Project

General

Profile

Actions

Emulator Issues #11495

closed

Metroid Prime - Raindrop effect on visor is not rendered correctly

Added by Rupeeclock over 5 years ago. Updated over 3 years ago.

Status:
Invalid
Priority:
Low
Assignee:
-
% Done:

0%

Operating system:
N/A
Issue type:
Bug
Milestone:
Regression:
No
Relates to usability:
No
Relates to performance:
No
Easy:
No
Relates to maintainability:
No
Regression start:
Fixed in:

Description

Game Name?

Metroid Prime (Gamecube)
Including a European version, and a US version with widescreen hack applied.

Game ID? (right click the game in the game list, properties, info tab)

GM8P01 (EU Region)
GM8E01 (US Region)

MD5 Hash? (right click the game in the game list, properties, info tab, MD5 Hash: Compute)

GM8P01 - b1379c44e0ebc521e18215de3e5dbeea (compressed as .gcz)
GM8E01 - 0d2ba934eec6a59ee1e155e0d039f194 (widescreen hack + trimmed ISO)

What's the problem? Describe what went wrong.

The raindrop effect on Samus' visor is not rendering correctly, where some droplets will display a portion of a black bar on the left side.
These droplets work by displaying a flipped-portion of what's drawn on screen relative to its position, this is usually noticeable by looking at droplets close to any geometry that contrasts the sky.

Issue will occur in both Direct3D 11 and OpenGL graphical backends.
Issue would occur on multiple different devices using Nvidia GTX or Intel HD integrated graphics.
Issue would occur regardless of multiple different graphic settings, including scaled EFB, EFB copies, XFB copies, internal resolution, ubershaders, texture cache accuracy, per-pixel lighting, fog, etc.

What steps will reproduce the problem?

At beginning of game in Tallon Overworld's landing site, the landing area has rainfall.
Walking forward or looking skyward will generate raindrop effects on the visor.
Raindrops will not generate when standing still, standing in water, under cover, or when looking at the ground.

Is the issue present in the latest development version? For future reference, please also write down the version number of the latest development version.

Dolphin 5.0-9213

Is the issue present in the latest stable version?

Yes, also occurs in stable 5.0 release from mid-2016.

If your issue is a graphical issue, please attach screenshots and record a three frame fifolog of the issue if possible. Screenshots showing what it is supposed to look like from either console or older builds of Dolphin will help too. For more information on how to use the fifoplayer, please check here: https://wiki.dolphin-emu.org/index.php?title=FifoPlayer

See attached screenshots from multiple devices.
Two FIFO recordings are attached in a single 7zip archive, one with Samus staring at the sky (raindrops), one with Samus submerged in water (no raindrops).
When recording the FIFO logs for the raindrops issue, the following warnings would repeat in alternating order until ignored:

FifoRecorder: Expected command to be 211 bytes long, we were given 323 bytes
FifoRecorder: Expected command to be 107 bytes long, we were given 163 bytes

Not ignoring these warnings would result in an excessively large log.
No such recording issue occurred when raindrops were not being rendered.

What are your PC specifications? (CPU, GPU, Operating System, more)

Dell XPS 15 9560:

  • Intel Core i7-7700HQ @ 2.80GHz 2.80GHz
  • NVIDIA GeForce GTX 1050
  • Intel(R) HD Graphics 630
  • Windows 10 Pro version 1803
  • 16.00 GB RAM

GPD Win2:

  • Intel Core m3-7y30 @ 1.00GHz 1.61GHz
  • Intel(R) HD Graphics 615
  • Windows 10 Home version 1803
  • 8.00 GB RAM

Desktop:

  • Intel Core i5-4690k @ 3.50 GHz 3.50 GHz
  • NVIDIA GTX 770
  • Intel(R) HD Graphics 4600
  • 8.00 GB RAM

Is there anything else that can help developers narrow down the issue? (e.g. logs, screenshots,
configuration files, savefiles, savestates)

See attached savestate for Dolphin version 5.0-9213, in five-part 7zip archive.


Files

Dell XPS 15.png (3.34 MB) Dell XPS 15.png Screenshot taken from Dell XPS 15 Rupeeclock, 12/22/2018 11:05 PM
GPD Win2 - Direct3D 11 5.0 stable.png (837 KB) GPD Win2 - Direct3D 11 5.0 stable.png Screenshot taken from GPD Win 2 (Direct3D 11 on 5.0 stable) Rupeeclock, 12/22/2018 11:05 PM
GPD Win2 - Open GL 5.0-9213.png (738 KB) GPD Win2 - Open GL 5.0-9213.png Screenshot taken from GPD Win 2 (OpenGL on 5.0-9213) Rupeeclock, 12/22/2018 11:05 PM
Desktop GTX 770 Direct3D 11 5.0-9213.png (762 KB) Desktop GTX 770 Direct3D 11 5.0-9213.png Screenshot taken from GTX 770 with Direct3D 11 on 5.0-9213) Rupeeclock, 12/22/2018 11:07 PM
GM8P01.7z.001 (5 MB) GM8P01.7z.001 5.0-9213 savestate 7zip archive Rupeeclock, 12/22/2018 11:21 PM
GM8P01.7z.002 (5 MB) GM8P01.7z.002 5.0-9213 savestate 7zip archive Rupeeclock, 12/22/2018 11:21 PM
GM8P01.7z.004 (5 MB) GM8P01.7z.004 5.0-9213 savestate 7zip archive Rupeeclock, 12/22/2018 11:21 PM
GM8P01.7z.003 (5 MB) GM8P01.7z.003 5.0-9213 savestate 7zip archive Rupeeclock, 12/22/2018 11:21 PM
GM8P01.7z.005 (3.29 MB) GM8P01.7z.005 5.0-9213 savestate 7zip archive Rupeeclock, 12/22/2018 11:21 PM
GM8P01 fifo logs.7z (1.53 MB) GM8P01 fifo logs.7z 5.0-9213 FIFO logs Rupeeclock, 12/22/2018 11:22 PM
GM8P01-16.png (731 KB) GM8P01-16.png 4.0-1146 Rupeeclock, 12/27/2018 12:40 AM
GM8P01-18.png (740 KB) GM8P01-18.png 4.0-1192 Rupeeclock, 12/27/2018 12:41 AM
Raindrops.png (2.93 MB) Raindrops.png JMC4789, 09/12/2020 09:10 AM
Actions #1

Updated by MayImilae over 5 years ago

  • Status changed from New to Accepted
  • Priority changed from Normal to Low

I've seen this before but couldn't bother making a report, so I'm just going to mark this as accepted right away!

That said, there is a high likelihood that this is EFB scaling related. That should be the first thing to check.

Actions #2

Updated by Rupeeclock over 5 years ago

Thanks, I've also just logged the following issue which is a device specific crash for Intel HD 615 graphics running Direct3D 11 in the same conditions.
https://bugs.dolphin-emu.org/issues/11496

I suspect that this crash is possibly related to this graphical issue, as the errors refer to DXTexture or Map failures in \core\videobackends\d3d\dxtexture.cpp

Most other graphics adapters or backends probably handle the error silently.

Actions #3

Updated by JMC4789 over 5 years ago

If I remember correctly, didn't Super Mario Sunshine have the same issue with its bubbles before Tev Fixes New?

Actions #4

Updated by Rupeeclock over 5 years ago

Where specifically do the bubbles in questions appear? I'd be happy to try and reproduce the graphical issue.
Is it bubbles in paint, air bubbles when Mario is under water, etc?

Actions #5

Updated by JMC4789 over 5 years ago

In the paint, but those bubbles work nowadays.

Updated by Rupeeclock over 5 years ago

I checked it out further and saw that the tev_fixes_new branch was merged in with 4.0-1192.

Tested Super Mario Sunshine on 5.0-9213 and the bubbles were still rendering correctly.
Then checked it in 4.0-1146, a build available prior to tev fixes, and the bubbles did not render correctly as expected.
Then checked Metroid Prime's raindrops on visor, and there was no rendering issue.
I then checked it again with 4.0-1192, but this also did not have any rendering issue.
These are also tested using Direct3D graphical backend.

It doesn't look like this problem was caused by tev_fixes, but also wasn't an issue prior to it.

Actions #7

Updated by JMC4789 over 3 years ago

If someone wants to bisect when this started, that'd be great. We know it worked at some point and it's supposedly broken now.

Actions #8

Updated by JMC4789 over 3 years ago

I found out what caused the regression - no longer needing EFB2RAM. Guess what fixes this issue? EFB2RAM.

Should probably put a note on the wiki I guess.

Actions

Also available in: Atom PDF