Viewing Issue Advanced Details
ID Category [?] Severity [?] Reproducibility Date Submitted Last Update
01954 Gameplay Minor Always Jun 29, 2008, 14:56 Aug 11, 2008, 04:23
Tester ejm973 View Status Public Platform MAME (Official Binary)
Assigned To Resolution Fixed OS
Status [?] Resolved Driver
Version 0.125u8 Fixed in Version 0.126u3 Build Normal
Fixed in Git Commit Github Pull Request #
Summary 01954: mspacman: Tunnel Slowdowns Perpetuated When They Shouldn't Be
Description Beginning with I believe the Apple board (level 5) there should be no more slowdowns for the ghosts in the tunnels. However, there is a slight slowdown persisting - that drops the ghosts back 1 ghost width with each pass through the tunnels. I have an original Ms. Pac arcade game and confirm that the tunnel slowdowns are not supposed to be persisting.

This has a couple rather significant effects: It interferes with a player's ability to group the ghosts, and it makes it easier for a player to escape the ghosts than would otherwise be possible under arcade gameplay.

Please see the following YouTube video showing how the gameplay should behave versus how it is behaving. This vid uses different versions than 125u8, however I have just this morning tested and confirmed that the bug persists with this latest version of MAME (125u8).

Note: MAME090b does NOT have this bug/anomaly, but later versions of MAME do.

Here's the video:
Steps To Reproduce See link to video above, which shows the bug versus normal gameplay. To reproduce this example, start up mspacman, activate the rack advance via dip switches to level 10, and experiment with the tunnels as shown in the video.
Additional Information
Github Commit
Flags Verified with Original
Regression Version 0.90u4
Affected Sets / Systems mspacman
Attached Files
 
Relationships
related to 04251Confirmed  mspacman: Ghost tunnel behavior different than arcade 
Notes
24
User avatar
No.01442
Haze
Senior Tester
Jun 29, 2008, 19:23
old versions of MAME used the bootleg patches for the original MSpacman, so effectively ran a bootleg version.

new versions emulate the protection on the original instead, and should be more accurate. dwidel was rather pedantic about this.

are you sure your board isn't a bootleg?
User avatar
No.01444
Tafoid
Administrator
Jun 29, 2008, 21:12
There is some discussion on the old forum here:
http://www.mameworld.info/ubbthreads/showthreaded.php?Cat=&Board=mamechat&Number=116190
User avatar
No.01447
ejm973
Tester
Jun 30, 2008, 02:34
Haze, that's a fair & good question, but yes I am certain the board and roms are original.

Also, beyond my current machine, I've had the opportunity to test three others in the past year (including two at Funspot arcade in May) and all behave the same.
User avatar
No.01448
Haze
Senior Tester
Jun 30, 2008, 05:59
you should probably try contacting dave widel then with the specifics
http://www.widel.com/

he basically documented the entire thing inside-out .. so is the only person who could tell you the reason for the differences in behavior you're seeing.
User avatar
No.01453
Fujix
Administrator
Jun 30, 2008, 15:07
I set it ack for more discussion.
If someone found something new, info and/or detail, or made progress, please post here.
User avatar
No.01522
PacWhiz
Tester
Jul 7, 2008, 16:16
I can confirm the 'tunnel-anomaly', as the video referred to above, is mine.
WolfMame 126 : anomaly is present.
WolfMame 106 : anomaly is present.
WolfMame 101 : anomaly is present.
WolfMame 100 : anomaly is present.
WolfMame 090 : no anomaly.
Mame 090 : no anomaly.
WolfMame 078 : no anomaly.

The same romset (mspacman) was used in all of the above cases.
I haven't tested versions in between 090 and 100, but the bug seems to have originated in one of the versions from 091-099.
Hope this helps!
User avatar
No.01523
Tafoid
Administrator
Jul 7, 2008, 16:27
There are binaries for all major releases at: http://mamedev.org/oldrel.html
Since you are able to test this quite readily with your knowledge - I would suggest downloading/trying it on those ready built binaries.
Start with 0.95 (the middle release) and check higher or lower based on results.
If you can find which major release the regression happens - this will make it that much easier to find a cause and possible fix. Thanks for the help!
User avatar
No.01524
PacWhiz
Tester
Jul 7, 2008, 16:29
Good plan, i'm on it ;)
User avatar
No.01527
PacWhiz
Tester
Jul 7, 2008, 17:04
edited on: Jul 7, 2008, 17:04
Ok, 0.91 is the version where The Tunnel Anomaly originated!
Plz fix! :) ... Good Luck & Thx :D
User avatar
No.01528
Haze
Senior Tester
Jul 7, 2008, 20:23
hmm we know when it changed, we know why it changed, just the general thoughts are that it's correct NOW and was wrong before.. I guess you're also saying it's wrong now then?
User avatar
No.01529
PacWhiz
Tester
Jul 7, 2008, 22:13
edited on: Jul 7, 2008, 22:16
Affirmative, it's wrong now.
Taking the arcade version as the best example, you will not find that 'tunnel anomaly' (i.e. the ghosts lagging in the tunnels after the pretzel board right up until the end) anywhere in any arcade.
Or at least, i'd never seen it until Mame.
So... you know when & why it changed, that's good!
Now change it back plz (^.^)

By the way... are you saying that it was ~consciously~ changed at the time?
User avatar
No.01531
Haze
Senior Tester
Jul 7, 2008, 22:25
well, previous versions of MAME didn't emulate the protection (they simply patched the game to run as a bootleg version), newer versions of MAME *do* emulate the protection, and this is the result, so it's not just a case of being able to revert the code, we need to figure out *why* with 'correct' emulation of the protection this happens.
User avatar
No.01532
PacWhiz
Tester
Jul 7, 2008, 22:56
Sry i'm not clear on what 'protection' implies; do you mean the part of the code which (tries to) prevent(s) copying and/or selling of the original rom? How can that affect the ghosts in the tunnels...?!
Me just simple player, me not understand (Layman's disease :P)
User avatar
No.01535
Haze
Senior Tester
Jul 7, 2008, 23:02
yes, the 'protection' device on an original mspacman board was designed to prevent bootlegs from being made. it has various effects on the gameplay, possibly including the one being pointed out here.

basically old mame versions ran mspacman as a bootleg, new versions run mspacman as mspacman.

based on the posts here it appears we're doing something wrong however.
User avatar
No.01536
PacWhiz
Tester
Jul 8, 2008, 00:14
edited on: Jul 8, 2008, 00:14
I see! Well, in the scope of things, it is rather minor, and only affects ghost-grouping and escape-tactics; altho this can be bothersome to Pac-freaks like myself (I only play WolfMame 090 now), who look for exact emulation of the arcade version of MsPac.
But you have done so much right! And for that, THX
The emulation is ~near~ perfect. ;)
User avatar
No.01582
ejm973
Tester
Jul 12, 2008, 14:47
Hello, Haze - I saw your comment on the MARPIRC forum that dwidel replied to you that his board and MAME are identical w/r/t the tunnels. One question and one comment:
Q: Can you think of any local (i.e. our computers) settings or variables that might be driving this one strange difference when we run it in MAME?
C: I will make a video of the effect on my Ms. Pac arcade game and then include in that video a close-up shot of my PCB & ROMs to confirm conclusively for the team that it is original hardware. I'm not as good with video editing as PacWhiz, so I may ask his help to add to my arcade vid a comparative vid of Ms. Pac MAME emulation in v125.

Thanks to everyone involved for your continued help with this!
User avatar
No.01826
Tafoid
Administrator
Jul 27, 2008, 17:24
I'm of the strong belief that this is a bug.

I tested noticed no stuttering in the tunnels on the 20th Anniversary Ms. Pac-Man/Galaga - which very well should be a 1:1 copy of the original (not like the namco museum/namco classics collection). Also, given the fact that no one has been able to show a PCB or claim their PCB shows this tunnel stutter that is currently in showing in MAME (all reported playings on real hardware do not show this tunnel delay) - I'm inclined to confirm this. As more evidence surfaces, the status can be changed.
User avatar
No.01829
Haze
Senior Tester
Jul 27, 2008, 20:39
Yeah.. I think it's a bug too, but more tests will be needed to fix it properly, according to dwidel the tunnel behavior depends on unmapped reads.. and obviously in some circumstances they must give different results to those obtained when testing..
User avatar
No.01857
Haze
Senior Tester
Jul 31, 2008, 10:32
please check the behavior in the latest build... see if that's fixed, or if anything else odd is happening. I made a slight change based on a suggestion by dwidel.
User avatar
No.01886
PacWhiz
Tester
Aug 2, 2008, 12:19
v126.3?
Anywhere i can dl it without subscribing to another bloody site?
User avatar
No.01887
Tafoid
Administrator
Aug 2, 2008, 12:45
http://www.emu-france.com/?page=fichiers&idFile=629
User avatar
No.01913
PacWhiz
Tester
Aug 3, 2008, 17:00
Eureka! You fixed it!! :D
Tested romset mspacman in Mame 0.126u3 (thx for that link Tafoid).
Great job guys, looks perfect!
No more "tunnel-anomaly", and as far as i can tell, the whole game now behaves EXACTLY like the original arcade game - once again, MANY THANKS! ;)
User avatar
No.01914
Tafoid
Administrator
Aug 3, 2008, 17:26
I'll mark this fixed then. Thanks for the follow up.
User avatar
No.02037
ejm973
Tester
Aug 11, 2008, 04:23
You guys are the BEST!!! Thanks so much for looking into this and fixing the issue. Great stuff.

A BIG thanks to all involved in this fix!!!

Eric