Enable weighted blending between calls to render #1148
Conversation
9919f2a to
2b28676
Compare
5cbdb41 to
9de8e76
Compare
|
I can see an argument that some might need to clear the weights between calls to renderer especially when they are using differerent cameras for an unrelated "scene". But i figured this would simulate a workflow that we often have where we "stream" in the data from a larger source and can't keep all the image data in GPU memory at the same time. |
9de8e76 to
4e1de2b
Compare
4e1de2b to
a7d4c3b
Compare
|
I think with my latest commit, the user has enough control in a fashion similar to other buffers to create complex effects with overlays. On the left is without the additions to The parameter names can probably use some work, but it seems pretty cool now. |
|
The triangle draw using the new weighted pipeline to highlight the fact hat you can fully utilize the weighted blending properties for any "scene", not just the "first" one. |
Small fixup found in the development of pygfx#1148
b90530b to
0262f62
Compare
Small fixup found in the development of #1148
This enables the blender to retain its state while it hasn't been completely cleared. This is useful if users are trying to render a scene manually, object by object, streaming in the objects from disk or other source slowly. One can recreate the same effect from the image_stitching.py example
0262f62 to
a60d200
Compare
Head branch was pushed to by a user without write access
|
screenshot has been uploaded from CI. |
thx thx. |
Probably not that important, but I'm a bit confused. The new validation example was not failing, but you uploaded a new image and it still works. Or is the new image only different in its metadata? |
|
Hmm. The last two commits were mine preceded by yours. Yours shows red. Likely because of aa changes. Then I just downloaded the screenshots and uploaded them. |
|
Ah, now I see. My small tweaks were fine, but merging with main likely affected aa. |
This enables the blender to retain its state while it hasn't been completely cleared.
This is useful if users are trying to render a scene manually, object by object, streaming in the objects from disk or other source slowly.
One can recreate the same effect from the image_stitching.py example
Part of this comes from our desire to display truely huge images.