pack vl rendering GUI dx11
Credits: vux, flateric
The nodes resemble the beloved vvvv drawing nodes and will make you feel at home immediately. CraftLie creates a list of draw descriptions and sends them over one output pin to vvvv. In vvvv a special node takes the descriptions and renders them into a Renderer (DX11) of your choice.
No output pin mayhem and much better performance if you want to render complex scenes from VL.
Placing the drawing nodes inside your VL data types makes much cleaner patches and shows the joy of object oriented patching. This makes it perfectly suited for UIs and other drawing tasks that have complicated visual structures and need dynamic instances.
Keep in mind that the rendering still takes place in vvvv. So you can use your vvvv shader knowledge to customize the Renderer (Craftlie) and draw the shapes in a different style. It's also possible to output many layers from VL and draw each with a different custom Renderer (CraftLie).
CraftLie is, like FeralTic the dx11 library it is based on, and anagram of the major dx11 developer @flateric. Due to the fake nature of the drawing, the name fits perfectly.
Check the folder CraftLie.1.x.x\vvvv\girlpower for:
1. An overview patch which shows all nodes in action
2. How to use instancing
3. Interactive particle patch by @sebl
4. CPU core race which stresses all cores using reactive nodes
5. Sprites particles in perlin force field
6. Dynamic UI demo that creates layered windows
wow, what's happening...checked this late last night and thought I was already in some delusional dream state ;) great contribution!
haha, didn't want to spoil all the fun in the description... ;) thanks for testing it.
This is fascinating demo of VL, but looking through the patch, to me, it is almost incomprehensible, some VL for dummy notes would be appreciated. The gui looks to be very useful, but how to add elements to it?
I still find VL to be as transparent as mud, as poor as I am at code, it seems to make more sense, which isn't the expected result of a node based GUI, surely the data flow should be visible, it feels like the part that makes sense is somehow hidden from view and you are left with magic nodes that do something secret, maybe its just me?
that is basically why you (and every other vvvv pro) should attend this:
anyway here is a tip on how to read the DynamicUI example:
it basically follows the super classic game development pattern for mainloops. as a vvvv developer you live in that world and you only need to understand one addition to the normal 'update' processing. the mainloop is split, or sliced, into several steps in the main patch:
1. check if there are any new input events
2. process all input events and update the state/logic of all elements accordingly
3. do normal "animation updates" like you are used to in vvvv
4. output layer to output pins and render it
> start over
step 2. is the new one for a vvvv developer and itself split into two steps, HitTest to check whether an element actually needs to process the input and ProcessEvents then does the actual logic and leaves the results for the Update stored in pads. so basically the pads connect the different mainloop slices by moving data/results from ProcessEvents to Update to animate/draw it...
so go into the patches and read them like this, someone from the outside is calling HitTest, if that returns true ProcessEvents will be called where the patch processes the UI logic. finally Update will be called to render animations and other stuff that was decided on ProcessEvents.
anonymous user login