Post by ars4326

Umihara no Naku Kawase Shun is a Japanese murder mystery game with some really innovative fishing segments. You Reading This is a very disappointed submission text reader who feels as though Samsara could have come up with a better intro.

This Run Is "Slower" Because Of More Accurate Emulation

To be more specific, this run has a longer input time because of the PSX BIOS being included. I'm putting this at the front because there has been more than one recent occasion where some blind TAS watcher votes No for the run being "slower" even though all of the gameplay segments are improved.

So maybe adding a long section where people will see it might help... but then again, I've overestimated the people on this forum before.

Game objectives

Emulator used: BizHawk 1.11.1
Aims for fastest in-game time, technically
I'm not doing water puns again
You remember what happened last time
Yeah, I thought so
I don't want the thread flooded with puns
WAIT
OH NO

Author's Encode

(removed, sorry!)

Comments

Umihara Kawase Shun, aside from a phrase you should always reply to with "gesundheit", is part of the Umihara Kawase series. Until recently, the Umihara Kawase series was Japan-only, but the most recent games are making their way stateside. Could this description get any more stentorian? Let me try.

Shun is the 2nd game in the series, obviously released on PSX, but it was packaged with the first game and ported to the DS, with a couple added features and some changes to the levels. Shun plays a bit slower than the original game, but the physics are mostly the same between the games aside from the line being a bit shorter.

This run is 516 frames faster than the published run in terms of in-game frames. Like the SNES Umihara Kawase run, I initially had this run finished, but decided to check one last time for alternate TASes and ended up finding a Nico TAS of the DS version. Even more coincidentally, the run I found was from TKDL, the same TASer whose run I found last time. Apart from the first, every Field in this run either matches or beats TKDL's times.

Stage Comparisons

Thanks to Spikestuff for making this section possible.

FIELD	Mukki	Samsara	Frames Improved
00	1002	843	159
11	601	557	44
23	458	438	20
14	523	507	16
15	803	796	7
35	1014	936	68
36	551	529	22
31	718	565	153
30	456	429	27

Field-by-Field Comments

Field 00

This Field is around a second slower than TKDL's TAS, and there's a very good reason why that is. If we hearken back to the run comments, I mentioned that the DS port has extra features and some changes to the levels. The extra features aren't really anything to write home about, because they're in Japanese and no one at my home would understand it, but Field 00 in the DS port actually has extra platforms and less enemies, which allows for a different and faster route through. This is the only Field in the run where that's relevant. The rest of the Fields in the run are unchanged in the DS port and thus completely comparable.

The bulk of the improvement is made up through better line control and the faster strategy at the end. The first swing up onto the platform, though it looks innocuous, was actually one of the tougher things in the run to get. Mukki wasn't able to get it in his run, but after working on it for an hour or so I finally managed to get Umihara to cooperate.

Field 11

I get a bigger swing than TKDL which lets me come in slightly faster, though I have to hook the ground to bring myself down in order to actually go into the door.

Field 23

One of the more interesting and tough to optimize strategies in the run. An odd little bit of physics abuse happens where jumping off of the slope actually lets me jump forward at full speed without the need to accelerate.

Field 14

Probably the stage I spent the second most time on overall. One of the problems with Umihara TASing is that there are so many places where you can potentially lose time, so if you're comparing to another TAS and you end up behind, then you pretty much have to go back through the entire stage several times over to see where you might've gone wrong. The end bit with the fast swing up onto the door platform was the trickiest part. It requires a pretty precise angle and really good line control.

Field 15

Thankfully this stage was already faster than TKDL's. Everything is mostly similar to Mukki's run except more optimized.

Field 35

The longest stage in the run! There's a lot of trickery with the line here. The last jump and swing in particular was a pain, but it paid off since I improved TKDL's time by a frame.

Field 36

A short stage that's a hell of a lot trickier than it looks, since it requires really precise control and a little bit of physics abuse.

Field 31

Spikestuff pointed out that a realtime run of this stage was about half a second faster than the published run, and that's what inspired me to give the game a look in the first place. I ended up taking that 30 frames of improvement and shaving another 2 seconds off of it by being able to slip past the right tadpole and put myself in a much better position to get under the left one at the earliest possible opportunity.

Field 30

This Field took about 8 hours of work total. Despite being one of the shorter stages in the run, it was also the hardest to properly optimize. Every swing, hook placement, and lengthening/shortening of the line is frame-precise, otherwise the entire strategy falls apart. I did find that it is possible to jump off of the slanted spikes if you land on them the right way, but it doesn't save any time.

Other comments

All in all, I always have fun TASing Umihara games. I'm still thinking about going back to the SNES version and seeing what I can do about improving Sightseeing and possibly trying to shave a few frames off of any%, but I feel like I have much bigger projects to worry about in the meantime.

Special Thanks

Biggest thanks to TKDL, the prime Umihara Kawase TASer out there. Without their TASes, both the SNES Sightseeing run and this run wouldn't be nearly as fast as they are
Extra thanks to Spikestuff for cluing me in on the Field 31 improvement and for finding the in-game framecounts of the published run so I could properly compare times
Shoutouts to arandomgameTASer, xy2_, ars4326, dekutony, and anyone else supporting me in the IRC for being chill and motivating
Extra shoutouts to moozooh, who can't fault me for deviating from the proper Sightseeing path this time. HA! IN YOUR FACE! pls enjoy the run :c

Screenshot Suggestion

Frame 4835, 6618, 7517

ars4326: Claimed for judging!

ars4326: Hello again, Samsara. Closer watching the two runs side-by-side (and compensating for the difference in loading times), it was evident that the gameplay optimizations made were very significant over Mukki's TAS. Good to see that every stage was improved upon (and that you also took the initiative to match, or exceed, TKDL's work, too). Overall, remarkable work and dedication on this one!

Accepting as an improvement (better emulation and optimization) to the published run!

Spikestuff: Publishing.