- F-ZERO GX: Tool-Assisted Speedruns -

TAS = (Tool-Assisted Speedrun / Superplay)


TASes are created with an emulator that supports specific tools, such as frame-by-frame speed, save states, and other additions that allows you to optimize the game as much as possible. Obviously, TASes aren't meant to be compared to real-time records, it's rather a demonstration of what the game looks like when it's pushed close to its technical limits.


TAS Techniques:


Broken Steering (BS)

This technique allows the steering radius to break in airborne state. As far as real-time goes, this has only been an instant flight killer whenever it occurs on accident, since it's practically impossible to control, or even trigger on demand. However, in a TAS, with the option of retrying segments infinitely, this can indeed be taken great advantage of. This allows us to cut corners further than before, and at the same time, remain in faster airborne state for longer.

What triggers this glitch ultimately is the horizontal axis of the control stick, with the requirement of having a high input value towards either left or right. Other conditions that needs to be met are still uncertain, but I can confirm that L&R, as well as the vertical control stick axis, has no direct influence on triggering the glitch.

Shaking - [Discovered by Mugg1991]

Shaking is essentially frame-perfect Snaking; alternating L+Left and R+Right every other frame, without timing error. This allows the machine to exceed beyond the speed limitation of traditional Snaking strides, as well as having quicker acceleration. There's also an additional side-effect that occurs from utilizing Shaking, which I prefer to call "Momentum Storage".

Momentum Storage (MS)  [Updated: 03.2020]

This is a technique which occasionally happens in real-time records, but there is a much more saturated effect that occurs from cancelling an extended Shaking line, often seen in TASes. Momentum Storage is additional speed that is temporarily stored by either Quick Turn drifting over an extended time, or by going into an unusual Quick Turn drift state, like Shaking. The speed stored will be unleashed as soon as the drift is interrupted, usually done by a change in movement, such as going into a neutral state, MTS, ground-to-air change and similar. An exception to this is to partially cancel the MS by going into boost state, which can be done without cancelling the drift state.

Momentum Throttle Air Shaking (MTAS) - [Discovered by CGN]

MT Air Shaking is a slightly adjusted version of Shaking on the ground. By performing Shaking in the air with the accelerator off, the decelrate is reduced to a minimal linear rate. This can especially be abused at high speeds, because the reduction in speed is actually less significant, due to the linear decelrate. However, this comes with a few limitations. Since we have to maintain Shaking state for this to work, turning ability is greatly reduced. As soon as a big change in cornering is required by the track layout, the tech must be interrupted, and the speed will also follow by going back to standard MT decelrate. Because of this, this tech is mostly useful on long straight paths where the track geometry does not increase in height.

Space Shaking - [Discovered by CGN]

If Shaking is done with the accelerator on while in the air, it can potentially gain infinite height, while also pushing the speed to its limit. Just like MT Air Shaking, the turning ability is mostly locked into a straight line and will therefore require trading a significant amount of speed in order to turn more sharply.

This is arguably considered as "Space Flying" tech, due to the possibility of infinite height gain.

Video: https://www.youtube.com/watch?v=2khDL-7FDVs

Collision Cancelled Deceleration (CCD) - [Discovered by CGN]

Collision Cancelled Deceleration or CCD for short, is when the speed decrease that's supposed to happen upon a collision is cancelled. This can happen by hitting a rail (or anything that's considered a collision) within a certain range of angles. However, for this to occur, it's also important that the accelerator is restarted on the frame after the collision has occurred. At really high speed, it's even possible to take some advantage of this upon colliding with just the track surface, without lifting the Y-axis of the machine prior to the impact. The main purpose of doing this tech is that we can avoid sacrificing speed and time caused by high-decel actions, such as lifting the Y-axis of the machine prior to where we plan to make contact.

The tech itself is not new, as it was first discovered and demonstrated in my former MCTR Snaking WR of 30"661 (see the MTS which is intentionally cancelled with the wall hit). Back then I did not call this tech a specific name, but I have now decided on a better name for it, since it's better understood. In addition to that, it was not realized it could be done while in the air, as mentioned above.

Airborne Shift Boost (ASB) - [Discovered by CGN]   [Updated: 03.2020]

Shift Boosting is something that has been thought to only be possible when the machine is in ground state. However, as recently discovered, it's possible to utilize a Shift Boost in airborne state, given that the correct conditions are met. Applying this technique allows us to extend, and even utilize more dives than before, as well as expanding the faster airborne travel.

The way it works from what I've noticed so far:

  1. The railing expands its gravity in a plane (how far they expand is still uncertain, but it's at least very, very far). To trigger a Shift Boost in the air, the machine must merge into this invisible plane.
  2. The machine cannot be aligned far away from a parallel alignment with the invisible gravity plane.
  3. In order to work, the machine must ascend, or be close to have height gain upon merging with the invisible gravity plane.

NOTE: I have yet to confirm if this works with edges that have no railings, but my theory so far is that it doesn't.

Another thing that's worth to mention is that the gravity plane only affect the machine during frames that meet "most" of the above requirements. I say most, because you can get influence from the gravity plane without getting a Shift Boost. It's just that if you're misaligned too much, there won't be any influence on the machine at all. If there's influence introduced as merging, it means the conditions are close to resulting in a Shift Boost.

UPDATE: After enough testing made while working on the Port Town Aero Dive TAS, it appears that ASB cannot be done on this track. This discovery suggests that only certain types of rails or track edges allow ASB to occur. So far we only have Mute City Sonic Oval confirmed to be working, but it's very possible that the Mute City type of railing in particular allows ASB to work. Further testing will still be needed before anything can be said with more certainty, whether ASB can be done or not on a said course. Details regarding this will eventually be updated again, once enough TASes has been created.



TAS Topic at F-ZERO Central:

http://www.fzerocentral.org/viewtopic.php?t=9884


CGN's TAS Playlist on YouTube:

https://www.youtube.com/playlist?list=PL9B347D67216AA868




CGN's TASes


NOTE:
The Dolphin emulator as of revision 4.0-2729 and beyond, emulates the physics of the game with 100% accuracy. Revisions before that emulates the physics with an accuracy of ~99%, meaning that it's not possible (in theory) to replicate the exact same TAS on console.


Definition of category restrictions:
http://crazygamenerd.web.fc2.com/FZGX_TA_Rules.html


TIME ATTACK


COURSE

CATEGORY

TIME

LAPS

MAX SPEED

MACHINE

DATE

VERSION

EMULATOR

REPLAY

Shaking

0'21"363

09"745 + 06"260 + 05"358

15367 km/h

FAT SHARK

27.02.2020

NTSC-U

5.0-11069

Work: 150 hours
Re-records: 62000
LAP 1: 78H, 33000 re-records
LAP 2: 37H, 17000 re-records
LAP 3: 34H, 12000 re-records

Average Speed:
6909 km/h (Total)
8196 km/h (Lap 3)

*Compared to "intended" track layout:
 8074 km/h (Total) (47913.5m / 21"363)
10727 km/h (Lap 3) (15964.8m / 05"358)

Tech Info:
4K 60FPS (3:2 Hack)
Video & Audio dumped with Dolphin

Console Verified:

COURSE

CATEGORY

TIME

LAPS

MAX SPEED

MACHINE

DATE

VERSION

EMULATOR

REPLAY

Max Speed

0'38"630

16"368 + 11"616 + 10"646

16830 km/h

Gallant Star-G4

18.07.2012

PAL

3.0-306

This was my first F-ZERO GX TAS that made use of the "TAS Input", allowing me to adjust analog values with absolute precision. Ultimately, this motivated me to work very hard on optimizing.

Unfortunately, there's an in-game desync that appeared near the end of lap 2. Might have something to do with the fact that this revision emulates the physics of the game at ~99% accuracy.

Work: 80 hours~
Re-records: 32500~

Max Speed

0'43"745

21"970 + 11"367 + 10"408

15783 km/h

FAT SHARK

30.07.2011

NTSC-U

2.0-7376

This was the first showcase of a Max Speed TAS. It's quite obsolete compared to what I can create now, but still a good watch for its time, I think?

No "TAS Input" available when I created this, so cornering sometimes lack precision.

Work: 50 hours~
Re-records: 20000~

COURSE

CATEGORY

TIME

LAPS

MAX SPEED

MACHINE

DATE

VERSION

EMULATOR

REPLAY

Shaking

0'11"245

05"851 + 02"770 + 02"624

15713 km/h

FAT SHARK

14.05.2016

NTSC-U

4.0-4480

The first TAS to showcase Broken Steering and Airborne Shift Boost. I don't think this game can be pushed much further than this...

Work: 115 hours~
Re-records: 60000~
*Very re-record heavy, due to lots of BS experimenting.

♪BGM: Super Monkey Ball
            World 8 - Clock Tower

Video: Dumped with Dolphin
*Res@FPS: RGB 852x480p@59.94FPS
Audio: Recorded from Nintendo Wii
*(running in GameCube mode)

Console Verified:


STORY MODE


CHAPTER

CATEGORY

TIME

LAPS

MAX SPEED

MACHINE

DATE

VERSION

EMULATOR

REPLAY

Shaking

0'22"877

10"635 + 06"863 + 05"379

  9846 km/h

BLUE FALCON

21.05.2011

NTSC-U

2.0-7376

I had to make two TASes of this, because I wrote down a mistake in the "input sheet" that I played back the TAS with. Saved time over revision 1, though:
23"482 (10''715 + 07''134 + 05''633)

I will make a much more optimized version on Very Hard at some point.

Work: 40 hours~
Re-records: 15000~

Shaking

0'15"854

12739 km/h

BLUE FALCON

29.03.2011

NTSC-U

2.0-6237

Super ancient revision of Dolphin, which didn't even support audio dumping.

This one has aged a bit, but at the time it was the first showcase of how crazy Shift Boosting is with perfect play.

Average Speed: 4434 km/h

Work: 25 hours~
Re-records: 10000~

Shaking

0'11"790

12413 km/h

BLUE FALCON

02.01.2012

NTSC-U

3.0-306

The .DTM file lost an input in the finished file, but I was lucky enough to locate it in a Hex-editor and repair it.

Work: 25 hours~
Re-records: 10000~

Max Speed

0'35"555

  2093 km/h

BLUE FALCON

17.08.2012

NTSC-U

3.0-306

This is sort-of a theoretical display of how far this chapter goes under Max Speed settings, without Shift Boosting.

I do believe Shift Boosting can be triggered here with enough patience.

*Peak speed of only using the booster and MT Boosting was 1615 km/h.

Work: 20 hours~
Re-records: 8000~

Obsolete TASes and other junk that I have made:
http://crazygamenerd.web.fc2.com/FZGX_GXTASJunk.html




-> BACK TO TOP OF THE PAGE <-




inserted by FC2 system