[Bf-viewport] Failled accurate soft shadow attempt

[Felipe Botero]

Hello,

I am no expert on the topic, but I know that in UE4 they use Distance
Fields for blurred shadows, this has the disadvantage that meshes' distance
fields need to be built, but it seems to solve the problem, and also allows
them to use the distance fields for AO.

Maybe having DF shadows disabled while in edit mode, but enabling them in
object mode maybe would work, and building the distance field in the
background if the mesh changed.

Also for skeletal meshes they also have smoothed shadows, but they don't
use mesh data but instead a PhysicsAsset data (which is basically the
capsules or other primitives used when simulating a ragdoll), maybe in
blender for those meshes the skeleton envelope could be used.

Thank you for your time, and keep up your fantastic work!

- Felipe Botero


On Mon, Sep 4, 2017 at 8:39 AM, Clément FOUCAULT <foucault.clem at gmail.com>
wrote:

> Hi guys,
>
> As you may have noticed I had some trouble one week ago and did not
> commited anything. This is because I was trying to get the most awesome
> shadows out of shadow maps.
>
> Foolish of me to think this had not been attempted before. I reckon having
> lost a bit of time before doing research.
>
> But long story short: There is no good, easy, fast solution for realistic
> soft shadows. By this, I mean contact hardening soft shadows. Fixed size
> penumbra shadows will be implemented (they already are).
>
> A more complete explanation of what I tried can be found here:
> https://wiki.blender.org/index.php/User:Hypersomniac/Dev_reports/Soft_
> shadows_failure#Soft_Shadow_Map_Faillure
>
> If someone is feeling more confident on this and want to take a shot at
> it, please do and get in touch.
>
> Clément
>
> _______________________________________________
> Bf-viewport mailing list
> Bf-viewport at blender.org
> https://lists.blender.org/mailman/listinfo/bf-viewport
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.blender.org/pipermail/bf-viewport/attachments/20170904/31c9afc2/attachment-0001.html>