My project makes heavy use of depth-of-field post processing, currently from the default pipeline. I notice, however, that the level of blur is tied to screen resolution:
It’s important to me to have the same experience regardless of the user’s screen resolution. Any ideas on how I can achieve this? Thanks in advance.
It’s alike the problem in this thread:
Hi everyone! I hope you are all doing well.
I have one question concerning how is managed the focal when going HD.
As you can see in these images the blur due to the focal is really changing depending on the hardware scaling level.
HardwareScalingLevel = 0.5:
[hd.PNG]
HardwareScalingLevel = 1:
[nothd]
I guess the quality of the rendering makes the focal more precise and thus there is less blur.
First, is it a normal behavior? And how could I avoid that change when going to HardwareScali…
The solution was to make the blur kernel depend on the texture size.
1 Like
sebavan
February 28, 2024, 9:19pm
3
Let me add that your screenshots are pretty ?
@Evgeni_Popov I wonder if there are anything we could do for window independent setup ? maybe a new mode to not break back compat ?
1 Like
Ha thx. Here’s the project for a little more context fyi: CameraSim 3D
Please let me know if you do anything to address this under the hood, otherwise I’ll try the solution from @Evgeni_Popov ’s thread.
sebavan
February 28, 2024, 10:16pm
5
cc @PirateJC and @PatrickRyan as they love photos
@sebavan I’ve updated our internal issue about improving DoF
1 Like
@Evgeni_Popov How can I follow any progress made on DoF?
This is an internal issue, so it’s not visible to the outside world.
We’ll be sure to make an announcement when the PR is ready, but don’t hold your breath, as it’s not on the priority list, and we also have to prepare the new Babylon version to come (by the end of this month).
Yep, understood. I don’t know how I could help, but I’d be eager to since DoF is central to my app.
If you can, you could try to implement another algorithm that would fix the problems identified in:
It would require a bit of research to find a suitable algorithm, and then to implement it, though.
1 Like