Viewing Issue Advanced Details
ID Category [?] Severity [?] Reproducibility Date Submitted Last Update
05423 Gameplay Minor Always Jan 11, 2014, 14:43 Oct 27, 2023, 18:31
Tester rcoltrane View Status Public Platform MAME (Official Binary)
Assigned To Haze Resolution Fixed OS Windows Vista/7/8 (64-bit)
Status [?] Resolved Driver
Version 0.152 Fixed in Version 0.261 Build Normal
Fixed in Git Commit 8d1c7fc Github Pull Request #
Summary 05423: stdragon, stdragona: 1p start screen - texts are not moving like they should
Description the 1p start screen scrolling texts are standing still in the center of the screen instead of moving left/right while the game is waiting the player to press start.
Steps To Reproduce
Additional Information
Github Commit
Flags
Regression Version
Affected Sets / Systems stdragon, stdragona
Attached Files
 
Relationships
There are no relationship linked to this issue.
Notes
19
User avatar
No.10497
rcoltrane
Tester
Mar 21, 2014, 17:10
Until MAME 0.143, the 1p start scrolling texts were working fine. On version 0.144, Angelo Salese made some changes do the Megasys1.c driver by removing some ROM patches for some games, including Saint Dragon, and this may be the cause of the scrolling texts not working anymore. This is the comment taken from Whatsnew 0144.txt:

"Added preliminary MCU handshake simulation sequence for astyanax,
hachoo, jitsupro, plusalph, stdragon and removed ROM patches for those
games [Angelo Salese]"

I've tested every single version since 0.143 and this version is the last one where the Saint Dragon 1p start texts were scrolling, which is the correct behavior confirmed against the original PCB.
User avatar
No.11395
Kale
Developer
Jan 25, 2015, 04:03
Interesting, since protection is actually unchanged from old behaviour, unless it effectively do ROM self-modifying code on main CPU, which I doubt a bit.
User avatar
No.12233
rcoltrane
Tester
Dec 10, 2015, 13:22
Very Good Kale, thanks a lot! This is one of my favorite games :)

What was causing this issue afterall?
User avatar
No.12245
haynor666
Tester
Dec 14, 2015, 14:40
edited on: Dec 14, 2015, 14:54
I've tested latest commit and now text is quite weird. It partially moves just like long time but there is also second copy of this text that still stands still. Both texts are flashing. Seems to be displayed alternately every frame. Is this ok ?


I've compiled mame x64 myself with old tools if this matter.

EDIT. Ok, now I see what might be problem - Haze is emulating sprite trail effect maybe that's problem right now.
User avatar
No.12246
Haze
Senior Tester
Dec 14, 2015, 16:10
don't think this is related to any trail effect.

I have no idea if the flickery effect is correct tho, would have to ask a PCB owner..

Jaleco did ugly things...
User avatar
No.12248
rcoltrane
Tester
Dec 14, 2015, 16:32
Here's a video I did with MAME 0.142, when the 1P Start screen was working fine. See if the current fix looks like this:

User avatar
No.12250
Haze
Senior Tester
Dec 14, 2015, 17:25
edited on: Dec 14, 2015, 17:26
the video is recorded at 30fps, the flickering happens at 60fps (it's every other frame), you can't compare the two. (this would also be a problem with any youtube videos recorded before youtube supported 60fps, hence real PCB owner verification being needed)
User avatar
No.12256
Kale
Developer
Dec 15, 2015, 19:04
edited on: Dec 15, 2015, 19:05
My fix was breaking few games, so I've reverted it.

It's also more annoying than expected, I'm guessing that the real HW could do raster interrupts and that wasn't known until now.
I've made an example video, but it definitely needs HW references more than before ...

User avatar
No.12257
rcoltrane
Tester
Dec 15, 2015, 22:21
Thanks for the video Kale. In fact your raster effect looks better than the effect shown in my previous video.

Anyway, since some guys on the Shmups forums have this board, I've posted a new topic there begging for someone to help us by providing a small footage of the real PCB booting and having a coin inserted. This way I think we can settle this 1P start screen behavior once and for all.

Speaking of raster interrupts, it may be the cause of the 3rd level boss to remain hidden if you don't blow up the last big structure, allowing the player to go straight to level 4 without any fight. This one however would be more difficult to track, since I don't think the guy will want to play the game until that part using his real PCB... I mean, they are PCB collectors, and they don't want to keep connecting their boards and playing them all the time. And those Mega System 1 A boards are very fragile from what I've read.
User avatar
No.12261
Haze
Senior Tester
Dec 17, 2015, 13:29
No, raster interrupts have nothing to do with a 3rd boss remaining hidden, raster effects are a visual thing.

I believe we've already established that to be real PCB behavior, that boss is conditional (it's considered one of the hardest bosses in the game, so being conditional makes sense really)

There is a chance the byte smearing I added for 64th street attract MIGHT change some behavior in the game because it does make byte accesses, but again, we've had PCB owners confirm that the boss really is conditional.
User avatar
No.12270
rcoltrane
Tester
Dec 21, 2015, 01:35
edited on: Dec 21, 2015, 01:43
MJR, from the SHMUPS forum replied to my request and made a video of the Saint Dragon PCB booting and 1P start screen raster effects! Here's the video:



As you can see, the recent code modifications are in fact an improvement!! It is almost 100% correct, except that in MAME, the bottom text remains static, while it should move on the opposite direction of the upper text. But you guys definitely made an improvement when comparing your latest video to that 1st video I posted here. Please keep going with these changes, you guys are in the right direction :)

Another thing worth mentioning in this video: The game inicializes faster than in MAME during boot. In MAME, the game stays at the dipswitch numbers for a longer period of time than the real PCB. The game boots faster on the PCB and it goes to the title screen pretty quick.
User avatar
No.12271
Haze
Senior Tester
Dec 21, 2015, 11:59
startup time might just be how long it takes the monitor to warm up, so it's taking longer but you don't see it? although there are reports that tshingen is the opposite (starts far too quickly in MAME)

either way, nice reference video ,good to know what the effect should look like.
User avatar
No.12272
Kale
Developer
Dec 22, 2015, 13:47
I've tried, and the only possible way to set that raster effect is to fire irq 2 at line 70 and irq 1 at about line 95.
It's of course a bit odd: HW is capable to program irqs at completely arbitrary lines instead of the usual vblank-in/out? And where's this control?
User avatar
No.12384
rcoltrane
Tester
Feb 4, 2016, 19:10
Does it help on something? I guess this chip may be used by one or more Jaleco games, including this one. If not, please ignore:

http://green.ap.teacup.com/junker/150.html
User avatar
No.12685
system11
Senior Tester
May 21, 2016, 05:01
That's one of the ones all the inputs pass through, they have three. They very, very frequently break off, I'd started to trace one out myself but this saves me finishing it - thanks for linking that!
User avatar
No.21337
ICEknight
Tester
Apr 21, 2023, 22:47
edited on: Apr 21, 2023, 23:52
Just found some less shaky footage showing the moving "Push Start" words on real hardware:


Additionally, the Arcade Archives release seems to get it right (unlike the Evercade version):



Regarding differences in startup time, it can be skipped by pressing any Fire button so it might be irrelevant if it's shorter in some videos.
User avatar
No.21808
Haze
Senior Tester
Oct 22, 2023, 15:56
i've added support for this, I don't see any breakage in other games.
User avatar
No.21813
hap
Developer
Oct 27, 2023, 18:00
Same commit as the tshingen bugfix?
User avatar
No.21814
Haze
Senior Tester
Oct 27, 2023, 18:07
was part of the one before https://github.com/mamedev/mame/pull/11629