Can you retry with the latest git version in opengl with accurate_colclip=1 in the ini file and "disable crc" in the hw hacks please?
Sure i'll do it when I get home
which ini file @refractionpcsx2
i tried that line in both gsdx ini files and disable crc in hw hacks under opengl same issue occurs
yeh the gsdx.ini
as long as you used openGL it would take effect
yeah i tried that with the one made 2 hours and 49? minutes ago nothing changed the issue still occurs even when using openGL @refractionpcsx2
ok thanks for testing, can you provide a GS Dump?
how do i do that? @refractionpcsx2
When you're on the scene with the problem, hold down shift+F8 for a couple of seconds, it will then put a BMP file showing the error and a .gs file in your "snaps" folder. 7zip them and upload them somewhere :)
ah here you go https://www.dropbox.com/s/axwz6ll5ntcxma4/snaps.7z?dl=0
Thanks :) You wanna add it to your list @gregory38 ? :)
Well it could be EE/VU bug too. So I won't look into it soon.
@kamijoutouma
Did zerogs have this problem ?
This question was asked before in one of the threads earlier and the answer was yes.
Also there was some extra info that about 25% of the screen was all black all the time in zerogs (specifically the bottom).
@ssakash @refractionpcsx2 @gregory38

Although i think the extra bit happens in all games.
@ssakash in case you're implying a core issue, just because ZeroGS also has the issue doesn't mean it's a core issue. It could still very much be a GS issue.
This isn't valid so I'm closing it. Either way it's technically playable according to Dabore in the forums (from Sept. 2015):
the last time i tried... it worked with a daily git version from the buildbot. you gotta use the opengl renderer and enable the HW depth option. it had lil glitches in the menus but the races worked. Smile
@MrCK1
Invalid label is for cases where the issue was never present in any revisions. In this case, this issue seems to be present in an older version of PCSX2, so it's not invalid technically. For issues which were present before in an older version and not present anymore, don't use an invalid label.
Also I'd prefer if tests are done on latest development builds before closing the issues, who knows? A regression might've occured after the time Dabore did his test.
I have discovered the cause of the bug. It's in fact a gsdx bug (and likely ZeroGS too).
This game clears depth by setting FBP = ZBP and the FRAME psm = ZBUF psm and drawing sprites 32 pixels wide and 32 pixels apart from each other. This is confirmed to work on real hw.
It works because if you divide the page in half the left side of the page and the right side of the page will address different blocks due to z formats and color formats swizzling.
However, since the game uses z24 for the frame psm there was a detail I was missing.
In the event that the a z psm is specified for FRAME the GS will internally use a color format for depth. This was also confirmed to be the case via hw test.
Thanks to efforts by water111 and refraction, I was able to fix this bug in Dobiestation.
See https://github.com/PSI-Rockin/DobieStation/pull/273
Most helpful comment
I have discovered the cause of the bug. It's in fact a gsdx bug (and likely ZeroGS too).
This game clears depth by setting FBP = ZBP and the FRAME psm = ZBUF psm and drawing sprites 32 pixels wide and 32 pixels apart from each other. This is confirmed to work on real hw.
It works because if you divide the page in half the left side of the page and the right side of the page will address different blocks due to z formats and color formats swizzling.
However, since the game uses z24 for the frame psm there was a detail I was missing.
In the event that the a z psm is specified for FRAME the GS will internally use a color format for depth. This was also confirmed to be the case via hw test.
Thanks to efforts by water111 and refraction, I was able to fix this bug in Dobiestation.
See https://github.com/PSI-Rockin/DobieStation/pull/273