-
-
Notifications
You must be signed in to change notification settings - Fork 1.6k
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
[BUG]: Tales of Rebirth software rendering regression #12100
Comments
can i have a memcard if possible |
heres a memcard for testing memcard_issue_12100.zip |
https://files.catbox.moe/xo3s3k.mp4 |
@macrosscape Can you post a gs dump of the issue? |
there's a gs dump. while I'm at it, here's a memory card. walk into the town with it set to software renderer. |
Just posting the debug progress so I/we don't forget what's happening: Reverting this gets around the issue: I compared the draws, first draw indicated in the log says 68, however when I compare the dumped draws it starts from 3 which is itex alpha draw 3 info: draw 68 info: Currently on master it breaks because inclusive_y_req is 0 in these cases TODO: Find out why inclusive_y_req is 0 in the conditions and if there's something we can do about it. |
Hi Lightning, I have also been looking at this bug, hope that is ok! Thanks for sharing the above, here is what I have found also: The glitch seems to be caused by bilinear filtering and goes away when bilinear filtering is disabled. Particularly, it seems to caused by some of the 4 UV coordinates used in the bilinear interpolation going outside the texture by 1 pixel and becoming pure black. This causes the alpha of the final color to be slightly less than 128, which causes it to fail the alpha test (the alpha test is set to GEQUAL with AREF=128). The blue streaks are from the background not being covered up. It is strange that this seems to only affect software rendering though. |
I did some tests with homebrew and it seems like the issue above affects all the renders (SW, VK, OGL, DX11) and both for triangles and sprites. Assuming this is a problem with bilinear filtering sampling outside the range of texture, is this already a well-known issue? I wonder if it is a limitation of PS2 that is being emulated accurately or a problem with the emulation. If anyone is able to test on a PS2, here are the ELFs. Also the GS dumps and a video showing the glitch. Hope this is useful. https://drive.google.com/file/d/1FjAbwvuCv462o1YSR7SB01AlJIjCIJWy/view?usp=sharing EDIT: These file can be ignored. I made a mistake in my homebrew code that caused the screen tearing, and was not related to problem in this game. |
I submitted as PR #12127 that may address the issue in this game. Though I haven't tested it very well so potentially it could break other things. Much appreciated if anyone is able to test it. |
Describe the Bug
Tales of Rebirth software rendering broke with PR #11324 in PCSX2 v1.7.5865 (June 4th 2024)
It's been broken for half a year and causes the backgrounds to tear when using software rendering.
DRAGONBLEAPIECE asked me to make this report.
Reproduction Steps
Walk around Sulz, the first town, or any town, with software rendering on.
Expected Behavior
No response
PCSX2 Revision
Latest
Operating System
Windows 10 (64bit)
If Linux - Specify Distro
No response
CPU
Ryzen 5 3600
GPU
rx570
GS Settings
No response
Emulation Settings
No response
GS Window Screenshots
No response
Logs & Dumps
No response
The text was updated successfully, but these errors were encountered: