- --
Viewing Issue Advanced Details
ID | Category [?] | Severity [?] | Reproducibility | Date Submitted | Last Update |
---|---|---|---|---|---|
08884 | Graphics | Major | Always | Jul 12, 2024, 17:44 | Jul 18, 2024, 13:50 |
Tester | rcoltrane | View Status | Public | Platform | MAME (Official Binary) |
Assigned To | Resolution | Open | OS | Windows 10/11 (64-bit) | |
Status [?] | Acknowledged | Driver | |||
Version | 0.53 | Fixed in Version | Build | 64-bit | |
Fixed in Git Commit | Github Pull Request # | ||||
Summary | 08884: all pitfight sets: Pit-Fighter - major graphics inconsistencies when compared to the real hardware | ||||
Description |
After some tests with my new video capture card, I saw that Pit-Fighter graphics that are processed by the game's internal scaling system (all background and sprite graphics that appear in between the two side bars, excluding the texts/mugshots layer) are different when compared to the real PCB. To demonstrate it, I captured the very same frame from one of the game's intro sequence that are the same in both MAME and PCB and put these two images into an animated GIF file here (see attached gif file at the end of this report). Both images have the very same size, so the side bars/top mugshots and texts are extremely aligned. This test allows us to see that all graphics that are processed by the scaling system (warehouse structure and wooden box behind the crowd, background crowd, main characters and floor) are all changed in one way or another. Graphics on the PCB are slightly better and more defined than in MAME. Positions of the crowd sprites are slightly different, their graphics are also different, the wooden box at the back of the crowd has more details on the pcb than on MAME, the main characters shadows, positions and graphics are slightly different too. The “vertical gaps” already reported here in the past, that are more prominent in MAME than in the real hardware could also be related to the same scaling system emulation. With this and other tests I did, I think it may have something to do with how graphics that pass through the scaling system are being handled. Or maybe the emulation of the grouth system itself is not perfect, and therefore it delivers different objects/positions on screen. |
||||
Steps To Reproduce | Video recorded straight from the original 3 player rev. 3 pcb. | ||||
Additional Information | Recorded intro loops up to the 6th loop without inserting a coin and took a screenshot from the 6th loop fight demonstration (kato vs heavy metal). If you do not insert a coin for the 1st time, all intro sequences/fight demonstrations are the same in both pcb and MAME. | ||||
Github Commit | |||||
Flags | Verified with Original | ||||
Regression Version | |||||
Affected Sets / Systems | all pitfight sets | ||||
Attached Files
|
![]() | ||||
![]() | |||||
Relationships
There are no relationship linked to this issue. |
Notes
1
![]() No.22232
rcoltrane Tester
Jul 18, 2024, 13:50
edited on: Jul 18, 2024, 13:52 |
I attached here today a patent document from Atari Games which describes with great amount of details how the whole raster scaling system hardware works. This may be helpful to improve the quality of the scaling system of the atari g1 hardware emulation in MAME. The pdf file name is "atari-raster-scaling-system-pat5363119.pdf". This patent began in 1981 and was updated until 1994. Two Atari developers (Doug Snyder and Sam Lee) who were involved with the development of Pit-Fighter (and are even credited in the game) are mentioned as responsible for the patent. |
---|