rWS NG 3lap

[calm ether]

This is the an old project, that I don't think has been worked on since this current rkg was sent to the other server sometime in 2023. Has been optimized till the turn before the zipper/after shroomless sc on lap 1. Previously a lap 1 was completed, but it is being redone due to finding a faster strat on the bridge ramp (the one used in this rkg).

You're welcome to improve on stuff in the rkg that has already been optimized. I think a better startslide has been found, but I couldn't keep the lead due to good QM on the first shroomspot I haven't been able to replicate. Not an expert in aligning QM between frames, so if the first mount can be cleared with a good lead this might be worth looking into.

Should add that I might return to this when I get time and energy to put in the effort, but I'll likely just continue from the point where the ghost is optimized to.

[light elbow]

i have a 1f improvement at the shommless sc ( by agc-ing with the flap )

and the turn before the zipper is almost synced with your old ghost that got a good turn and a good zipper

[glossy dust]

Great job!

[light elbow]

Thanks !

[light elbow]

finally got the stick at the turn before the zipper ( vs my previous ghost )

vs the flap for the zipper line

[light elbow]

improved the zipper, got the good turnskip, and synced the flap bump strat and synced rs ending

there might be a better strat at the landing of the warparound after the bump, but i couldn't get anything better that had a good angle for last turn

is rs ending qualified NU ?

[calm ether]

Welp, I got news that I used SG on the last ramp unknowingly, so it doesn't count as NG, but it should still count as NU.

Also, awesome job!

[light elbow]

Thanks !

[light elbow]

ok, i will try to find an other last ramp strat

[calm ether]

Also, in case you're trying to sync up ghosts for lap 2 and 3 in the NU category, then this already has a good first bump that comes right off after the lap 1 ending you synced, so all the work needed here should be syncing (I've never synced runs before, but I did dig this up).

Also I'm just going to list the contributors I could find over the years by looking into the rksys archives (will be updated):
RS, Alego, Ejay, Wade, Kierio, Marth, theancientpoop

some additional contributors from the flap bumps:
ellucinea (a bit uncertain on this one), LirWm, ego

[abstract orbit]

doubtful i have any inputs if the zipper has been improved again

[languid meteor]

My improvement to the bumps wasn’t used because of

[glossy dust]

holy shit

32.0

[glossy dust]

what makes it interesting is that ODB could be slower on rWS, which means this SG could actually be the fastest strategy for NU 3lap

[calm ether]

Yeah, but I don't have experience with ODB and especially superhopping, so I'll not be able to confirm this quite easily.

[light elbow]

got a good ramp with almost no sg input ( only 2f )

[glossy dust]

i think it's possible to get a lower landing

[light elbow]

when i tested it was slower

[glossy dust]

is that speedometer XZ

[light elbow]

yeah

[glossy dust]

it should definitely be faster to land at like >90 XZ than 85 XZ or something

[light elbow]

i'm gonna try

[glossy dust]

wait a bit

[glossy dust]

sending

i improved even more upon the above ghost

rWS NG 3lap - Improved first wraparound landing and carried until boost ramp

@light elbow

[light elbow]

Great job !

[glossy dust]

thank you

[glossy dust]

vs alego

obviously the ramp didn't quite work, but should be able to save 1f from this

might be able to save more time if it's possible to get fewer airtime frames at ~25.2

ideally we just get 1f or 2f airtime, which means we can reo drift immediately and wheelie sooner

[glossy dust]

I think I'm 0.014 ahead, so we should definitely be able to save 0.002

[light elbow]

resynced ending with ejay ghost

[glossy dust]

hell yeahhhhh

good stuff

[glossy dust]

i think this ending will also be useful for UR 3lap (laps 1 and 2) (if ODB doesn't work out), if you adjust alignment after landing to be more leftwards

[light elbow]

Thanks !

[light elbow]

more leftwards to be able to start a sg at the landing ?

[glossy dust]

No

More leftwards to bounce off the left fence at the bump

So you can turn around, shroom, and do the banner bypass / ultra

I basically just mean crossing the finish line on the left side of the road

[light elbow]

oh ok

[glossy dust]

Does that make sense

Have a look at the bkt if not

[light elbow]

yes 👍

[light elbow]

i asked this because i read NU instead of UR lol

[raven wyvern]

Fwiw just looking at who worked on the TAS won’t necessarily lead you to an accurate author list — I can help make an accurate one for the eventual name display if needed (I have past research into the old lap 1 authors)

[calm ether]

ok, cool. My memory of it all has been fading for a while.

[light elbow]

improved a bit the ramp, need to get a better stick at bump l2

[light elbow]

also if ramp can be improved by 1~2ms it could sync with rs l2 first bump

[rough needle]

~3 ms trick improvement
part of it is by landing further right, but i also gained ~0.8 xz ev more before the trick. i saw that you already had 4-5 xz so i left it. not sure where the line should be drawn in regards to nu vs ng

[light elbow]

got good first bump trick with corvid ghost, ghost is rs l2 from this ghost #1317929657567416413 message

[rough needle]

very nice. good luck on resyncing this monster of a run

[glossy dust]

It looks slightly not as good as RS's one

But very good attempt

[rough needle]

does it? to me it looks like he's goes wider while shrooming, catches up to rs before ramp, then has a worse line after the rfh trick part (which is fixable)

[light elbow]

after rfh i can have the same line as rs, i tried this line to get a tigher exit for the big ramp rfh

[light elbow]

Thanks !

[light elbow]

ty !

[light elbow]

btw this ghost is ahead by 0.5f to l1 for the big ramp, so or we would need to lose that time, or manage to gain enough to catch the next frame qm

[glossy dust]

Yes, but I think the QM at the end of the mud before the rfh means that he catches up to RS anyway (as you said)

So I don't think it matters

[light elbow]

improved ramp + synced big ramp rfh, shroomless sc turn + zipper turn

will focus on syncing the zipper a bit later today

[glossy dust]

I noticed when watching the zippers back that the wheelie hops aren't perfectly spaced

As in, some hops have a few more frames of airtime than the others

I wonder if that's a big deal

[calm ether]

I remember from when I optimized the zipper that the fastest rehops didn't align with the ending too well, so I had to compromise.

[raven wyvern]

Yeah spacing is just as important as speed here

[light elbow]

synced zipper, and almost synced bump after, still need to get the 2nd bump low trick ( ghost is l1 )

[smoky talon]

[glossy dust]

That 2nd bump trick is so tough

[light elbow]

yeah

[light elbow]

ghost is l1

got the 2nd bump trick, now need to get a good warparound, will try tomorrow

[glossy dust]

Oh shit you got the ultra low trick

We tried hard to get that on the flap and nobody was able to lmao

I wonder if it's not as good though, for the reason that we might need to be in a boost during the wraparound for it to work

[light elbow]

you can modify some input before the trick to get a one like the l1

[light elbow]

got a good warparound ( even better than l1 )

i think it might be possible to save a frame compared to the l1 one

[light elbow]

got a frame of timesave compared to l1, need to improve a bit before first boost ramp of l3 to get same qm as l2

[glossy dust]

do you plan on saving that frame on lap 1 as well or nah

[light elbow]

i dont think it possible, the qm is too different beetween l1 and l2

[glossy dust]

yeah

good work!!

[light elbow]

Thank you !

[rough needle]

the first rfh trick on l1 loses over a frame compared to l2 too

but i guess it's late to fix these two things

[light elbow]

yeah

[light elbow]

i mean i could, but re-resyncing l2 and start of l3 will be paintful

but it's maybe not possible anyway because of little bump and first boost ramp qm

[rough needle]

i don't see why it wouldn't be possible. it's prob not worth to try though. imo, the current bkt is so outdated it's better to just finish l3 and perfect it in the future with better tools, maybe some new strat gets found too

[crude prism]

How come?

[light elbow]

synced first boost ramp, big boost ramp, shroomless sc and zipper turn

[light elbow]

synced zipper + turnskip + bump trick + ending almost as fast as the flap ( 1 or 2 ms slower )

[glossy dust]

it's real

rWS NG 3lap is real

RS won't believe his eyes when he wakes up

currently a 1.962 BKT timecut

[rough needle]

optimizing the ending is kinda pointless because the line between ng and nu is very blurry

[glossy dust]

SG inputs on the lap 3 ramp don't really help with speed

it's to lower your trick height

[light elbow]

yeah

[glossy dust]

we can at least try and get it to be equal to the flap

fantastic work alego. you're literally revolutionising TASing

[light elbow]

i mean 4~5 ev is pretty low ( for l1 and l2 trick ), i think the flap rfh high trick entry is more than this iirc

[glossy dust]

a new era of efficiency

[light elbow]

Thank you Ejay !

[crude prism]

Which is fucking crazy, that's like perfect usage of AGC with crazy skills at TASing to match it that fast

[calm ether]

This is amazing!

[glossy dust]

we need rs to save the final millisecond now

to complete the story

[acoustic atlas]

this is so sick, fantastic work alego once again fr

[abstract orbit]

alego are u god

[rotund crest]

Fr. Alego will be the future best TASer

If he isn’t already

[raven wyvern]

Alego is Akari if Akari came back

[crude prism]

He is already

[light elbow]

Thanks !

[glossy dust]

Are there any sections of the flap that can be improved?

Now that lap 3 of the 3lap has been (almost) finished

[light elbow]

i dont think so, ( maybe the warparound, but i'm not sure )

[glossy dust]

in such a short time you've fully optimised rSGB NG, MC NG, rWS NG, and rDH NG

superb

all 8 categories

[light elbow]

saved 1ms on the ending, it's 0.000002 race competion away from the .392

[glossy dust]

@calm ether

Now's your time to shine

[crude prism]

delay the last wheelie by a frame (im very serious)

it's gonna save 0.00000001

[rough needle]

higher racecomp is better (you have crossed further into the finish line)

i recommend using the timedifference (finish) section though

[crude prism]

so it means i'm slower here?

[rough needle]

yes

[crude prism]

woops

then I lost the inputs where it was higher 😭

[rough needle]

you should always check this part instead of racecomp

[crude prism]

yeah I should but for some reason i messed up with ghost inputs so couldn't compare with it

let's let RS cooks

[calm ether]

I do have a small improvement on the ramp, but I need some time to figure out how not to lose it all on the landing

[calm ether]

might have to check for illegal inputs, but I'll let others be the judge.

think this should be correct rkg

[crude prism]

how does it feel for you to be done with rWS?

[light elbow]

Nice job !

[rough needle]

there are no illegal inputs. were you talking about the +7 -7 alterations? i think that was dolphin's input smoothing making that seem illegal

[calm ether]

I was worried about accidental sg

[rough needle]

oh yeah i am worried too there is zero consensus lmao

just keep doing your thing mate someone will figure it out

[calm ether]

gained .000005 race completion (over previous image, ghost should be this one)

not sure how much race completion is needed for a .001 save

[glossy dust]

Sick

How much XZ EV is gained here compared to the NG flap

[calm ether]

I don't know about flap, but the current ghost I sent goes up to 6 XZ EV on the ramp (and 9.4 XZ EV in the air, but that isn't due to supergrind). I also lost some XZ speed at the end of the ramp, where I did the +7 -7 alterations for alignment, and I don't remember doing sg like inputs earlier during the rfh. I think it can be safely put into ng category.

[glossy dust]

yeah that's super fine lol

[calm ether]

XZ EV goes down to like 2-3 ish when I exit.

and a few frames after down to 0.5

If anyone is willing to try to save anymore milliseconds, then go ahead. If not then I think it's about time to consider the project finished.

[raven wyvern]

I might take a look 👀 (I’ll probably be unsuccessful but I just wanna try my hand at getting in on this epic project)

[glossy dust]

ofc

[calm ether]

get the name in the game

[glossy dust]

classic kierio improvement

[calm ether]

keep the meme going

[raven wyvern]

When was the last time I even did that

Was it in June?

[glossy dust]

last year

[raven wyvern]

Shy Guy beach I think

[glossy dust]

yeah wasn't sgb the last collab tas you worked on

[raven wyvern]

came out of hiatus briefly just to save 0.001

I think so

All my other credit since then has been from work I did years ago

[glossy dust]

case in point

[raven wyvern]

that too

[calm ether]

ok, I'll wait a bit before calling this finished then.

[raven wyvern]

this ghost is 0.658356 ms from 390

ive got to 0.486389 ms from 390

i think i can still push further

[graceful torrent]

The 0.0002 improvements

[raven wyvern]

this one is 0.461426 ms away, i cant seem to do any better

i dont trust that ive reached the limit though

i got the big improvement from .6 to .4 rather randomly so it feels very likely that it could happen again, i just haven't found it

[graceful torrent]

The 0.00002 improvements

[smoky talon]

HUELIGEEEEEEEE LET'S GOOOOO

[abstract orbit]

[raven wyvern]

but uh yeah im done as much as i can

so someone else will need to try to get the 390 🫡

[novel hollow]

wait, how do you see how close you are to the next ms

[raven wyvern]

Exact finish

The reason that one code needs to be active for infodisplay to work

[calm ether]

So, since it's been a few days since the Kierio improvement, are we ready to call this project finished?

[light elbow]

yeah i think

also who upload the video ?

[rough needle]

i took a brief look at it some time ago and couldn't improve at all

[calm ether]

Well, I'd like to since I haven't posted a video in a while
Otherwise I'd go with you due to your effort in syncing up with the ghost

[raven wyvern]

Was this before or after my improvement?

[rough needle]

after i think

[raven wyvern]

Alright

[calm ether]

I should just record with XZ speed-o-meter right?

[glossy dust]

vehicle speed

IV

[calm ether]

You'd say that's better?

[glossy dust]

we're not using any exploit that increases XZ speed but not IV

XZ speed is very jittery and hard to analyse/watch back

so there's no reason to use it for a no glitch run

[graceful torrent]

IV gives much more information about what the vehicle is actually doing

And xz/xyz are only really useful in comparison to iv as well

[calm ether]

I don't have a good IV speed-o-meter code

[glossy dust]

#resources gecko codes

[calm ether]

Oh, wait I actually had the code, but mistoke it for XYZ speed-o-meter

How many decimals would be ideal for this?

I'll mess around with the position for a bit too

[glossy dust]

with 2 dp you can see wheelie turning patterns which is nice

[graceful torrent]

10

[abstract orbit]

does rws use wheelie turning

[graceful torrent]

All the decimals

[glossy dust]

maybe before the last turn hop

do you plan on doing name display rs

[calm ether]

Yes, this is a first for me though. Also a first for me with a new editor.

And a while since I last recorded, so it may take a bit before a video is ready.

[abstract orbit]

2 dots per

[calm ether]

Oh, wait, the ghost should be sent for approval somewhere first, right?

[abstract orbit]

no i think you can do it with the video

verification w video and rkg

[calm ether]

yeah, but I need the rkg verification in the system

[glossy dust]

wdym

[calm ether]

I should create a ticket or something

[raven wyvern]

yes! thanks for sending your run in now rather than later

we should always be sending a run in when the recording process starts rather than when it finishes

because that's the point at which we've decided the run is complete

[calm ether]

I'm running into the problem of not being able to keep track of all the authors

Now that I've created the ticket

[raven wyvern]

send the ghost we can figure out authors later

[calm ether]

ok cool

I can figure out the the author list, as I'll have to do name display anyways.

[raven wyvern]

i can help if you need

i have an author list from the old wips already

[calm ether]

I'd gladly take it.

[glossy dust]

Ejay inputs every lap:
landing from boost ramp trick -> miniturbo release after shroomless dirt cut

miniturbo release before zippers -> landing from first stunt ramp trick

second stunt ramp trick landing -> finish line

[calm ether]

I'm not too sure how many inputs alego has taken from me for AGC syncs

Otherwise I think I have a hand in the last ramp and everything from beginning till end of zipper each lap

[raven wyvern]

if i understand how the syncing process went, you should still have inputs in the entire tas

though the part near the end may have been different because of that new stick strat

so maybe everything except that?

[calm ether]

Not on the bumps after the zipper and the wraparound the turn after those bumps I think

[raven wyvern]

yea

oh thats what you said already

[glossy dust]

Alego synced the zippers and double trick from the NG flap, which is all my input

[raven wyvern]

haha

[glossy dust]

And he probably changed some frames here and there to make the AGC sync perfect

[calm ether]

it this a good spot for the speed-o-meter?

[raven wyvern]

i would say yes

if you're including input display, that would go on the left

so right works for speedometer

also it's been a while since i've seen someone refer to it as "speed-o-meter"

i wonder when that changed

[calm ether]

what does people say now instead?

[raven wyvern]

just "speedometer"

no hyphens

[glossy dust]

people have always said speedometer without hyphens

that's the first form of the word

[raven wyvern]

yeah it's just the gecko code used to be called "Speed-O-Meter"

the one that went in the milliseconds

i guess it must have changed from like 2019/2020 onwards

when the "Speedometer XYZ" code started being used instead

[glossy dust]

I literally have never heard anyone call it that besides RS

I don't think "it changed"

[raven wyvern]

ejay did you just delete all instances of you saying it

tis gone

thats also why people call it som

[glossy dust]

*call it that repeatedly

[calm ether]

how do I get the wii-wheel on the ghost display?

[raven wyvern]

the full input range code on the code document in #resources comes packaged with an extra code which saves all ghosts as driven with wheel

[calm ether]

@glossy dust How do I edit this text back?

[raven wyvern]

https://www.tt-rec.com/msgeditor/

[glossy dust]

lmao

[raven wyvern]

search for "ghost has been created" and you should find the right one

[glossy dust]

i'm not telling you

[raven wyvern]

once you select the message, you can then edit the text

once youve put in your custom text, you press start then copy (to copy the code)

you can then add that code

if you want to remove the custom text entirely, just turn off whatever code youre using

[calm ether]

It's an old relic from RR and TF recordings

[glossy dust]

why is it on your dolphin

if i recorded them

[calm ether]

idk

[raven wyvern]

maybe you sent him your gecko code config once

and forgot that was there

[glossy dust]

nup

[calm ether]

not sure where in the gecko code it would be now, because I now have something close to the one from #resources

[raven wyvern]

hey ejay

i notice that youre typing

are you doing the same thing im doing

[glossy dust]

people who sent rWS NG files in rksys-archive

• RS
• Marth
• Ejay
• Kierio
• Wade

additional people who contributed in tas-files

• Ancient

additional people who contributed in here

• Alego
• Corvid

additional flap authors:

• Ego
• Ellucinea
• LirWm

[raven wyvern]

yup youre doing the same thing as me

if you had said something i could have saved you the time

im already on this

[calm ether]

guys...

[glossy dust]

it only took like a minute and a half

anyway it all depends on what sections alego referenced from the old lap 1 and the flap

[raven wyvern]

i have a pretty good idea on most of them

theres only one part im unsure about

(the zipper)

all he said was "improved the zipper"

#1317929657567416413 message

if that involved referencing the flap first, then ejay, rs, and ellucinea all get input as in the flap

if not then it's just alego (and maybe rs for the beginning of the zipper)

[calm ether]

Perhaps we should just ask
Hey @light elbow , could you please tell us from what you AGC synced the sections of the run from, thank you

[raven wyvern]

i was just meaning for that one part

[calm ether]

Also, what's the difference between recording on dolphin 5.0 and dolphin 4.0-8572?

[raven wyvern]

all other times it's really clear

[glossy dust]

Wade and Ancient don't have inputs in this 3lap

because RS redid lap 1

[raven wyvern]

alego referenced the old lap 1

so wade does have input

ancient's inputs were only in lap 2

so that was scrapped early

[abstract orbit]

doubtful that I have inputs

[raven wyvern]

you will if alego referenced the flap zipper

[abstract orbit]

I wanna say he redid the turn before

[calm ether]

I could try and record on dolphin 4.0-8572, but I'd need a different code config so that this dolphin doesn't crash

[glossy dust]

no he didn't

not for the zipper

[raven wyvern]

there was no turn before zipper to reference in the new lap 1

[glossy dust]

yes there was

[raven wyvern]

that only went to the shroomless cut

look at the first post in this thread

[glossy dust]

https://www.youtube.com/watch?v=vGs0MdPJE44

this was posted july 2022

rws redid from the startslide in may 2022

[raven wyvern]

nope

start slide improvement was scrapped

[calm ether]

Nevermind, just needed to fix the codehandler

[glossy dust]

but i'm still right anyway

he redid everything from the boost ramp onwards in april after finding the new boost ramp trick strat

[raven wyvern]

yes

he started to

and only got up to the shroomless sc

[glossy dust]

he got up to here

[raven wyvern]

only optimised* to the shroomless sc

alego agcd a run with a good turn before zipper and zipper, which could only be the old lap 1

the new lap 1 never got that far

this is the run rs started the thread with fwiw

[glossy dust]

#1317929657567416413 message

depends what "old ghost" means here

[raven wyvern]

.

[calm ether]

I've only gotten a good zipper entry once

so unless something got fixed the second time around, then there could only be one ghost of 'mine' to copy this from

[glossy dust]

i always thought you did the zippers on lap 2 as well before deciding the backtrack to lap 1

but upon viewing your lap 2 rkg in dolphin, it stops after the first zipper wheelie

[calm ether]

btw, for the recording, is it preferred to race against ghost and copy it so that name is displayed the entire run, or just have the name displayed in the beginning by not racing a ghost?

[raven wyvern]

i would recommend it not being displayed since you have name display instead

idk what other people like though

[languid meteor]

have the mii change to whoever is doing the inputs 👍

[raven wyvern]

more than one person can be doing inputs though

[languid meteor]

combine the miis

[raven wyvern]

NOTE: These are all just rough areas to help get a full author list, we can get more precise after to have accurate name display

A: 2020-2022 lap 1-2 collab

RamoelleuS: everything
Marth: first ramp to shroomless sc (excluding during shroom), bumps section to end
Ejay B: post-zipper to end, shroom l2
Wade: turn before zipper
ancient: first ramp and shroom l2

B: 2022-2023 new bridge ramp strat

RamoelleuS: bridge ramp and shroomless sc (overwrote marth's input?)

C: 2022 start slide improvement (unused)

Ramoelleus: start and first ramp trick
kierio04: start to first ramp

D: 2024 flap collab

Alego11: everything
Ejay B: bridge ramp to end
ego: double bump trick, reo drift to end
ellucinea: 6th zipper wheelie to double trick
RamoelleuS: end of 1st shroom to bridge ramp, zipper wheelies (agc'd from C)
LirWm: reo drift to end

E: 2024-2025 alego lap 1 resync

Alego11: bridge ramp trick and shroomless sc (agc'd from D), turn before zipper to end
Ejay B: bridge ramp trick and shroomless sc (agc'd from D), last turn to end (agc'd from A)
RamoelleuS: start to ramp, turn before zipper to double trick (agc'd from A), last turn to end (agc'd from A)
Wade: turn before zipper (agc'd from A)
ego: double bump trick (agc'd from D)
Marth: last turn to end (agc'd from A)

F: 2025 3lap collab (expanded for readability)

Alego11:

• bridge ramp trick and shroomless sc l1/2/3 (agc'd from D/l1/l2)
• turn before zipper l1 to zipper l3
  Ejay B:
• bridge ramp trick and shroomless sc l1/2/3 (agc'd from D/l1/l2)
• wraparound landing to last ramp l1
• last ramp to end l1 (agc'd from E)
• shroom l2/3 (agc'd from A/l2)
• last turn to last ramp l2 (agc'd from l1)
• last turn to end l3 (agc'd from D)
  RamoelleuS:
• start to bridge ramp trick l1
• turn before zipper to double trick l1/2/3 (agc'd from A/l1/l2)
• last turn to end l1/2 (agc'd from A/l1)
• last ramp to end l3
  Wade:
• turn before zipper l1/2/3 (agc'd from A/l1/l2)
  ego:
• double bump trick l1/2/3 (agc'd from D/l1/l2)
• last turn to end l3 (agc'd from D)
  Marth:
• last turn to end l1/2 (agc'd from A/l1)
  Corvid:
• last ramp to end l1/2 (agc'd from l1)
• start to first ramp l2/3 (agc'd from l2)
  ancient:
• first ramp and shroom l2/3 (agc'd from A/l2)
  LirWm:
• last turn to end l3 (agc'd from D)
  kierio04:
• last ramp to end l3 (0.000 improvement)

oh wait you submitted the ghost with my ending "improvement"

[calm ether]

yep 😄

[raven wyvern]

lovely, i get credit for saving 0.000

that's kinda nice because it means everyone who worked on rWS at some point between 2020 and 2025 gets credit in the 3lap

[graceful torrent]

Aw man I was hoping for surprise contributions like I get on rbc

Or rdh

[raven wyvern]

so yeah it's either 10 or 11 authors depending on whether alego agc'd the zipper section of the flap

[calm ether]

Yeah, so I can't record on dolphin 4.0-8572, because dolphin keeps crashing on me when I try to record in high quality, so I'm just going to record on my usual dolphin version.

[glossy dust]

what's your usual version

please don't record on 5.0 stable

you must use a newer beta version of 5.0 to record

[raven wyvern]

interesting fact, this will be the first TAS to have contributions from six different years

[graceful torrent]

Why submit to the weak beta version when I can be an epic alpha dolphin recorder

[glossy dust]

damn it he's using 5.0 stable isn't he

[raven wyvern]

fwiw they arent called beta versions anymore

[calm ether]

The one provided in #resources

[raven wyvern]

yeah lua core is based on 5.0 stable

it's fine for TASing but not the best for recording

use one of the release versions

[glossy dust]

what do you want me to say

the bitrate option first entered dolphin graphics config at some newer 5.0 beta version

[raven wyvern]

saying beta versions now will probably confuse people into downloading the dev versions

[glossy dust]

there's just no word that says beta on the site anymore?

[raven wyvern]

only in the blogs from before they switched to the release cycle

[calm ether]

I'll try to set up the recent dolphin release for recording.

[glossy dust]

https://cdn.discordapp.com/attachments/768082766104625194/1339723068318416956/image.png?ex=67cf65d6&is=67ce1456&hm=4a555bde90a692272e94be683035199ec97a53fe50df25808bd882d871f43163&
https://cdn.discordapp.com/attachments/768082766104625194/1339723163155697717/image.png?ex=67cf65ec&is=67ce146c&hm=57d6dabb81c8d2faa2992333825a457645f1bde53ade6508b4107f448a62f187&
https://cdn.discordapp.com/attachments/768082766104625194/1339723213630079037/image.png?ex=67cf65f8&is=67ce1478&hm=b33c591a0c7e0d2464c04f76b8a2caa57ba2b096cfb681c8bc879ee9902dc133&
https://cdn.discordapp.com/attachments/768082766104625194/1339723299302801519/image.png?ex=67cf660d&is=67ce148d&hm=cbba149d9496a590715b098782161b18218d624f07b2bf3d91bc79c894edd8c2&

if your monitor isn't 16:9 then Stretch to Window might need to be changed to another setting FYI

[calm ether]

my monitor isn't 16:9

I've been looking at this page https://www.tt-rec.com/howto/?page=recdolphin

though the newest dolphin has a few different options

Also, luckily I don't need to configure hotkeys and controller settings

seems like I have an extra option

Also, what custom textures and graphics mods are you using?

[glossy dust]

#tas-general message

[calm ether]

Where do I unpack this?
edit: found it nevermind, sry for ping

[abstract orbit]

same

[smoky talon]

Great minds think alike 😎

[glossy dust]

Load > Textures > RMCE01

[raven wyvern]

it took me ages to figure out which setting here is best

the settings i use are:

• Aspect Ratio: "Custom (Stretch)"
• Custom Aspect Ratio: 16:9
• Resolution Type: Aspect Ratio Corrected Internal Resolution

[calm ether]

https://www.youtube.com/watch?v=UsubhtSmYeo

I'll rest for now and look into the video quality later. overall image rendering and audio seems good, haven't checked for lags yet.

[calm ether]

ok, I checked for lags, there doesn't seem to be any, so no need for rerecording, yay

[glossy dust]

oh no

the 15 second menu screen is back

video looks good 👍🏻

[rough needle]

this is why no name display is better.

[raven wyvern]

name display or not i still needed to do that to figure out what all the authors were

[glossy dust]

you cannot in good faith pick the ONE tas that has continued for so long and use that as an example of why name display is bad

[raven wyvern]

this is one of the most extreme examples ever because the contributions were spread across SIX YEARS

[glossy dust]

when 99% of name displays are so easy to make

[rough needle]

nah ejay is right i'm just saying it to start shit lmfao

[glossy dust]

as hinted at by the full stop at the end

[rough needle]

i have talked to this guy too much

[raven wyvern]

I know it's already been recorded, but did we consider having the video be a race against the 1:37 (it would show how massive 2 seconds of improvement is)

If that doesn't seem like something that should be in the main encode, then that's ok, but I highly recommend that be made regardless

Sorry for not mentioning this earlier, I just completely blanked on that being a possibility

[light elbow]

i agced the zipper from rs finished l1

this ghost #1317929657567416413 message

[raven wyvern]

thanks!

@light elbow so just to be sure, you only agc'd the flap for:

• the bridge ramp and shroomless sc (lap 1)
• the bump sticks and wraparound (lap 1)
• the last turn and last ramp (lap 3)

everything else was either from RS's old lap 1 or recycled from previous laps

is this correct?

[light elbow]

yeah 👍

[raven wyvern]

awesome

tas is added to the site, just the video left and this project will be finally finished after over 1500 days :D

[glossy dust]

you changed your site name to kierio04?

[raven wyvern]

yeah ive been meaning to do that since i got site admin again (ive actually wanted to do it for several years now) but only remembered to do that today (when i edited rs to ramoelleus)

[calm ether]

Wait, does zak’s ghost actually sync?

[rough needle]

all ghosts on mkwtas.com sync

[light elbow]

yeah

[shut kestrel]

insane job everyone i do not understand how this syncing magic that alego did works but well done

bonkers that I get credit for my contribution, feels like ages ago

[calm ether]

I'm trying to implement input display, but I've noticed that the input display is a bit off at times. Right now while editing I'm going frame by frame, and it looks to me so far that there are no lags in the video, yet the input display recording seems to skip some frames, so I'm trying to stitch it together.

[raven wyvern]

Yeah that’s a result of input display being 60fps, and the framedump being 59.94fps

That’s easily fixed by setting the input display video to 99.9% speed

[calm ether]

I noticed it skipped a frame during a rfh on lap 1 bridge

so unless I manually edit it, the input display will be wrong at times.

[raven wyvern]

Hm that’s odd

[raven wyvern]

Could you try this anyway, I’m curious if that fixes it (your editor might be removing frames to try and compensate for different frame rates)

[calm ether]

And this is probably one of the easiest ones to notice (as I noticed I pressed the jump button two frames in a row according to input display).

[calm ether]

it actually does fix it, wow.

[raven wyvern]

Yay! 🙌

[calm ether]

the frame got added back in.

also, I gave up on the new editor I wanted to try. It wouldn't take .avi files and even after converting to .mov it would only take the input display due to codec. So I went back to old editor instead.

[raven wyvern]

Whatever works

😊

[calm ether]

It seems like it'll still be a bit off from time to time. Now on the last rfh ramp lap 3 input display repeats jump inputs twice. Whats interesting is that the buttom timer (the one in the picture) still changes, so one of these inputs must be off. But after this repeat the inputs sync with the run, and the inputs also synced with the run during the rfh brigde ramp (this is where I made sure to sync input display with the run in the first place). So what I think (not an expert at all, so I'm probably wrong here) is happening is that the input display recording internally in 60 fps tries to fix its accuracy to a 59,94 fps display (or something similar) which introduces some points like I noticed at the last rfh ramp. In total this means the input display will probably be 1 frame off at times.

[glossy dust]

This is a hacky way of doing it that doesn't always work @calm ether

The better way is to go to every 999th frame of the input display video and delete the duplicate frame

You'll know it's a duplicate because the timer that comes with the input display will freeze for 2 days

I actually have a version of input display that comes with a proper frame counter on the input display video (instead of a seconds and milliseconds one)

I can send you that if you'd like, to check for accuracy

Setting it to 99.9% speed might make it sync but it could actually get rid of the wrong frame (not the duplicate)

[calm ether]

I don't get this. setting it to 99.9% speed makes the input display video take longer while erasing duplicates make it shorter. and right now the input display takes too little time to get through it all.

shouldn't these things not fundamentally have the opposite effects (spitballing here)?

[glossy dust]

wait lol

yeah wtf?

why is your input display shorter than your video, I've never seen that before

try with the version I use and see if it works maybe

https://drive.google.com/file/d/1m_cWl_IvvCAXkGLwFZdzJkM2o5M5a6lz/view?usp=sharing

note: do NOT record in 59.94fps mode

choose 60fps

[calm ether]

sure, plz send (you already sent as I was writing this, and I'm too stubborn to not post message)

[glossy dust]

otherwise the program will write like 5 million PNGs to your hard drive until you run out of storage

(not kidding)

[calm ether]

tf

[glossy dust]

happened to me and it made 3 million images in a folder before I realised what happened and stopped it

Anyway yeah setting the video to 99.9% speed to make it longer is just going to create incorrect, duplicate frames

[calm ether]

I tried your version and it seems like I run into the same issue here

[glossy dust]

Are you able to work out what frame(s) of input display are getting skipped

[calm ether]

This one got skipped (before I put input display to .999 speed, shoud be equivalent to frame 762 since it starts counting frames from -4.000 sec and jumps 0.17 seconds).

[glossy dust]

if it happens every ~13 seconds or so then i guess you could just use some editing trickery to manually add that frame in, if you know what the inputs are

you'd only have to do it 8 times or so

copy and paste a frame from somewhere else in the run that matches

why is it always you that runs into recording issues man

[raven wyvern]

i think everyone gets these issues it's just rs has such a good attention to detail that he notices them

[glossy dust]

no

first time i've seen this

everyone always has the problem where input display is longer than the video

not shorter

[raven wyvern]

oh ok

[glossy dust]

also i'm mainly referring to all the issues he had in 2019 with framedumps lagging or being incompatible with his editor

or low quality

it's the reason why i was the recorder for a few runs

[calm ether]

the thing is, this frame does actually appear when I put it on .999 speed, and input display video lenght matches the length of the run when put into .999 speed.

[glossy dust]

not sure if you've decided on it already, but i would make the size of the input display much smaller

maybe by 25%

[calm ether]

hey, my frame dumps does actually not have any issues this time.

[glossy dust]

you never realise how big it truly looks until the video's published

[glossy dust]

this just makes 0 sense

[calm ether]

like this?

[raven wyvern]

i recommend trying to match this input display size/position, since youll need room for 10 names on the name display

https://www.youtube.com/watch?v=zsrLa71ADoA

[glossy dust]

what about the issues you were complaining about before

[glossy dust]

or disable minimap and put the names on the right side

i wouldn't match the position exactly from that video, the input display is too low

you don't need to have the position that low because there are less names

[calm ether]

It's because I'm still seeing duplicates in another part of the run.

But taking out all the duplicates would mean that input display is still too short (in terms of video length compared to my actual run).

[raven wyvern]

what i thought happened (which is why i suggested the speed change before) is that:

• he imported the 59.94 fps framedump
• importing the 60fps input display video into his editing software cut out frames to align with the 59.94 fps timeline
• setting the input display video to 0.999 speed made it 59.94 fps, aligning it with the framedump correctly
  • however, when this was done, the process wasn't perfect, and it gave some duplicated frames instead of restoring missing frames

[glossy dust]

what if you make the timeline 60fps

[raven wyvern]

that will probably cause there to be duplicate frames in the framedump lol

[glossy dust]

under that logic won't there be duplicate frames anyway

since youtube processes the upload as 60fps

[calm ether]

if you look at the images I gave, the duplicates are not in the video, but in the recording.

that's what I'm referring to

[raven wyvern]

i dont quite know how 59.94fps encoding works, since there should be a duplicate frame like once every 16 seconds if it's converted to 60fps

and ive never observed that

[calm ether]

all 3 images are different, but two share timing and other two share input.

[glossy dust]

out of the first two, pick the frame that fits into the hop sequence and discard the other

and then leave the .417 one in

[raven wyvern]

yeah ^ if youve managed to find the duplicate frames, you can actually just manually delete them

oh but you said that would make the input display too short, right?

[calm ether]

perhaps putting the video to 0.998001 speed and then erase duplicate frames is the solution

[tough citrus]

maybe convert 60fps input display -> image sequence -> 59.94fps input display

[calm ether]

I noticed that the frame count of the particular duplicate I have does not align with any multiple of 999 in the run. (just checked on dolphin).

also this is the last frame of input display recorded. The timer (95.300 --> 1:35.300) does not align with the end time of my run either. However the inputs recorded are more in sync with the end of the run after I put input display on .999 speed.

[glossy dust]

What a mess

I doubt there's a fix for you at this stage

What video editor are you using

[calm ether]

shotcut

this case is pretty odd

going frame by frame, notice how it created a duplicate here

and then subsequently skips a frame here

these occurances happen like .2 after one another.

and you may notice that the input display shows something different in each of these images, so this is something introduced in the input display video somewhere (unless the skipped frame is there and I can't access it in my editor).

You know what, I'm just going to take the input display as is, and then note there might be some slight inaccuracies in the video discription. It should either be completely accurate or 1 frame off for the entirety of the run (will be my conjecture).

[tough citrus]

You could also modify PyRKG to output in 60fps (CONFIG.py, line 1)

Wait nvm I misread sorry

PyRKG outputs 59.94 by default

[calm ether]

https://github.com/Bsquo/PyRKG
Is this the PyRKG you were talking about?

[abstract orbit]

interesting

not sure what editor you're using but yes I generally change input display speed from 60 to 59.94 (x0.999) and then delete duplicate frames

[calm ether]

Ok, I've noticed that input display was originally recorded in 62.5 fps (according to my editor).

[tough citrus]

I was talking about this one: https://github.com/AtishaRibeiro/PyRKG/ but Bsquo's fork is fine

Both have the framerate set on line 1 of CONFIG.py

[raven wyvern]

ftr when most people use input display they dont use pyrkg

so unless this can be easily translated to how it works in the gui app, i dont think this is going to help

[tough citrus]

Ohhh makes sense

[raven wyvern]

https://github.com/AtishaRibeiro/InputDisplay/releases/tag/1.2 we use this

[tough citrus]

WHAT
I didn't even know this existed

[raven wyvern]

https://www.tt-rec.com/inputdisplay/ people have gotta learn to click the other tabs of tt-rec hahahaha

[tough citrus]

💀💀

[calm ether]

ayy, it even comes with a free ffmpeg

[calm ether]

ok, got the input viewer recording to work. Just had to slow it down to somewhere like 0.95 speed and then remove a frame every .4 second.

I only fully checked that it matched on every rfh, because that was the easiest to tell.

[rough needle]

from my experience with the input display; every 999th frame it duplicates a frame, and every ~499 frames later (roughly, it's inconsistent) it skips a frame. or maybe the opposite. either way setting the video speed to 0.999x definitely does not fix the issue, for me it made it even worse.

but i can't explain all the other weird inconsistencies you (or ejay) have been getting, for me that's all the program gets wrong, it still finishes at the same time as the run

[calm ether]

noticed a little error in masking away the timer on input display
https://www.youtube.com/watch?v=S0O37CejiOQ

also music test

[raven wyvern]

ghost comparison! wooooooooo

[calm ether]

I still need to do name display and add 'km/h' text under my vehicle speed. Would be good to know how you generate these texts.

[raven wyvern]

First of all, you want this font

Download and install it

Then you can use that in your editing software

[glossy dust]

I don't get the 499th frame thing that you do

[raven wyvern]

The way name display works is just by having a layer of text for each name, and applying 33% opacity whenever there's no contribution

[glossy dust]

I love how I told someone I arbitrarily used 33% opacity once in like 2019 and now it's the exact percentage everyone uses lol

[calm ether]

And what about the 'km/h' text (the same font as 'TIME')?

[raven wyvern]

i know you told me that

[rough needle]

it duplicates and removes random frames outright. it alternates between these two every 500 frames roughly. so it stays roughly in sync with the run the entire time

[raven wyvern]

theres a mkw font generator site nokky made

[rough needle]

at least for me but idk. i wish there was a good input display that normal peasants can use (so not pyrkg)

[raven wyvern]

lemme find it

[calm ether]

Also, I'm taking a break on editing further for now, I'd just like to know ahead of time.

[raven wyvern]

https://nokky-mkwii-text-generator.streamlit.app/ you can make any text with this

ejay probably has a hq "KM/H" image though

[glossy dust]

I do

I will send it when I get home

In 1 hour

Or maybe I can find it in the server

[raven wyvern]

#off-topic message

theres a bunch of moderately hq images here

[glossy dust]

No

Low quality

That's very pixelated

[raven wyvern]

yeah i was just noting that link in case

i found it

[glossy dust]

LOL

#tas-focused message

Here

[raven wyvern]

corvid do you know what that image is from

[rough needle]

no

[raven wyvern]

oh youre gonna have to give me another fire emote

https://www.youtube.com/watch?v=df2e9AKxsHA

[rough needle]

i will need to start using hardstyle for my future runs. it seems like everybody in the community loves it!

[raven wyvern]

https://www.youtube.com/@N3vermind/videos just watch every video in this channel and you'll change your mind!!

[calm ether]

I'll figure out the name displays later.

this is the general layout I think I'll be going with (name displays on the left)

[abstract orbit]

motion track the names to daisy's front wheels

jk I think that looks good, you could maybe move input display slightly down and slightly right if you wanted to align it with the other hud elements

[calm ether]

I'm lowkey considering keeping input display this high due to YT pause screen.

[covert elbow]

@calm ether

i have the km/h symbol

if you want to

https://cdn.discordapp.com/attachments/710835440156147743/937933496721240064/NICE_KMH.png?ex=67d11323&is=67cfc1a3&hm=baf316aacccd3276945ae5826233800a93b0286f8a9c2eb470b09264ec79923f&

[glossy dust]

We already sent it lol

[covert elbow]

o

[calm ether]

Finally found a way to get the font to work properly with good shadows.

[calm ether]

Don't mind me, I'm just compiling all info for name display

[calm ether]

.

[raven wyvern]

Just the F section of my first message should be what you need

[calm ether]

Is that up to date?

[raven wyvern]

Yes

It’s not precise though

As in, it doesn’t say precisely where the start and end of the listed contributions should be

[glossy dust]

why doesn't Ellu have input

[raven wyvern]

the flap was never referenced for the zipper

[glossy dust]

I thought the double trick was

[raven wyvern]

ellucinea’s inputs ended before the double trick

[calm ether]

I could start by getting everyone's name on display and check layout.

[raven wyvern]

First question

How do you want to organise names

Alphabetically, or by contribution?

[calm ether]

probably alphabetically, that's easier to determine

Also it'll take a couple days before upload is ready

[raven wyvern]

I mean if you want it by contribution we can figure out what works best

In fact I have an idea for visualising it

[rough needle]

i love you so much

[calm ether]

Finally added all the names (have to change input display position a bit). I took the names directly from https://mkwtas.com/

how bout this

[covert elbow]

i like alego's position for it

and it's size

[glossy dust]

Just put me as Ejay

[raven wyvern]

i would make the names a bit smaller

looks good though

[glossy dust]

This is why right side name display is just better

Who needs the minimap for no glitch categories anyway

It looks way too cramped on the left

[smoky talon]

Microscopic font

[calm ether]

I gave a little more room to the names

[raven wyvern]

ghost comparison means something because of the minimap

[raven wyvern]

also here's what i was talking about

blue = syncing
green = direct/reference
yellow = agc

wait theres a mistake

[graceful torrent]

Lol 10 author tas

[calm ether]

While that's cool, I can't color code the name display. That's going to be a bit too confusing to the viewer for my liking. Maybe I could link to it in the discription. (I'm talking about the sheet below)

[graceful torrent]

Second most only to sgb I think?

[raven wyvern]

this isn't for name display itself

this was just to help figure out who has most contributions

[crude prism]

Maybe only use Alego, since Alego never adds the "11" #contributing

[rough needle]

i like alphabetical

[raven wyvern]

if youre wondering, the row heights actually correspond to IGT

[graceful torrent]

What do the colors correspond to

[raven wyvern]

.

[graceful torrent]

Oh

[raven wyvern]

i would probably say:
rs, alego, ejay, marth, ancient, corvid, ego, wade, lirwm, me

(this is still important to know because the bkt sheet lists authors by contributions)

[rough needle]

someone will need to sit down with me someday and explain to me the point of the contributions dick measuring contest

i guess i've spammed this thread with dogshit enough but i will never get it

[glossy dust]

I'm not a fan either

But nobody likes my pyramid idea anyway

[glossy dust]

in what world did ancient contribute to this tas more than Corvid did

[rough needle]

i think everybody's contributions are valuable and it's just toxic to decide who did more arbitrarily

[rough needle]

honestly i did next to nothing, i almost don't want the credit for this tas

[raven wyvern]

• corvid saved 3ms on the last ramp of lap 1, which got reused lap 2
• ancient saved probably more than that on the first ramp of lap 2, which got reused on lap 3

[glossy dust]

So it's based on timesave?

I thought it was the length of track someone helped with

Based on your chart thing

[raven wyvern]

well you have to keep in mind most of corvid's contribution is like 5 seconds of wheelie

ancient's contribution was during the shroom

[glossy dust]

See what I mean, it's impossible to quantify

So why bother in the first place

[raven wyvern]

because i wanted to

[calm ether]

I guess it's so that when people read the name list on mkwtas.com they see the biggest contributors first and the ones who helped with one thing or two at the bottom, and you're more likely to read and remember the top names first. Otherwise I wouldn't see any point to it either, and it doesn't matter at all for name display (we already fade out, where people don't contribute)

[glossy dust]

contribution worth is entirely subjective

[raven wyvern]

[raven wyvern]

name list on mkwtas.com isn't ordered by contributions

[rough needle]

see but the names list on mkwtas is stupid

bc it goes by order of whose account was made first

[raven wyvern]

it's ordered by who got a bkt on the site first lol

[rough needle]

i will always be at the bottom (or near the bottom)

[calm ether]

welp whoops I guess

[acoustic atlas]

this, actually

[calm ether]

then I have no clue

[raven wyvern]

i was talking about https://docs.google.com/spreadsheets/d/1-214dOJVTJyymS3eGcowJAMefI5y0QoS8R_L8ech6CI/edit#gid=451594463

[raven wyvern]

yeah i agree

[acoustic atlas]

number of contributions? value of those? trying to compare stuff? so relative and subjective

[raven wyvern]

but imo it's the most relevant order to list names in

where people's names are in the alphabet has nothing to do with the tas

but thats just my opinion

🤷‍♂️

[rough needle]

so you're saying the most relevant order for a list is your opinion?

[acoustic atlas]

not to be a bitch but prob not the right place

[glossy dust]

We have two methods to order name display

1. functional purpose
2. aesthetic purpose

Since the 1 functional purpose (contribution worth) is impossible to judge, then that leaves 2 as a default

The most aesthetic way to order the display is either by lettering convention (alphabetical), or by length (shortest name to longest name)

[calm ether]

not sure if this run should be categorised as cyan or blue tbh. Then again it's hard to put the same standards on every category

ok, compared to some other categorised runs I'd say blue

[abstract orbit]

real

[graceful torrent]

What if the names just appeared in random spots on the screen every time they lit up

Problem solved

[calm ether]

what if all the names were put on the same spot.

[hushed zodiac]

Loving the Demetori music

[dense dawn]

What if a name only showed up if they had inputs on the sections

[glossy dust]

annoying for people who want to see the whole author list at once

And if they skip a tiny section of the run they may not know a particular author worked on a section

[calm ether]

we do have a description for that

also, I guess this is how the names will look when faded

[calm ether]

ok, now I need to figure out good start/end points for each contribution. I'm not sure how precise I should be in this.

[calm ether]

guess I'll be following this

Guess the bridge trick starts here?

[raven wyvern]

One sec lemme explain each row

I combined some parts into one label

[calm ether]

👀

[raven wyvern]

note that the labels represent the start of each section

• slide - fade-in
• start - beginning of lap
• first ramp (hop/drift) — this isn’t on there, but is actually where corvid’s inputs end, not the trick
• first trick (animation)
• shroom
• shroom end - landing from hop out of offroad
• bridge trick (animation)
• bridge trick (landing)
• turn before zipper (mt release from shroomless sc)
• turn before zipper (drift)
• zipper (hop after stick)
• double trick (zipper hit)
• double trick (landing)
• bumps - first trick (animation)
• second trick (animation)
• wraparound (hop)
• wraparound (landing)
• last turn (hop)
• last ramp (rfh)

@calm ether if you’re unsure about any of these then ask and I can give you exact IGT/screenshot