#1641 closed defect (fixed)
Assembly versions of HQ2x and HQ3x scalers
| Reported by: | SF/kaminari | Owned by: | sev- |
|---|---|---|---|
| Priority: | normal | Component: | Graphics: Scalers |
| Version: | Keywords: | ||
| Cc: | Game: |
Description
Since the recent update of these filters on 2004-05-20 (23:08), some distorsion around the cursor is now clearly noticeable. The graphics right under the cursor are correctly filtered, but just move the cursor slowly and notice how the graphics get slightly distorded in its wake (try it on the game selection screen for example).
It's been happening since scummvmwin32_20040521.exe and later. Works perfectly in previous builds.
Ticket imported from: #961365. Ticket imported from: bugs/1641.
Attachments (1)
Change History (9)
comment:1 by , 20 years ago
comment:2 by , 20 years ago
| Owner: | set to |
|---|
by , 20 years ago
| Attachment: | hq_asm.diff added |
|---|
comment:4 by , 20 years ago
It seems the assembly versions handles directly the special extremum points (0, 0), (w-1, 0)... to avoid reading out of graphics buffer. This isn't needed as those special cases are directly handled in the SDL backend by allocating extra bytes.
The attached patch simply changes the asm version to act like the C one, ie without checking the special points. Last thing, I didn't tested that much, hence why I prefer posting the patch than directly committing it :)
comment:5 by , 20 years ago
| Status: | new → closed |
|---|
comment:7 by , 20 years ago
| Resolution: | → fixed |
|---|
comment:8 by , 5 years ago
| Component: | --Unset-- → Graphics: Scalers |
|---|
I see this with the Windows snapshot, as well. I'll have to remember to try it on my Linux box at home later, but I would assume the behaviour is the same there.
In this case it is, of course, only rescaling the area of the screen that is affected by the cursor movement. It might just be my imagination, but it looks to me as if it isn't scaling the edges of the image. Could there be some off-by-one error somewhere?