doup
@aillarra.bsky.social
I'm toying with a little optimization for my #Bevy entity inspector: marker components now show at the top as pills.
The check is naive and has false positives (e.g. Frustum and VisibleEntities). I can improve it slightly but there is not much more room with the current BRP spec/implementation.
The check is naive and has false positives (e.g. Frustum and VisibleEntities). I can improve it slightly but there is not much more room with the current BRP spec/implementation.
May 25, 2025 at 6:54 PM
I'm toying with a little optimization for my #Bevy entity inspector: marker components now show at the top as pills.
The check is naive and has false positives (e.g. Frustum and VisibleEntities). I can improve it slightly but there is not much more room with the current BRP spec/implementation.
The check is naive and has false positives (e.g. Frustum and VisibleEntities). I can improve it slightly but there is not much more room with the current BRP spec/implementation.
Ideally the only extra code should be to improve the UX, not because of misalignment between the value and the schema. For example, "Node" could use a custom widget like this for the Flex values:
May 23, 2025 at 7:15 PM
Ideally the only extra code should be to improve the UX, not because of misalignment between the value and the schema. For example, "Node" could use a custom widget like this for the Flex values:
After taking a break from BiRP (a #Bevy BRP inspector) for a couple of weeks, today I tried components editing. I got something basic working (styling will come later).
But I'm frustrated that the returned values and schema don't align, making it impossible to create something generic. 👿
But I'm frustrated that the returned values and schema don't align, making it impossible to create something generic. 👿
May 23, 2025 at 7:15 PM
After taking a break from BiRP (a #Bevy BRP inspector) for a couple of weeks, today I tried components editing. I got something basic working (styling will come later).
But I'm frustrated that the returned values and schema don't align, making it impossible to create something generic. 👿
But I'm frustrated that the returned values and schema don't align, making it impossible to create something generic. 👿
🚀 Announcing BiRP! My *read-only* #Bevy BRP Inspector is now on GitHub.
BiRP lets you inspect your game entities, components, and resources using the Bevy Remote Protocol.
Hope you find it useful! (Ah! and don't miss the mascot! 😄)
Get it here: github.com/doup/birp
BiRP lets you inspect your game entities, components, and resources using the Bevy Remote Protocol.
Hope you find it useful! (Ah! and don't miss the mascot! 😄)
Get it here: github.com/doup/birp
May 9, 2025 at 7:38 PM
🚀 Announcing BiRP! My *read-only* #Bevy BRP Inspector is now on GitHub.
BiRP lets you inspect your game entities, components, and resources using the Bevy Remote Protocol.
Hope you find it useful! (Ah! and don't miss the mascot! 😄)
Get it here: github.com/doup/birp
BiRP lets you inspect your game entities, components, and resources using the Bevy Remote Protocol.
Hope you find it useful! (Ah! and don't miss the mascot! 😄)
Get it here: github.com/doup/birp
Some more progress with the shading. The normal map is back! The skin color is configurable via the extended material and I've been toying with more vibrant colors for the rim lights.
Oh! And I remade the topology of the shoulder area which has improved the deformation… 👌
#bevy
Oh! And I remade the topology of the shoulder area which has improved the deformation… 👌
#bevy
February 20, 2025 at 3:02 PM
Some more progress with the shading. The normal map is back! The skin color is configurable via the extended material and I've been toying with more vibrant colors for the rim lights.
Oh! And I remade the topology of the shoulder area which has improved the deformation… 👌
#bevy
Oh! And I remade the topology of the shoulder area which has improved the deformation… 👌
#bevy
I've removed the (directional) rim light, and instead I'm applying the rim light in the character shader. The pros are that the ground doesn't get weird reflections anymore and each character rim light has a different hue.
Now… why is the normal map broken? 😂
#bevy
Now… why is the normal map broken? 😂
#bevy
February 17, 2025 at 7:53 PM
I've removed the (directional) rim light, and instead I'm applying the rim light in the character shader. The pros are that the ground doesn't get weird reflections anymore and each character rim light has a different hue.
Now… why is the normal map broken? 😂
#bevy
Now… why is the normal map broken? 😂
#bevy
It's working again, I'll fix tomorrow the models orientation (they should be facing each other).
I took another approach, instead of exporting the data from inside #blender (including coodrs conversion 😵💫), I've exported to GLTF and then extracted the animation/pose data from there. Muuuuch simpler.
I took another approach, instead of exporting the data from inside #blender (including coodrs conversion 😵💫), I've exported to GLTF and then extracted the animation/pose data from there. Muuuuch simpler.
February 12, 2025 at 5:40 PM
It's working again, I'll fix tomorrow the models orientation (they should be facing each other).
I took another approach, instead of exporting the data from inside #blender (including coodrs conversion 😵💫), I've exported to GLTF and then extracted the animation/pose data from there. Muuuuch simpler.
I took another approach, instead of exporting the data from inside #blender (including coodrs conversion 😵💫), I've exported to GLTF and then extracted the animation/pose data from there. Muuuuch simpler.
I had to rework the sculpt (for normal maps) and the weights for the skeleton, but know I'm starting to like it. I think I'll be back to the Bevy integration soon… maybe I'll give a try to Bevy Cobweb to improve the UI of my app. 🤔
(character color still pending…)
#blender #bevy
(character color still pending…)
#blender #bevy
January 27, 2025 at 7:30 PM
My #bevy project is slowly moving, I've my puppet rigged (I still need to tweak the bone weights) using #blender Rigify plugin. This is captured in Blender's viewport and it's using a kuwahara filter to give it a painterly effect.
Now I just have to import it in Bevy and do the same… 🙃
Now I just have to import it in Bevy and do the same… 🙃
January 14, 2025 at 5:07 PM
I've just copied the code I had for the "longest path" ranking algorithm for non-compound graphs (basic Sugiyama). Not bad for a first approximation, but I still need to separate top/bottom border and base nodes into separate ranks and apply "minimum length" and "weight" constraints.
December 22, 2024 at 8:01 PM
I've just copied the code I had for the "longest path" ranking algorithm for non-compound graphs (basic Sugiyama). Not bad for a first approximation, but I still need to separate top/bottom border and base nodes into separate ranks and apply "minimum length" and "weight" constraints.
This is the visualization of my test data (I've done the layout manually). In the previous image the dashed lines represent the "hierarchy", and the solid lines represent the "connections".
December 22, 2024 at 6:41 PM
This is the visualization of my test data (I've done the layout manually). In the previous image the dashed lines represent the "hierarchy", and the solid lines represent the "connections".
I've made tiny progress on the graph layout implementation, I'm following more closely the "Layout of Compound Directed Graphs" paper. Now I have the "extended nesting graph" (graph hierarchy + connections) data structure ready.
Note that for each sub-graph there is a top/bottom "border" node.
Note that for each sub-graph there is a top/bottom "border" node.
December 22, 2024 at 6:41 PM
I've made tiny progress on the graph layout implementation, I'm following more closely the "Layout of Compound Directed Graphs" paper. Now I have the "extended nesting graph" (graph hierarchy + connections) data structure ready.
Note that for each sub-graph there is a top/bottom "border" node.
Note that for each sub-graph there is a top/bottom "border" node.
The human puppet is slowly taking shape. 👯♂️
I've simplified some parts and I'm not worrying so much about topology (as it's not going to be subdivided). I'll probably simplify the lower leg a little bit more. But first I want to finish the forearm and hand.
#blender
I've simplified some parts and I'm not worrying so much about topology (as it's not going to be subdivided). I'll probably simplify the lower leg a little bit more. But first I want to finish the forearm and hand.
#blender
December 13, 2024 at 6:22 PM
The human puppet is slowly taking shape. 👯♂️
I've simplified some parts and I'm not worrying so much about topology (as it's not going to be subdivided). I'll probably simplify the lower leg a little bit more. But first I want to finish the forearm and hand.
#blender
I've simplified some parts and I'm not worrying so much about topology (as it's not going to be subdivided). I'll probably simplify the lower leg a little bit more. But first I want to finish the forearm and hand.
#blender
Behold a leg! I'm trying to do clean topology… although I'm not sure if it's that important for a game asset (I'm not going to subdivide it); I suppose it's more important to have enough mesh in the parts that are going to deform? 🤔
#blender
#blender
December 9, 2024 at 7:27 PM
Behold a leg! I'm trying to do clean topology… although I'm not sure if it's that important for a game asset (I'm not going to subdivide it); I suppose it's more important to have enough mesh in the parts that are going to deform? 🤔
#blender
#blender
I also need to compute the "nesting graph" which is like the "nesting tree" but it's mirrored over the leaf nodes. For each sub-graph (container boxes in the first image) there is a top and bottom node which represent the top/bottom borders of the container. ✅ Done.
(image generated with graphviz)
(image generated with graphviz)
December 2, 2024 at 6:12 PM
I also need to compute the "nesting graph" which is like the "nesting tree" but it's mirrored over the leaf nodes. For each sub-graph (container boxes in the first image) there is a top and bottom node which represent the top/bottom borders of the container. ✅ Done.
(image generated with graphviz)
(image generated with graphviz)
☝️ That's my test compound graph. So, before I can assign ranks (vertical positions) I need to compute a couple of intermediate structures.
The "nesting tree" for the hierarchy of the graph. Which also gives me the depth information for each node and sub-graph. ✅ Done.
The "nesting tree" for the hierarchy of the graph. Which also gives me the depth information for each node and sub-graph. ✅ Done.
December 2, 2024 at 6:12 PM
☝️ That's my test compound graph. So, before I can assign ranks (vertical positions) I need to compute a couple of intermediate structures.
The "nesting tree" for the hierarchy of the graph. Which also gives me the depth information for each node and sub-graph. ✅ Done.
The "nesting tree" for the hierarchy of the graph. Which also gives me the depth information for each node and sub-graph. ✅ Done.
Some progress today. I was porting Dagre but my test graph was panicking. I've tested the same graph in Dagre and I also get an error… so maybe there is an issue in their algorithm? Weird. Anyway, I went back to the "Layout of Compound Directed Graphs" paper and I'm following that for now.
December 2, 2024 at 6:12 PM
Some progress today. I was porting Dagre but my test graph was panicking. I've tested the same graph in Dagre and I also get an error… so maybe there is an issue in their algorithm? Weird. Anyway, I went back to the "Layout of Compound Directed Graphs" paper and I'm following that for now.