WST
She/Her
Active player (490)
Joined: 10/6/2011
Posts: 1697
Location: RU · ID · AM
This is just for fun. Sadly, I don’t have better level maps (but already looking for). Link to video
S3&A [Amy amy%] improvement (with Evil_3D & kaan55) — currently in SPZ2 my TAS channel · If I ever come into your dream, I’ll be riding an eggship :)
Editor, Player (69)
Joined: 1/18/2008
Posts: 663
Feel so old...and small... I like this but it makes me sad :(
true on twitch - lsnes windows builds 20230425 - the date this site is buried
Joined: 8/21/2012
Posts: 29
Location: Russia, Moscow
Monitors... Gotta catch 'em all! Link to video
WST
She/Her
Active player (490)
Joined: 10/6/2011
Posts: 1697
Location: RU · ID · AM
What the hell happens at 29:58?
S3&A [Amy amy%] improvement (with Evil_3D & kaan55) — currently in SPZ2 my TAS channel · If I ever come into your dream, I’ll be riding an eggship :)
marzojr
He/Him
Experienced player (762)
Joined: 9/29/2008
Posts: 964
Location: 🇫🇷 France
WST wrote:
What the hell happens at 29:58?
There are hidden static monitors all over Scrap Brain 2; these hidden monitors have the same role that invisible solid objects do in other levels. It is probable that either the level designers for SBZ2 didn't know about the invisible solid objects or that they didn't exist when objects were placed in SBZ2.
Marzo Junior
Player (178)
Joined: 5/13/2012
Posts: 51
I TASed Green Hill 2 in a rather unconventional way :D
WST
She/Her
Active player (490)
Joined: 10/6/2011
Posts: 1697
Location: RU · ID · AM
Nice. I did exactly the same (in this game and also in some hacks), but never put it into real use :) For the viewer it looks like Sonic went to Earth orbit :) You may want to see one of my glitchfests (Amy in Sonic 3), also showing an «orbital» way of beating the level :)
S3&A [Amy amy%] improvement (with Evil_3D & kaan55) — currently in SPZ2 my TAS channel · If I ever come into your dream, I’ll be riding an eggship :)
Joined: 9/22/2013
Posts: 7
Location: Ukraine
http://tasvideos.org/userfiles/download/12421200089184835 Looks like you can hang on right slope of tube for a while.
WST
She/Her
Active player (490)
Joined: 10/6/2011
Posts: 1697
Location: RU · ID · AM
Gvozdevik wrote:
http://tasvideos.org/userfiles/download/12421200089184835 Looks like you can hang on right slope of tube for a while.
I beleive this is known; it often happens to Amy, but since she’s able to climb higher in that tube, the terrain does not allow her to jump.
S3&A [Amy amy%] improvement (with Evil_3D & kaan55) — currently in SPZ2 my TAS channel · If I ever come into your dream, I’ll be riding an eggship :)
Joined: 9/3/2011
Posts: 26
Why isn't the GH2 wrap consistent? After getting the level wrap, Sonic will always be at the right edge of the level. You need to press left at a certain point in time so that Sonic runs into the tube and rises above the lower level boundary. That's my understanding of this, but I must be missing something; pressing left the exact same number of frames after getting the wrap will sometimes work, and sometimes not. I don't have a clue as to why.
WST
She/Her
Active player (490)
Joined: 10/6/2011
Posts: 1697
Location: RU · ID · AM
Naegleria wrote:
I don't have a clue as to why.
This game has quite complex physics; depending on your X/Y subpixel positions, you may stuck in the tube longer or shorter (within the same number of frames when you press left).
S3&A [Amy amy%] improvement (with Evil_3D & kaan55) — currently in SPZ2 my TAS channel · If I ever come into your dream, I’ll be riding an eggship :)
Joined: 9/3/2011
Posts: 26
Can the x/y subpixels make this trick literally impossible? Or is there always atleast 1 frame where Sonic is above the boundary?
Expert player (3645)
Joined: 11/9/2007
Posts: 375
Location: Varberg, Sweden
Naegleria wrote:
Can the x/y subpixels make this trick literally impossible? Or is there always atleast 1 frame where Sonic is above the boundary?
If you can tell exactly how many frames after the level wrap you start holding left, and for how many frames you hold left I can give the likelyhood of it to succeed. And FYI only the y-subpixels are variable here, x-subpixels reset after a level wrap.
feos wrote:
Only Aglar can improve this now.
Joined: 9/3/2011
Posts: 26
In a TAS setting I can tell the exact number of frames, but in real time I use visual cues in the level as the camera scrolls by. My goal here is to find the most consistent way to perform this trick in real time. I hold left, and keep holding left, until I hear Sonic in the tube. Holding left is the easiest method, but does holding left and then jumping over the upslope before the tube give a better chance of success? I'm not sure if it slows Sonic below the speed cap upon entering the tube, or if jumping over the slope messes up the Y subpixels so that theres actually worse chances of surviving. For any given Y subpixel, is it still possible to dodge the lower level boundary? If so, whats the number of frames where Sonic is above the boundary and will not be killed?
Expert player (3645)
Joined: 11/9/2007
Posts: 375
Location: Varberg, Sweden
So far I've gotten that if your y-subpixels are between 201 and 235, then you will survive if you just hold left all the way, but yeah the odds of this is only 35/256; you have a nice 24 frame window to pull it of though. It'll be hard to do further analyses of strats that can consistently be pulled off in real time, but if you have certain interval for the number of frames you push left I can quickly check if there's a consistent pattern for these.
feos wrote:
Only Aglar can improve this now.
Joined: 9/3/2011
Posts: 26
My visual cues to press left are the following: 1. When the crab located shortly after the speed shoes goes off screen. (I try to time it so its when the crab is half-onscreen and half-offscreen) 2. The same thing, but using the ice block instead of the crab. 3. When the screen stops scrolling down after getting the wrap. I've had varied success with these. Both of which I press and hold left until the screen is past the waterfall. Does pressing left+jump increase the frame window? Does jumping over the upslope increase the frame window?
Joined: 8/4/2011
Posts: 26
Location: Sweden
Aglar, I'm interested to know how you find these boundaries for the subpixels. I didn't think y-subpixels mattered for this trick so I went ahead and tried to come up with something using TeeNTee's latest S1 TAS (14:25.23) Just holding left worked for no frames at all in his TAS (Y = 943:250) and instead tried to see what ranges I could get out of the release-method TNT used. My results were (all frame data is in 38XY format):
LEGEND
*******************
Inputs: left, release, left

------------------------
frame of first input
------------------------
frame of release
range where holding left again allow you to clear the level
*******************



------------------------
left at 46
------------------------
release at 48
nothing

release at 49
55-65+

release at 50
58-65+

release at 51
nothing



------------------------
Left at 47
------------------------
release at 49
nothing

release at 50
leniency for left: 55-65+

release at 51
leniency for left: 58-70

release at 52
nothing
------------------------



------------------------
Left at 48
------------------------
release at 50
nothing

release at 51
55-65+

release at 52
59-65+

release at 53
nothing
------------------------



------------------------
Left at 49
------------------------
release at 51
nothing

release at 52
59-65+

release at 53
59-65+

release at 54
nothing
------------------------

------------------------
Left at 50
------------------------
release at 52
nothing

release at 53
57-65+

release at 54
59-65+

release at 55
nothing
------------------------

(the same pattern, wait 3-4 frames then having a big window for left again seems to be reasonable to assume for a number of frames ahead but 4 frames is plenty for consistency so I do not bother checking them)


x-position at random wraps 8231:0 always
This seem to work well for arbitrary y-subpixels but again, I don't have a method of finding the ranges other than randomly jumping on the wall before wrapping and hoping I get a weird value so it's hard to tell when I'm playing on console if my inputs are bad or how much of an effect subpixels give. It's also 4 am and I don't feel like playing Sonic right now.
WST
She/Her
Active player (490)
Joined: 10/6/2011
Posts: 1697
Location: RU · ID · AM
You can always force your Y-subpixels to certain value from a Lua script if you need to perform some kind of research. At least this is how I do it myself, when I need to analyze possibilities of doing certain trick before starting to do in in real TAS.
memory.writebyte(0xFFD00E, 255)
Using this way, you can find which range of Y-subpixel values give you ability to perform certain trick (and even do it automatically, if you can write a condition which is true, if your trick succeeded, and false otherwise).
S3&A [Amy amy%] improvement (with Evil_3D & kaan55) — currently in SPZ2 my TAS channel · If I ever come into your dream, I’ll be riding an eggship :)
Joined: 9/3/2011
Posts: 26
How would I set a lua script up? I've done it with dolphin before, but someone else already wrote the script.
Joined: 8/4/2011
Posts: 26
Location: Sweden
Ah ,thank you WST. I always wondered how you forced bits to change in Gens rr because I thought lua was redundant and that you could do it in the memory value search somehow. Naegleria, if you use gens_11b (-s1) that you can find on this site, just create a new text document and paste the code there. Then you go into Gens > Tools > Lua scripting and open your text document. Look around on this site to find useful and fun Lua scripts! EDIT: So I tried out the method I posted in my last message and I'm a bit disappointed. After you do the wrap and the left/release/left input you keep holding left until there's a string of frames where your y-(sub)pixels are stable. In this range, the subpixels that work are 0-117 and 227-255, meaning you have 57% of all subpixels that work. Not a whole lot considering the inputs are hard as is. The ranges seem to be independent of the frame you start to hold left though.
Joined: 9/3/2011
Posts: 26
GH2 seems to be the most reasonable level to hunt for consistent wrap strats, once we have something for that, I think we should look for a method on the L3 and SB2 wrap. I've looked into replicating the L3 wrap, but I'm not sure that one is feasible for realtime runs. EDIT: https://www.youtube.com/watch?v=-CJ4nnoLB4k Marble 1 wrap. Is getting hit into the wall just a matter of getting the timing and being on the right pixel, or are sub-pixels relevant here? Are sub-pixels relevant for the wrap in M1, M2, M3, L3, SL1, SL2, or SB2? EDIT #2: For the Marble 3 TAS route, I couldn't get the same X velocity the TAS gets (8000+) when attempting to replicate it. I would press Right on the same Y pixel and left on the next frame, played around with it, and could never get more than 6000. Whats happening here?
Joined: 9/3/2011
Posts: 26
Scratch that, I've figured out most of that. Now my only questions are how to perform the SB2 wrap (from spinning around on the wheel) and the subpixels required for it.
WST
She/Her
Active player (490)
Joined: 10/6/2011
Posts: 1697
Location: RU · ID · AM
Naegleria wrote:
Scratch that, I've figured out most of that. Now my only questions are how to perform the SB2 wrap (from spinning around on the wheel) and the subpixels required for it.
That particular trick is based on skipping the 16-pixel-wide wall between X=506 and X=522, which means that your X pixel position should not appear within this range. Thus, more speed you have — more chances to have this trick done.
S3&A [Amy amy%] improvement (with Evil_3D & kaan55) — currently in SPZ2 my TAS channel · If I ever come into your dream, I’ll be riding an eggship :)
Joined: 9/3/2011
Posts: 26
Hmm, I noticed the TAS taps left on the second wheel... is this to set the pixels properly to cross the wall? If so, what are the odds of getting this trick if you attain the same speed but dont have the same pixel alignment?
WST
She/Her
Active player (490)
Joined: 10/6/2011
Posts: 1697
Location: RU · ID · AM
Yeah, the purpose of that single press is proper alignment. 16 pixels = 4096 subpixels. So, if your speed is 4096+m subpixels per frame, the probability of having the trick done is m/4096. Which is high enough only if the speed is greatly more than 4096. However, there are some other factors affecting your success, which can make the probability higher. First, I beleive, there is a const maximum speed that you can gain there. Second, the speed after the jump cannot be affected by pressing d-pad keys. Third, there can be a point on the wheel, giving you quite consistent result (with the maximum speed). Amyway, I cannot be sure ­— I am a tool-assisted speedrunner, and I addicted to having full control of the situation — if I fail, I just redo until success… So I doubt I can really help you. The best people to ask are Tee-N-Tee, mike89, wersterlobe, Fuzzerd… I’m sure they have their own technic of doing certain tricks… Because all these digits and subpixels are interesting only when you are making a TAS, but they mean nothing when you are playing with your controller… (upd) I just realized that my logics were wrong; I am not good at all these calculations, as I completely forgot everything about probability calculations excepting that it cannot be >1
S3&A [Amy amy%] improvement (with Evil_3D & kaan55) — currently in SPZ2 my TAS channel · If I ever come into your dream, I’ll be riding an eggship :)