(Link to video)
・Dual core off
・idol skip off
・DSP LLE recompiler
・no memory card
・Direct 3D11
・intel(R) HD Graphics family
Uses 4 gamecube controllers
・MD5 checksum 3b07a4bb22db926b177e207f9d7f0d87
Uses credit warp by arbitrary code execution.
cutscene underflow glich causes it.
explanation of cutscene underflow glich(quote from noki doki) "The game has 8 “slots” to store active cutscenes, numbered 0 to 7. It also keeps track of the slot number of the last cutscene that started and the slot number of the last cutscene that ended. When a cutscene starts, the last started cutscene slot number increases, cycles back to 0 if it would go past 7, and the new cutscene gets loaded into that slot. Once the cutscene ends, the last ended cutscene slot number increases in turn, becoming equal to the last started cutscene slot number. Since both slot numbers are equal, the game knows it ended as many cutscenes as it started, so there are no more cutscenes to process.
Since there’s usually no way to end more cutscenes than you start, and the slot numbers are being used in a cycle, any time the last started and ended slot numbers don’t match up, that must mean there are cutscenes left to process, even if the last ended cutscene seems to be ahead of the last started cutscene. But when a Shine Get cutscene starts during another cutscene, for some reason, it ends twice, leaving the game with -1 cutscenes left to play, which it sees as 8-1=7.
hat if you replay a cutscene in a different area than it was originally played? For most cutscenes, nothing interesting happens: either they play normally, or they end instantly, depending on whether the associated camera movement filename was overwritten in the transition or not. But some cutscenes retain a callback to execute before and after the cutscene (even if the camera movement couldn’t be found), and a pointer to a relevant object to be passed to the callback. Most notably, Shine spawns keep a pointer to the Shine being spawned.
At the beginning of a Shine spawn cutscene, its callback will lead to one of the Shine’s virtual functions, TMapObjBase::makeObjAppeared, being called. When replaying the cutscene with cross-area cutscene underflow, its Shine pointer will be stale and the virtual call can be redirected to arbitrary code. Let S be the Shine pointer; *S will be read as a pointer to the Shine’s virtual table, and *(*S + 0x100) will be set as the instruction pointer. "
setting right bell cutscene on slot4 while peach is kidnapped. the game has some delfino plaza status, each plazas objects have different memory address. so the value of right bell cutscene(8134A65C) while peach is kidnapped will be related turbo gate address while poluted piranha appearances on ricco harbor.
By looking turbo gate on screen, the value of (8134A65C) can be changed. so setting the value to be related to controller address.(this part was made by su)
at the last, get a shine while blue coin cutscene is playing. then input codes by 4 controlles for credit warp.(this part was made by su)

ThunderAxe31: Claiming for judging.
ThunderAxe31: Excellent job! Accepting for publication as a new branch.
fsvgm777: Processing.
Also changed the branch to "game end glitch" to be consistent with other TASes of this kind.

TASVideoAgent
They/Them
Moderator
Joined: 8/3/2004
Posts: 14884
Location: 127.0.0.1
Active player (434)
Joined: 2/5/2012
Posts: 1690
Location: Brasil
way too much insanity here lol
TAS i'm interested: megaman series: mmbn1 all chips, mmx3 any% psx glitched fighting games with speed goals in general
GamesFan2000
He/Him
Joined: 1/4/2019
Posts: 75
Location: Canada
Who knew that spawning a blue coin as you collected a shine would f___ with the game that much? Another game destroyed by computer science, another yes vote.
fsvgm777
She/Her
Senior Publisher, Player (221)
Joined: 5/28/2009
Posts: 1185
Location: Luxembourg
Could you also share your graphics settings? I assume they're the same as with the current any% publication, but I'd like to be sure.
Steam Community page - Cohost profile Oh, I'm just a concerned observer.
Challenger
He/Him
Skilled player (1639)
Joined: 2/23/2016
Posts: 1036
This is a milestone. Excellent job! Yes vote.
My homepage --Currently not much motived for TASing as before...-- But I'm still working.
Editor, Player (163)
Joined: 4/7/2015
Posts: 330
Location: Porto Alegre, RS, Brazil
What an amazing feat! I'm glad we could all hear the clown honking too.
Games are basically math with a visual representation of this math, that's why I make the scripts, to re-see games as math. My things: YouTube, GitHub, Pastebin, Twitter
Skilled player (1706)
Joined: 9/17/2009
Posts: 4952
Location: ̶C̶a̶n̶a̶d̶a̶ "Kanatah"
The progress was completely quiet after the 34 minute proof of concept, so I was legitimately surprised that a sub 10 minute route was possible. Congratulations both of you for finally figuring it all out! I know absolutely nothing about the PowerPC instruction set, but I would love to see what the payload was for the credits jump. How feasible is this for RTA? What happens after the credits? Does the game's story still work correctly?
Joined: 9/22/2012
Posts: 15
Location: Oregon
This is unhinged. Probably the easiest Yes I've ever voted.
fsvgm777
She/Her
Senior Publisher, Player (221)
Joined: 5/28/2009
Posts: 1185
Location: Luxembourg
We've been trouble getting this run to sync (it desyncs shortly after the FLUDD cutscene, where the Goo Piranha doesn't show up at all, but nothing looks off before that), even on Intel HD Graphics 3000 (which I recall was the one that got the published any% TAS to sync) from my testing, as well as Intel 4xxx from other users. zelpikukirby, could you please provide your graphics settings and which driver version you're running?
Steam Community page - Cohost profile Oh, I'm just a concerned observer.
Skilled player (1706)
Joined: 9/17/2009
Posts: 4952
Location: ̶C̶a̶n̶a̶d̶a̶ "Kanatah"
fsvgm777 wrote:
We've been trouble getting this run to sync (it desyncs shortly after the FLUDD cutscene, where the Goo Piranha doesn't show up at all, but nothing looks off before that), even on Intel HD Graphics 3000 (which I recall was the one that got the published any% TAS to sync) from my testing, as well as Intel 4xxx from other users. zelpikukirby, could you please provide your graphics settings and which driver version you're running?
In case they don't reply, they have a twitter account as well. https://twitter.com/zelpikukirby
Emulator Coder, Judge, Experienced player (608)
Joined: 2/26/2020
Posts: 698
Location: California
Got this to sync, it seems this is just a case of the SRAM.raw file having Mono set instead of the default Stereo, causing the desync. Also does not appear to have hardware dependent sync.
GJTASer2018
He/Him
Joined: 1/24/2018
Posts: 251
Location: Stafford, NY
CasualPokePlayer wrote:
Got this to sync, it seems this is just a case of the SRAM.raw file having Mono set instead of the default Stereo, causing the desync.
May I ask what's the point of even changing that setting in the first place? Is there some performance reason why someone would want to do this?
c-square wrote:
Yes, standard runs are needed and very appreciated here too
Dylon Stejakoski wrote:
Me and the boys starting over our games of choice for the infinityieth time in a row because of just-found optimizations
^ Why I don't have any submissions despite being on the forums for years now...
Emulator Coder, Judge, Experienced player (608)
Joined: 2/26/2020
Posts: 698
Location: California
For whatever reason it was changed may be due to some kind of personal preference, perhaps playing some game that changed it for whatever reason, maybe changing it long ago for whatever reason and forgetting it. Dolphin doesn't wipe the SRAM.raw when making a new movie, so this could have been set long ago.
fsvgm777
She/Her
Senior Publisher, Player (221)
Joined: 5/28/2009
Posts: 1185
Location: Luxembourg
CasualPokePlayer wrote:
Got this to sync, it seems this is just a case of the SRAM.raw file having Mono set instead of the default Stereo, causing the desync. Also does not appear to have hardware dependent sync.
Confirming sync by setting the audio to Mono in SRAM.raw (either via booting from IPL or hex editing the specific bit). The whole thing with the GPU turned out to be a red herring in this case, as I synced it on a laptop with an AMD CPU and an nVidia RTX 4060 GPU.
Steam Community page - Cohost profile Oh, I'm just a concerned observer.
Joined: 3/9/2024
Posts: 3
Location: France
I reset my Dolphin Lua settings, only changed the ones mentioned in the submission, then switched the audio to Mono in the IPL, but the airstrip plant still won't come out. Any suggestions on what else could be causing the desync?
Joined: 10/1/2013
Posts: 98
Location: My Basement
Technically impressive in every sense of the phrase. Even ignoring the ACE it's quite novel and entertaining. Yes vote. I do wish the description was a little more fleshed out, with better grammar and some more explanation of the "turbo gate", payload, and controller reading.
Skilled player (1706)
Joined: 9/17/2009
Posts: 4952
Location: ̶C̶a̶n̶a̶d̶a̶ "Kanatah"
NokiDoki wrote:
I reset my Dolphin Lua settings, only changed the ones mentioned in the submission, then switched the audio to Mono in the IPL, but the airstrip plant still won't come out. Any suggestions on what else could be causing the desync?
Hey, you're the one who initially discovered that this was possible right? Nice to see you here! Did you ever figure out how to sync it? I would love a more detailed explanation of the payload, and input, as well as any ideas on if this will ever be RTA-viable.
Joined: 3/9/2024
Posts: 3
Location: France
That's me! I'm working on an explanation video about this, so being able to play the DTM up to the interesting part would be pretty helpful, but I'm not sure what I can do to sync it... As for RTA-viability, the weird-looking sidestep just before going out of bounds around the 9:30 mark is manipulating something to be at a very precise position on screen, but unlike my LOTAD from 2 years ago, that thing should be consistent, so I don't think it's unrealistic for a setup to be found, and the rest is precise but likely doable
Skilled player (1706)
Joined: 9/17/2009
Posts: 4952
Location: ̶C̶a̶n̶a̶d̶a̶ "Kanatah"
NokiDoki wrote:
That's me! I'm working on an explanation video about this, so being able to play the DTM up to the interesting part would be pretty helpful, but I'm not sure what I can do to sync it... As for RTA-viability, the weird-looking sidestep just before going out of bounds around the 9:30 mark is manipulating something to be at a very precise position on screen, but unlike my LOTAD from 2 years ago, that thing should be consistent, so I don't think it's unrealistic for a setup to be found, and the rest is precise but likely doable
Do you have the right file? The md5 posted in the submission text is: 3b07a4bb22db926b177e207f9d7f0d87 What are your settings, along with hardware?
Joined: 3/9/2024
Posts: 3
Location: France
Welp... I'd been using my previous download of Dolphin Lua Core v3.5.1 with only the User folder renamed to User_, thinking it'd be equivalent to a fresh download, but typing it out made me think that might be the problem, and sure enough, an actual fresh download with the settings mentioned in the submission and audio set to Mono in SRAM does the trick. 😅 Thanks!
Post subject: Movie published
TASVideoAgent
They/Them
Moderator
Joined: 8/3/2004
Posts: 14884
Location: 127.0.0.1
This movie has been published. The posts before this message apply to the submission, and posts after this message apply to the published movie. ---- [5882] GC Super Mario Sunshine "game end glitch" by zelpikukirby & su in 09:50.40