Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Don't render children of entities with is_visible set to false #865

Closed
wants to merge 4 commits into from
Closed

Don't render children of entities with is_visible set to false #865

wants to merge 4 commits into from

Conversation

ambeeeeee
Copy link
Contributor

Created in response to #838, would likely need to be manually merged with #861 due to changing the structure of the file that PR modifies.

@karroffel karroffel added A-ECS Entities, components, systems, and events C-Feature A new feature, making something new possible A-Rendering Drawing game state to the screen labels Nov 14, 2020
@ambeeeeee
Copy link
Contributor Author

ambeeeeee commented Nov 15, 2020

Currently looking into leveraging Changed<T> in the visible checking system.

@CleanCut
Copy link
Member

Since I was doing this manually here and here I thought I'd give this a try (you need to connect at least one gamepad if you want to try it out). Unfortunately, it doesn't seem to work for me!

I did a quick read-through of the PR, and nothing jumps out as obviously wrong to me, but I haven't looked closely yet.

Copy link
Member

@CleanCut CleanCut left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LOL. Nevermind. It was me. I changed the bevy source just before I stashed my local changes, so my bevy wasn't pointing to this PR during my testing. Once I point to the right place, it works great.

I read through the PR, and it all seems correct to me. It seems a bit verbose, especially in the number of arguments to the recursive functions...but I couldn't think of a decent way to simplify it. Perhaps someone else will see a way to simplify it.

@ambeeeeee
Copy link
Contributor Author

ambeeeeee commented Nov 16, 2020

Since I was doing this manually here and here I thought I'd give this a try (you need to connect at least one gamepad if you want to try it out). Unfortunately, it doesn't seem to work for me!

I did a quick read-through of the PR, and nothing jumps out as obviously wrong to me, but I haven't looked closely yet.

A big goal for me with this PR is not to modify the is_visible of the children, which you obviously had to do, due to the feature not existing in bevy itself. I'm happy with how it came out and I hope it's satisfactory for cart :)

@AlexCMueller
Copy link

AlexCMueller commented Nov 16, 2020

I don't have a good understanding of the engine's internals so I don't know how feasible my first thought of approaching this would be, but I notice that this solution requires some recursion while mine wouldn't if it works: would it be possible to, for each entity, determine if it should be drawn by searching up through the parents until it either reaches the root entity or finds an entity with is_visible set to false?

something along the lines of

let mut is_visible = true;

while let Some(parent) = entity.get_component::<Parent>() {
    if let Some(draw) = parent.get_component::<Draw>() {
        if !draw.is_visible {
            is_visible = false;
            break;
        }
    }
}

if is_visible {
    // drawing code here
}

does this have performance issues relative to amber's solution that I'm not seeing? unexpected behaviors regarding entities that should be drawn but aren't? etc...

EDIT: very big performance issue, as written this is an infinite loop...

@ambeeeeee
Copy link
Contributor Author

ambeeeeee commented Nov 16, 2020

Searching up through parents would most likely be less performant in general.

I based my solution off of the transform_propagate_system, so barring any horrific performance issues I didn't notice, I believe it would be acceptable due to previous use of recursion within bevy. Additionally, when I hit an entity with is_visible set to false I stop iterating on its children since it no longer matters what their values are.

@AlexCMueller
Copy link

I'm starting to see the advantages of a recursive parent-to-children approach, especially in my specific use case where I have a single root node holding potentially hundreds of children sprites. Thanks for taking on this issue!

@cart
Copy link
Member

cart commented Nov 17, 2020

Just pushed a slight refactor that reduces the overall size/complexity of the implementation. In general this looks good, but we do have a problem that will block merging: we now only draw hierarchies like this:

e1 (has Draw) -> e2 (has Draw)  -> e3 (has Draw) 

We no longer draw cases like these:

// e3 is not drawn
e1 (has Draw) -> e2 (no Draw)  -> e3 (has Draw) 

// e2 and e3 are not drawn
e1 (no Draw) -> e2 (has Draw)  -> e3 (has Draw)

This breaks the assumptions we make when loading GLTF files as scenes (note that the load_gltf.rs example no longer draws anything)

// e1 is generated from a GLTF node that isn't intended to be "drawn", only to transform its children
e1 (Transform, GlobalTransform) -> e2 (PbrComponents { Draw, Transform, GlobalTransform, ...})

We have options here:

  1. I think my ideal behavior would be that we dont require a "Draw" chain and just implicitly treat entities without Draw as "visible" / still attempt to draw their children. Unfortunately if we use the approach in this pr, I think it would require removing the With<Draw> filter and walking the entire entity tree / checking component existence at each entity. The alternative is querying every drawn entity and then "walking up" the tree to determine if that entity should be drawn. A naive implementation would be extremely expensive and almost certainly worse than just visiting every entity in most cases. With memoization we could cut down on the compute cost, but with O(n) memory usage (where "n" is the number of entities in World).
  2. In cases like the GLTF example, we could require that Draw components are inserted.

@ambeeeeee
Copy link
Contributor Author

Through discussion on discord we have analyzed the issue and come to the conclusion it needs to be re-implemented. I'm closing this PR in favor of a future one coming in the next day or two from me, as this one's git history isn't entirely relevant anymore.

Relevant discord link.

@ambeeeeee ambeeeeee closed this Nov 18, 2020
bors bot pushed a commit that referenced this pull request Jul 15, 2022
…support (#5310)

# Objective

Fixes #4907. Fixes #838. Fixes #5089.
Supersedes #5146. Supersedes #2087. Supersedes #865. Supersedes #5114

Visibility is currently entirely local. Set a parent entity to be invisible, and the children are still visible. This makes it hard for users to hide entire hierarchies of entities.

Additionally, the semantics of `Visibility` vs `ComputedVisibility` are inconsistent across entity types. 3D meshes use `ComputedVisibility` as the "definitive" visibility component, with `Visibility` being just one data source. Sprites just use `Visibility`, which means they can't feed off of `ComputedVisibility` data, such as culling information, RenderLayers, and (added in this pr) visibility inheritance information.

## Solution

Splits `ComputedVisibilty::is_visible` into `ComputedVisibilty::is_visible_in_view` and `ComputedVisibilty::is_visible_in_hierarchy`. For each visible entity, `is_visible_in_hierarchy` is computed by propagating visibility down the hierarchy. The `ComputedVisibility::is_visible()` function combines these two booleans for the canonical "is this entity visible" function.

Additionally, all entities that have `Visibility` now also have `ComputedVisibility`.  Sprites, Lights, and UI entities now use `ComputedVisibility` when appropriate.

This means that in addition to visibility inheritance, everything using Visibility now also supports RenderLayers. Notably, Sprites (and other 2d objects) now support `RenderLayers` and work properly across multiple views.

Also note that this does increase the amount of work done per sprite. Bevymark with 100,000 sprites on `main` runs in `0.017612` seconds and this runs in `0.01902`. That is certainly a gap, but I believe the api consistency and extra functionality this buys us is worth it. See [this thread](#5146 (comment)) for more info. Note that #5146 in combination with #5114 _are_ a viable alternative to this PR and _would_ perform better, but that comes at the cost of api inconsistencies and doing visibility calculations in the "wrong" place. The current visibility system does have potential for performance improvements. I would prefer to evolve that one system as a whole rather than doing custom hacks / different behaviors for each feature slice.

Here is a "split screen" example where the left camera uses RenderLayers to filter out the blue sprite.

![image](https://user-images.githubusercontent.com/2694663/178814868-2e9a2173-bf8c-4c79-8815-633899d492c3.png)


Note that this builds directly on #5146 and that @james7132 deserves the credit for the baseline visibility inheritance work. This pr moves the inherited visibility field into `ComputedVisibility`, then does the additional work of porting everything to `ComputedVisibility`. See my [comments here](#5146 (comment)) for rationale. 

## Follow up work

* Now that lights use ComputedVisibility, VisibleEntities now includes "visible lights" in the entity list. Functionally not a problem as we use queries to filter the list down in the desired context. But we should consider splitting this out into a separate`VisibleLights` collection for both clarity and performance reasons. And _maybe_ even consider scoping `VisibleEntities` down to `VisibleMeshes`?.
* Investigate alternative sprite rendering impls (in combination with visibility system tweaks) that avoid re-generating a per-view fixedbitset of visible entities every frame, then checking each ExtractedEntity. This is where most of the performance overhead lives. Ex: we could generate ExtractedEntities per-view using the VisibleEntities list, avoiding the need for the bitset.
* Should ComputedVisibility use bitflags under the hood? This would cut down on the size of the component, potentially speed up the `is_visible()` function, and allow us to cheaply expand ComputedVisibility with more data (ex: split out local visibility and parent visibility, add more culling classes, etc).
---

## Changelog

* ComputedVisibility now takes hierarchy visibility into account.
* 2D, UI and Light entities now use the ComputedVisibility component.

## Migration Guide

If you were previously reading `Visibility::is_visible` as the "actual visibility" for sprites or lights, use `ComputedVisibilty::is_visible()` instead:

```rust
// before (0.7)
fn system(query: Query<&Visibility>) {
  for visibility in query.iter() {
    if visibility.is_visible {
       log!("found visible entity");
    }
  }
}

// after (0.8)
fn system(query: Query<&ComputedVisibility>) {
  for visibility in query.iter() {
    if visibility.is_visible() {
       log!("found visible entity");
    }
  }
}
``` 


Co-authored-by: Carter Anderson <[email protected]>
inodentry pushed a commit to IyesGames/bevy that referenced this pull request Aug 8, 2022
…support (bevyengine#5310)

# Objective

Fixes bevyengine#4907. Fixes bevyengine#838. Fixes bevyengine#5089.
Supersedes bevyengine#5146. Supersedes bevyengine#2087. Supersedes bevyengine#865. Supersedes bevyengine#5114

Visibility is currently entirely local. Set a parent entity to be invisible, and the children are still visible. This makes it hard for users to hide entire hierarchies of entities.

Additionally, the semantics of `Visibility` vs `ComputedVisibility` are inconsistent across entity types. 3D meshes use `ComputedVisibility` as the "definitive" visibility component, with `Visibility` being just one data source. Sprites just use `Visibility`, which means they can't feed off of `ComputedVisibility` data, such as culling information, RenderLayers, and (added in this pr) visibility inheritance information.

## Solution

Splits `ComputedVisibilty::is_visible` into `ComputedVisibilty::is_visible_in_view` and `ComputedVisibilty::is_visible_in_hierarchy`. For each visible entity, `is_visible_in_hierarchy` is computed by propagating visibility down the hierarchy. The `ComputedVisibility::is_visible()` function combines these two booleans for the canonical "is this entity visible" function.

Additionally, all entities that have `Visibility` now also have `ComputedVisibility`.  Sprites, Lights, and UI entities now use `ComputedVisibility` when appropriate.

This means that in addition to visibility inheritance, everything using Visibility now also supports RenderLayers. Notably, Sprites (and other 2d objects) now support `RenderLayers` and work properly across multiple views.

Also note that this does increase the amount of work done per sprite. Bevymark with 100,000 sprites on `main` runs in `0.017612` seconds and this runs in `0.01902`. That is certainly a gap, but I believe the api consistency and extra functionality this buys us is worth it. See [this thread](bevyengine#5146 (comment)) for more info. Note that bevyengine#5146 in combination with bevyengine#5114 _are_ a viable alternative to this PR and _would_ perform better, but that comes at the cost of api inconsistencies and doing visibility calculations in the "wrong" place. The current visibility system does have potential for performance improvements. I would prefer to evolve that one system as a whole rather than doing custom hacks / different behaviors for each feature slice.

Here is a "split screen" example where the left camera uses RenderLayers to filter out the blue sprite.

![image](https://user-images.githubusercontent.com/2694663/178814868-2e9a2173-bf8c-4c79-8815-633899d492c3.png)


Note that this builds directly on bevyengine#5146 and that @james7132 deserves the credit for the baseline visibility inheritance work. This pr moves the inherited visibility field into `ComputedVisibility`, then does the additional work of porting everything to `ComputedVisibility`. See my [comments here](bevyengine#5146 (comment)) for rationale. 

## Follow up work

* Now that lights use ComputedVisibility, VisibleEntities now includes "visible lights" in the entity list. Functionally not a problem as we use queries to filter the list down in the desired context. But we should consider splitting this out into a separate`VisibleLights` collection for both clarity and performance reasons. And _maybe_ even consider scoping `VisibleEntities` down to `VisibleMeshes`?.
* Investigate alternative sprite rendering impls (in combination with visibility system tweaks) that avoid re-generating a per-view fixedbitset of visible entities every frame, then checking each ExtractedEntity. This is where most of the performance overhead lives. Ex: we could generate ExtractedEntities per-view using the VisibleEntities list, avoiding the need for the bitset.
* Should ComputedVisibility use bitflags under the hood? This would cut down on the size of the component, potentially speed up the `is_visible()` function, and allow us to cheaply expand ComputedVisibility with more data (ex: split out local visibility and parent visibility, add more culling classes, etc).
---

## Changelog

* ComputedVisibility now takes hierarchy visibility into account.
* 2D, UI and Light entities now use the ComputedVisibility component.

## Migration Guide

If you were previously reading `Visibility::is_visible` as the "actual visibility" for sprites or lights, use `ComputedVisibilty::is_visible()` instead:

```rust
// before (0.7)
fn system(query: Query<&Visibility>) {
  for visibility in query.iter() {
    if visibility.is_visible {
       log!("found visible entity");
    }
  }
}

// after (0.8)
fn system(query: Query<&ComputedVisibility>) {
  for visibility in query.iter() {
    if visibility.is_visible() {
       log!("found visible entity");
    }
  }
}
``` 


Co-authored-by: Carter Anderson <[email protected]>
james7132 added a commit to james7132/bevy that referenced this pull request Oct 28, 2022
…support (bevyengine#5310)

# Objective

Fixes bevyengine#4907. Fixes bevyengine#838. Fixes bevyengine#5089.
Supersedes bevyengine#5146. Supersedes bevyengine#2087. Supersedes bevyengine#865. Supersedes bevyengine#5114

Visibility is currently entirely local. Set a parent entity to be invisible, and the children are still visible. This makes it hard for users to hide entire hierarchies of entities.

Additionally, the semantics of `Visibility` vs `ComputedVisibility` are inconsistent across entity types. 3D meshes use `ComputedVisibility` as the "definitive" visibility component, with `Visibility` being just one data source. Sprites just use `Visibility`, which means they can't feed off of `ComputedVisibility` data, such as culling information, RenderLayers, and (added in this pr) visibility inheritance information.

## Solution

Splits `ComputedVisibilty::is_visible` into `ComputedVisibilty::is_visible_in_view` and `ComputedVisibilty::is_visible_in_hierarchy`. For each visible entity, `is_visible_in_hierarchy` is computed by propagating visibility down the hierarchy. The `ComputedVisibility::is_visible()` function combines these two booleans for the canonical "is this entity visible" function.

Additionally, all entities that have `Visibility` now also have `ComputedVisibility`.  Sprites, Lights, and UI entities now use `ComputedVisibility` when appropriate.

This means that in addition to visibility inheritance, everything using Visibility now also supports RenderLayers. Notably, Sprites (and other 2d objects) now support `RenderLayers` and work properly across multiple views.

Also note that this does increase the amount of work done per sprite. Bevymark with 100,000 sprites on `main` runs in `0.017612` seconds and this runs in `0.01902`. That is certainly a gap, but I believe the api consistency and extra functionality this buys us is worth it. See [this thread](bevyengine#5146 (comment)) for more info. Note that bevyengine#5146 in combination with bevyengine#5114 _are_ a viable alternative to this PR and _would_ perform better, but that comes at the cost of api inconsistencies and doing visibility calculations in the "wrong" place. The current visibility system does have potential for performance improvements. I would prefer to evolve that one system as a whole rather than doing custom hacks / different behaviors for each feature slice.

Here is a "split screen" example where the left camera uses RenderLayers to filter out the blue sprite.

![image](https://user-images.githubusercontent.com/2694663/178814868-2e9a2173-bf8c-4c79-8815-633899d492c3.png)


Note that this builds directly on bevyengine#5146 and that @james7132 deserves the credit for the baseline visibility inheritance work. This pr moves the inherited visibility field into `ComputedVisibility`, then does the additional work of porting everything to `ComputedVisibility`. See my [comments here](bevyengine#5146 (comment)) for rationale. 

## Follow up work

* Now that lights use ComputedVisibility, VisibleEntities now includes "visible lights" in the entity list. Functionally not a problem as we use queries to filter the list down in the desired context. But we should consider splitting this out into a separate`VisibleLights` collection for both clarity and performance reasons. And _maybe_ even consider scoping `VisibleEntities` down to `VisibleMeshes`?.
* Investigate alternative sprite rendering impls (in combination with visibility system tweaks) that avoid re-generating a per-view fixedbitset of visible entities every frame, then checking each ExtractedEntity. This is where most of the performance overhead lives. Ex: we could generate ExtractedEntities per-view using the VisibleEntities list, avoiding the need for the bitset.
* Should ComputedVisibility use bitflags under the hood? This would cut down on the size of the component, potentially speed up the `is_visible()` function, and allow us to cheaply expand ComputedVisibility with more data (ex: split out local visibility and parent visibility, add more culling classes, etc).
---

## Changelog

* ComputedVisibility now takes hierarchy visibility into account.
* 2D, UI and Light entities now use the ComputedVisibility component.

## Migration Guide

If you were previously reading `Visibility::is_visible` as the "actual visibility" for sprites or lights, use `ComputedVisibilty::is_visible()` instead:

```rust
// before (0.7)
fn system(query: Query<&Visibility>) {
  for visibility in query.iter() {
    if visibility.is_visible {
       log!("found visible entity");
    }
  }
}

// after (0.8)
fn system(query: Query<&ComputedVisibility>) {
  for visibility in query.iter() {
    if visibility.is_visible() {
       log!("found visible entity");
    }
  }
}
``` 


Co-authored-by: Carter Anderson <[email protected]>
ItsDoot pushed a commit to ItsDoot/bevy that referenced this pull request Feb 1, 2023
…support (bevyengine#5310)

# Objective

Fixes bevyengine#4907. Fixes bevyengine#838. Fixes bevyengine#5089.
Supersedes bevyengine#5146. Supersedes bevyengine#2087. Supersedes bevyengine#865. Supersedes bevyengine#5114

Visibility is currently entirely local. Set a parent entity to be invisible, and the children are still visible. This makes it hard for users to hide entire hierarchies of entities.

Additionally, the semantics of `Visibility` vs `ComputedVisibility` are inconsistent across entity types. 3D meshes use `ComputedVisibility` as the "definitive" visibility component, with `Visibility` being just one data source. Sprites just use `Visibility`, which means they can't feed off of `ComputedVisibility` data, such as culling information, RenderLayers, and (added in this pr) visibility inheritance information.

## Solution

Splits `ComputedVisibilty::is_visible` into `ComputedVisibilty::is_visible_in_view` and `ComputedVisibilty::is_visible_in_hierarchy`. For each visible entity, `is_visible_in_hierarchy` is computed by propagating visibility down the hierarchy. The `ComputedVisibility::is_visible()` function combines these two booleans for the canonical "is this entity visible" function.

Additionally, all entities that have `Visibility` now also have `ComputedVisibility`.  Sprites, Lights, and UI entities now use `ComputedVisibility` when appropriate.

This means that in addition to visibility inheritance, everything using Visibility now also supports RenderLayers. Notably, Sprites (and other 2d objects) now support `RenderLayers` and work properly across multiple views.

Also note that this does increase the amount of work done per sprite. Bevymark with 100,000 sprites on `main` runs in `0.017612` seconds and this runs in `0.01902`. That is certainly a gap, but I believe the api consistency and extra functionality this buys us is worth it. See [this thread](bevyengine#5146 (comment)) for more info. Note that bevyengine#5146 in combination with bevyengine#5114 _are_ a viable alternative to this PR and _would_ perform better, but that comes at the cost of api inconsistencies and doing visibility calculations in the "wrong" place. The current visibility system does have potential for performance improvements. I would prefer to evolve that one system as a whole rather than doing custom hacks / different behaviors for each feature slice.

Here is a "split screen" example where the left camera uses RenderLayers to filter out the blue sprite.

![image](https://user-images.githubusercontent.com/2694663/178814868-2e9a2173-bf8c-4c79-8815-633899d492c3.png)


Note that this builds directly on bevyengine#5146 and that @james7132 deserves the credit for the baseline visibility inheritance work. This pr moves the inherited visibility field into `ComputedVisibility`, then does the additional work of porting everything to `ComputedVisibility`. See my [comments here](bevyengine#5146 (comment)) for rationale. 

## Follow up work

* Now that lights use ComputedVisibility, VisibleEntities now includes "visible lights" in the entity list. Functionally not a problem as we use queries to filter the list down in the desired context. But we should consider splitting this out into a separate`VisibleLights` collection for both clarity and performance reasons. And _maybe_ even consider scoping `VisibleEntities` down to `VisibleMeshes`?.
* Investigate alternative sprite rendering impls (in combination with visibility system tweaks) that avoid re-generating a per-view fixedbitset of visible entities every frame, then checking each ExtractedEntity. This is where most of the performance overhead lives. Ex: we could generate ExtractedEntities per-view using the VisibleEntities list, avoiding the need for the bitset.
* Should ComputedVisibility use bitflags under the hood? This would cut down on the size of the component, potentially speed up the `is_visible()` function, and allow us to cheaply expand ComputedVisibility with more data (ex: split out local visibility and parent visibility, add more culling classes, etc).
---

## Changelog

* ComputedVisibility now takes hierarchy visibility into account.
* 2D, UI and Light entities now use the ComputedVisibility component.

## Migration Guide

If you were previously reading `Visibility::is_visible` as the "actual visibility" for sprites or lights, use `ComputedVisibilty::is_visible()` instead:

```rust
// before (0.7)
fn system(query: Query<&Visibility>) {
  for visibility in query.iter() {
    if visibility.is_visible {
       log!("found visible entity");
    }
  }
}

// after (0.8)
fn system(query: Query<&ComputedVisibility>) {
  for visibility in query.iter() {
    if visibility.is_visible() {
       log!("found visible entity");
    }
  }
}
``` 


Co-authored-by: Carter Anderson <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
A-ECS Entities, components, systems, and events A-Rendering Drawing game state to the screen C-Feature A new feature, making something new possible
Projects
None yet
Development

Successfully merging this pull request may close these issues.

5 participants