Topic by TakaM

Posted: 5/18/2009 5:22 PM
(Edited: 5/19/2009 5:23 PM)

Post subject: Sonic Rush

Quote

TakaM

Joined: 3/25/2008
Posts: 64

I figured I might as well start the topic for this one. Zone 1 Act 1 first attempt: DSM | Youtube (manually corrected framerate) Zone 1 Act 1 second attempt: DSM | Youtube (manually corrected framerate) | Youtube (1 screen version) Zone 1 Act 1 third attempt: DSM | Youtube (manually corrected framerate) I just checked my time trial on my Sonic Rush cart and it's faster (42.13), I think I used one of my earlier routes for this TAS, I really wish I realized this before I went to the trouble of fixing the framerate lol... Oh well, hopefully this means getting it done within the 30's is possible. Tricks: When grabbing those things that spin you around in the hollow tree trunks, normally you keep spinning until the cable winds up and you hit the bumper pad, but you can just jump off the cable as soon as you're inside the trunk. You have to be careful not to pass through the bumper pad while you're still in the foreground though. I also should have came to a stop before hitting the giant ring at the end, the slower you're going, the less it spins.

.II

Posted: 5/18/2009 5:29 PM

Quote

Comicalflop

Editor, Experienced player (730)

Joined: 6/13/2006
Posts: 3300
Location: Massachussetts, USA

According to this, the WR is 37.93". So I think a route change is in order to get a lower time. Rather than slowing down before the ring to make it spin less, you may want to make the run go for fastest in-game time, which some other sonic TASes do. That way you can compete for lower time, the same way speedrunners do, and not have to worry about the circle spinning and can focus on speed instead. If you want, I can take that .dsm and make a proper AVI with audio.

☣ Homepage ☣ Retired

Posted: 5/18/2009 5:34 PM

Quote

TakaM

Joined: 3/25/2008
Posts: 64

Ahh thanks for that video, that is roughly the same route I used on my cart (although it's more awesome) Don't bother converting that DSM, I'll make a new run tomorrow.

.II

Posted: 5/18/2009 5:40 PM

Quote

Atma

Former player

Joined: 9/1/2005
Posts: 803

I'd suggest holding off on doing this for now as I'm fairly sure the current rerecording version of desmume still suffers from a hang in the second set of stages unless you're using the bios files/swis. If you can somehow manage to work around these hangs, that'd be interesting, but given that they don't happen at a specific place or time, you're going to have fun if they're directly on your path.

Posted: 5/18/2009 5:58 PM

Quote

Comicalflop

Editor, Experienced player (730)

Joined: 6/13/2006
Posts: 3300
Location: Massachussetts, USA

I don't know why the file size is 74 MB.... weird. http://www.mediafire.com/download.php?nenqnynyolz Is the game supposed to be this laggy? Desmume counted 1,036 lag frames; that's quite a lot for a short level.

☣ Homepage ☣ Retired

Posted: 5/18/2009 6:09 PM

Quote

Mitjitsu

He/Him

Banned User, Experienced player (532)

Joined: 4/24/2006
Posts: 2997

Can't wait till this run appears on the workbench. The WIP despite being unoptimized was more enjoyable than any of the Sonic Advance games.

Posted: 5/18/2009 6:12 PM

Quote

TakaM

Joined: 3/25/2008
Posts: 64

Atma, thanks for the info. I'm not even sure if I'll attempt the entire game, at the very least I plan on getting zone 1 done. And naw Comicalflop, I think there's only one spot I get noticeable lag in act 1 on my cart, but it's out of the way and not an issue.

.II

Posted: 5/19/2009 8:15 AM
(Edited: 5/19/2009 8:47 AM)

Quote

TakaM

Joined: 3/25/2008
Posts: 64

Took another swing at 1-1: DSM | Youtube (manually corrected framerate) | Youtube (1 screen version) Still a second behind the world record but I'll get there.. Let me know what you think of the 1 screen version, not sure if the larger resolution is worth the extra confusion. Also, I noticed on one of the lag frames, the rings rotated 1 frame despite nothing else in the game moving.

.II

Posted: 5/19/2009 8:20 AM

Quote

upthorn

He/Him

Emulator Coder, Active player (388)

Joined: 3/24/2006
Posts: 1802

Comicalflop wrote:

Rather than slowing down before the ring to make it spin less, you may want to make the run go for fastest in-game time, which some other sonic TASes do.

This. And I believe that all published Sonic TASes do this

How fleeting are all human passions compared with the massive continuity of ducks.

Posted: 5/19/2009 8:38 AM

Quote

Flygon

Joined: 11/4/2007
Posts: 1772
Location: Australia, Victoria

Just watched the upload, and goddamn, that... has to be some of the fastest Sonic gameplay I've seen, even compared to the Sonic 3 and Knuckles TAS up.

Posted: 5/19/2009 8:48 AM

Quote

mike89

Active player (301)

Joined: 8/8/2005
Posts: 296
Location: NSW, Australia

I hate to mention it, folks, but 37"93 isn't even close to the real-time record. I'm not sure how much the route has changed from the video in question, but I can't imagine it'd be much.

Posted: 5/19/2009 5:23 PM

Quote

TakaM

Joined: 3/25/2008
Posts: 64

Those times are insane Mike, I can't really see where I can shave 2 seconds off but I guess it's there somewhere... Zone 1 Act 1 third attempt: DSM | Youtube (manually corrected framerate) Scratching away at that timer...

.II

Posted: 5/19/2009 7:39 PM

Quote

Derakon

Joined: 7/2/2007
Posts: 3960

How much validation does Sonic Center do? I could easily see people lying about the times they get just so they can be at the top of the list. Alternatively, if the people at the top are known to be good speedrunners of other Sonic games (where the runs can be more easily validated than on the DS), that lends their times more credibility. Of course, given that link ComicalFlop gave to a verified run that's faster than the TAS, it's clear you still have some more improvements to find. :) I can see how figuring out where they are would be difficult given the quality of the "encode" of that video, though. I don't suppose it's possible that the emulator's timer is off compared to reality? It should be possible to check by timing some simple action (like running horizontally from the start of the level until you hit some landmark) in the emulator and in the game.

Pyrel - an open-source rewrite of the Angband roguelike game in Python.

Posted: 5/19/2009 8:37 PM

Quote

moozooh

Senior Moderator

Joined: 8/4/2005
Posts: 5770
Location: Away

Even if the emulator lags, the in-game timer will lag together with the game. They are inseparable.

Warp wrote:

Edit: I think I understand now: It's my avatar, isn't it? It makes me look angry.

Posted: 5/19/2009 10:16 PM

Quote

Derakon

Joined: 7/2/2007
Posts: 3960

For normal lag, yes. I was suggesting that there might be an emulator bug. Obviously I have no evidence for such a claim, though.

Pyrel - an open-source rewrite of the Angband roguelike game in Python.

Posted: 5/20/2009 1:09 AM

Quote

Acheron86

Joined: 8/23/2008
Posts: 417

Derakon wrote:

For normal lag, yes. I was suggesting that there might be an emulator bug. Obviously I have no evidence for such a claim, though.

An emulator bug that specifically screwed up the in-game timer with no other noticable consequences is very unlikely. As far as verifying if such a bug existed, you'd just put a real-time speed run next to a TAS for a few seconds and see if the timer is the same at the same point in both runs (assuming the TAS copied the other run). I'm more inclined to believe some of those times are fictitious. I know Sonic runners are hardcore, but that doesn't mean they can't or wouldn't lie. On the other hand, I also know very little about the verification process of the Sonic WR site; if they require actual game video, that's sufficient evidence and we could just watch how it's done to know why the times now in the TAS are off by such a margin.

I will not use self-reference in my signature.

Posted: 5/20/2009 3:36 AM

Quote

TakaM

Joined: 3/25/2008
Posts: 64

There may be something weird with the timer, I have discovered two lag frames where minor things animate despite nothing else in the game moving. I've had the rings spin and some particles move while the timer was still on the same number. Does that mean anything?

.II

Posted: 5/20/2009 8:34 AM

Quote

Miles

Joined: 11/21/2006
Posts: 94

Four people have times that are low 36.xx; personally I think it's unlikely that all four of them are lying. (and one of them is the person with the 37.93 video) Also, Rush TASing has barely begun. I really don't think people should accuse the realtime records of being fake so quickly.

Posted: 5/20/2009 8:54 AM

Quote

TakaM

Joined: 3/25/2008
Posts: 64

Yeah, I highly doubt they're fake, although I wish they were :P

.II

Posted: 5/20/2009 11:28 AM

Quote

upthorn

He/Him

Emulator Coder, Active player (388)

Joined: 3/24/2006
Posts: 1802

Derakon wrote:

How much validation does Sonic Center do? I could easily see people lying about the times they get just so they can be at the top of the list. Alternatively, if the people at the top are known to be good speedrunners of other Sonic games (where the runs can be more easily validated than on the DS), that lends their times more credibility.

SkyLights is known to be credible.

How fleeting are all human passions compared with the massive continuity of ducks.

Posted: 5/20/2009 2:47 PM

Quote

Comicalflop

Editor, Experienced player (730)

Joined: 6/13/2006
Posts: 3300
Location: Massachussetts, USA

For your benefit, I found some addresses: (all of these are 1 byte, signed) X Speed: 020907E9 020907F5 Y Speed: 020907EB 020907F1 I'd watch all 4, just to make sure that in any situation you're watching one that is more accurate. If you're constantly falling behind the realtime records, that probably means that you're route is flawed. I would watch the video that I linked to, and maybe join TSC forums and ask what the route was used to achieve the 36.'' times. Also, I did some testing and I think you should be using air dashes in the same way that Nitsuja does in the Sonic Advance 2 run, where you jump and immediately air dash. It brings your speed to 14 briefly, and higher when you are moving at even faster speeds, and I think it'd be a helpful move that will enable you to beat the realtime records. 'A' 1 frame then 'R' the next frame seems to be most optimal. If you want, I can give the first level a try, and see if I can find any improvements or optimize some stuff.

☣ Homepage ☣ Retired

Posted: 5/20/2009 3:01 PM

Quote

TakaM

Joined: 3/25/2008
Posts: 64

I'd love to see someone else make a run at this. And I found Skylight's Youtube profile, he's on regularly so I just sent him a message there. edit- Just tested out spamming the air dash, got to the first spring pad several frames earlier, I didn't really optimize much, so it looks like this will really help. I was kinda hoping that kinda movement wouldn't be needed, but oh well :P

.II

Posted: 5/20/2009 5:58 PM

Quote

Comicalflop

Editor, Experienced player (730)

Joined: 6/13/2006
Posts: 3300
Location: Massachussetts, USA

You might want to put this on hold for the moment; I'm finding a ton of game breaking glitches and exploits that has me currently a full 2 seconds ahead of the 37.93" video, barely 10 seconds in. I'm going to keep experimenting and working on lvl 1 and see what the final time is. Optimal air dashing requires the tension boost to be used throughout for it to be effective, since it will prevent your speed from going down to 10, but rather keep it at 12 after the first 1 frame 16 speed burst. So far, I've only used it in succession at the beginning of the level, and it's not worth it once you're above 13 speeds.

☣ Homepage ☣ Retired

Posted: 5/20/2009 9:20 PM

Quote

josh_l.

Player (101)

Joined: 3/25/2004
Posts: 189
Location: Austin

Man, if you guys get through the stages TOO fast we won't be able to hear much of the awesome soundtrack :( Seriously though, looking forward to this.

http://girlyyy.com/ The El Viento TAS is the only decent thing I submitted here.

Posted: 5/21/2009 10:40 AM

Quote

TakaM

Joined: 3/25/2008
Posts: 64

Comicalflop has uploaded his WR shattering video on Youtube Awesome run, couple questions though: at around 13 seconds you do a much more vertical jump than I normally do, was this necessary? You can avoid moving onto the top screen which I think costs you one frame every time you switch screens. at around 20 seconds (in game timer) you miss the invincibility powerup, the reason I pick it up is for the enemy after exiting the cable spin thing inside the tree trunk, I think it's faster to use him to get up (only possible with invincibility) rather than have to use the second boost ring, I guess not? at around 26 seconds, you jump down into the ground, is it quicker than jumping over the enemies and avoiding the collision pause? Also, when doing tricks, try to end every string of tricks (before landing or using R to double jump/start a new string) with the A button to maximize points and earn extra boost. Awesome run btw.

.II