of legal reasons about the trademark, Xenko has to do a re-branding and will be called Stride soon. We'll of course follow along and will rename the library to VL.Stride.

If you want to know more about something specific or have any remarks or requests, don't hesitate to leave a comment below.

That’s it for now, looking forward to NODE20 and more exciting patches in 2020.

yours,
devvvvs