Ambassador, Experienced Forum User, Moderator, Published Author, Site Developer, Player
(155)
Joined: 3/17/2018
Posts: 358
Location: Holland, MI
This was my main childhood mario kart game, I remember using Yoshi and Toad a lot and specifically trying to beat the times on Sky Garden with the intended shortcuts, so it's awesome to see Yoshi used for that track and to see such a major skip available!
Ambassador, Experienced Forum User, Moderator, Published Author, Site Developer, Player
(155)
Joined: 3/17/2018
Posts: 358
Location: Holland, MI
Welcome PoltArtGeist! I don't know squat about HGSS but I can tell how much saving 9 minutes over RTA means in a Pokemon speedrun, especially one with this kind of RNG mechanics. Can't wait to watch it in full.
Ambassador, Experienced Forum User, Moderator, Published Author, Site Developer, Player
(155)
Joined: 3/17/2018
Posts: 358
Location: Holland, MI
MUGG wrote:
The latest Kirby's Dream Land submission runs the game in GBC mode.
It was found that doing so and using a specific color palette is faster due to less lag.
This method itself wasn't even fully explored as the runner stated he didn't want to play through the game 16 times, each time using a different palette.
As a side note, some palettes are not so nice-looking...
While it seems like a legit thing to do, will this be something that is of significance when deciding if one run obsoletes another?
Shouldn't plain GB games be played only in GB mode?
This is a very good question and it has several nuances as applied to the Game Boy:
Typically speedrunners try to use the fastest iteration of original hardware available, or in the case of TAS rather to emulate the fastest iteration - an example is with XBox disc drive or hard drive variants affecting game load times. Because of how the Game Boy's backwards compatibility was designed, "original hardware" exists for GB games all the way through the GBC and GBA platforms, since the GBC cpu was designed to be back-compatible with GB games and the GBA includes an entire GBC cpu. There's an additional complexity of many games being designed for the -SGB- even though the SGB ends up in many ways a problematic platform. Playing some of the SGB enhanced GB games on GBC causes delays from palette swap behavior, but at the same time timesaves can be gained from the palette behavior as using a non-default palette can reduce lag and also create more possible RNG variance, which itself can lead to other timesaves. Some of this can be seen in the last few updates to the Red Save Glitch movie (known as Any%/Save Corruption in RTA) where timesaves were being discovered on the order of a handful of frames and clock cycles.
At the same time, Nintendo with the Game Boy Player accidentally created one of the most convenient verification setups of any console system, making it highly preferable to have TASes done in GBA mode for ease of verification.
The previous solution to these complexities has been to allow whatever the author prefers for their submitted movie, noting the high potential for verification if written for GBA mode but also noting that verification is not a requirement for submission. I'll also note that since many RTA communities standardize on the Game Boy Player for submission, TASing towards that platform also often aids in RTA comparison. I think the author choice continues to be a good solution as otherwise you just have a mess of options of different ways to standardize.
Ambassador, Experienced Forum User, Moderator, Published Author, Site Developer, Player
(155)
Joined: 3/17/2018
Posts: 358
Location: Holland, MI
As Alyosha hinted at, his and Mugg's TASes of this game have been a frequent part of live console verification demos and I'm super excited to see such a major improvement. I can't wait to show people all the new improvements.
Ambassador, Experienced Forum User, Moderator, Published Author, Site Developer, Player
(155)
Joined: 3/17/2018
Posts: 358
Location: Holland, MI
TASVideos x TASBot were guests at Long Island Retro 2024, showing TAS console verification on the Game Boy Player + GBI platform live on the show floor all weekend, and holding a panel about new approaches to game preservation and TASing with my web emulator project T3Boy and my Pokemon Yellow ROMHack "Fort Knox Edition". Our first edited content from the event is our interview with Justin Scerbo of Eon Gaming. Eon makes one of the HDMI adapter products that we use to capture Game Boy Player footage from the Gamecube with high quality. I've manually corrected the closed captions for the video as the mics we used clipped out a lot and it's been requested before in comments. We'll have two other interviews to edit, one with Frank Cifaldi of the Video Game History Foundation and also one with Try from My Life in Gaming, and then we'll have the panel talk to upload as well.
Link to video
Ambassador, Experienced Forum User, Moderator, Published Author, Site Developer, Player
(155)
Joined: 3/17/2018
Posts: 358
Location: Holland, MI
I had the privilege to discuss this with ikuyo prior to posting and in person with DrD2k9 due to the timing of our Long Island Retro expo showing. I have a few thoughts:
I believe that curation is too core to the site to completely abolish. Some level of curation is necessary to have a place for things to be confirmed as syncing and confirmed to beat known records. dwangoAC's recent DEF CON talk is a great example of the importance of this curation layer.
From talking with DrD2K9, this should already be similar to the official process, but my opinion would be that judging should be minimized to replicating sync and either confirming with a game community that an improvement appears legitimate or if no community exists just to do a visual check that the TAS appears like a good faith attempt to set a baseline record. In this way I would love larger game communities to be handed more direct judging capability over their own games, and to have our current judges act more as the backup pool for games without baselines and final say on resolving disputes. This would allow for example the street fighter community to curate TASes of tens of different individual characters without burdening the judge system.
If that were to streamline the judging system, it leaves publication as a bottleneck. For that I would suggest having an alternate publication track channel "tasvideoschannel2" where the submitter or other community member provides the encode and we don't dig into the minutiae of the video quality. In other words the main channel could host for example the fastest character "any%" but the authors or community could supply encodes for channel 2 for as many other characters as desired or for those baseline uncontested records.
Ambassador, Experienced Forum User, Moderator, Published Author, Site Developer, Player
(155)
Joined: 3/17/2018
Posts: 358
Location: Holland, MI
It doesn't do as much generally for the N64 but rcombs' 15 hour Watch for Rolling Rocks .5 A Press verification is the longest I know of which is an achievement itself for the stability of the tooling.
Link to video
On GB I would also highlight the DK94 verification as an important one for the validation of procedures to stabilize uninitialized RAM.
I would also argue the whole PC TASing scene is quite interesting as to how it questions the very idea of what a verification is - usually it applies only to a single SKU of a system, eg the GBA/GBP variant of the Game Boy Color, but my Backyard Baseball TAS for example is replicable across systems of wildly different hardware capabilities. There's not really any interesting accuracy insight that could be added by eg sending the inputs over mock mouse/keyboard over usb instead of through libTAS since you're then validating lag behavior of the OS layer much more than the hardware. So it kinda shows how our concept of verification is tightly tied to these earlier bare metal systems without an OS layer or ones where the OS is very meager and without any/many revisions.
Ambassador, Experienced Forum User, Moderator, Published Author, Site Developer, Player
(155)
Joined: 3/17/2018
Posts: 358
Location: Holland, MI
Hi! If you want you should jump in the Discord linked in the "Chat" section of the site here: https://tasvideos.org/LiveChat
There's some specifics to GB stuff especially one's like Kirby's Dreamland that we've already verified that I can get into more detail on in a chat environment.
Ambassador, Experienced Forum User, Moderator, Published Author, Site Developer, Player
(155)
Joined: 3/17/2018
Posts: 358
Location: Holland, MI
I've hit two significant technical milestones with T3Boy!
First, I implemented touch controls that work without having to raise and repress fingers and also work with multiple touch points, which enabled me to record this GBA/GBP accurate RNG manip for a perfect NidoranM live on stream:
https://youtube.com/shorts/pqlOGZ4MVvw?feature=share
Second, I implemented working rudimentary GBI input log playback. This is huge for me because part of my design goal of a previous project, rgbscaler.com, was to minimize the data transfer necessary to watch TASes while bringing perfect video quality via lossless encoding. This new system actually playing the real TAS in-browser is an entirely different level of insane data transfer savings and quality, with users having the option for example of whether to emulate the GBC lcd palette or not, and then being able to continue playing themselves from any point in the TAS. To host TAS movies for GB/GBC would now require only hosting this site and the GBI input logs, which is awesome to have a backup from YouTube even if not necessary at this moment.
Ambassador, Experienced Forum User, Moderator, Published Author, Site Developer, Player
(155)
Joined: 3/17/2018
Posts: 358
Location: Holland, MI
I agree with the playground distinction for this run. It's a unique technical accomplishment, but the goal choice is not ideal for entertainment value and doesn't match well with existing precedent. A "maximum dex in this game" TAS aka "Emerald 213" would match a longstanding RTA category and have a good entertainment value, personally for me closer to 7/10. Likewise with Coop Diploma we have existing gen 1 and gen 2 TASes and I could see that as a 7/10. Living dex I would drop to 5/10 because of the repeat catches needed and Professor Oak's Challenge I would drop to a 3 or 4 because of the very repetitive grinding. Doing both together is a tad more interesting at a 4 or 5 for me.
Ambassador, Experienced Forum User, Moderator, Published Author, Site Developer, Player
(155)
Joined: 3/17/2018
Posts: 358
Location: Holland, MI
https://t3boy.vercel.app/https://github.com/TiKevin83/T3Boy
T3Boy is an experimental webassembly based Game Boy emulator based on our gambatte speedrun core in BizHawk, thanks to Stringflow's help making WASM builds and prototyping the render loop. It's open source and uses the T3 Stack of React/Next.js/Typescript/Tailwind/tRPC for client UI and API connectivity. The tRPC API enables, when signed in via discord auth integration, cloud backed game saves and savestates per-game. Other current features include remappable keyboard and controller input support along with enough accuracy that I can perform the standard GBA/GBP Yellow nido manip on it.
The way the render loop works may be unstable for tracking the speed at which the emulator runs over long time frames and has terrible frame pacing so I don't intend for it to be usable for submissions but as is it works well enough to get people practicing and learning without being able to download or install an emu. The long term goal is eventually to have this project be also able to replay GBI movies in-browser to provide a much higher quality and lower bandwidth solution to watching GB TASes which simultaneously can then be played through after watching.
Ambassador, Experienced Forum User, Moderator, Published Author, Site Developer, Player
(155)
Joined: 3/17/2018
Posts: 358
Location: Holland, MI
Bigbass wrote:
I think it would be useful to come up with some written guidelines for determining when a warning is beneficial. If anyone knows of some reputable sources for learning more about photosensitivity and epilepsy, particularly in regard to videos, possible triggers and risks involved, that would be quite useful.
If anyone here actually has epilepsy or a related disorder, I'd be very curious in hearing their opinions on this subject. Such as what would be most helpful for them when viewing a movie, and what sorts of situations we should look out for. The most common trigger I've heard of is rapidly flashing lights, but it seems there is more to it than that. Does the whole screen have to flash, or just a small part of it? Does the flashing have to be sustained, or would a single flash warrant a warning? Is a notice in the video description or publication text sufficient, or would placing a momentary message at the start of the video be best?
I would look generally for the WCAG criteria - flashing of more than a 10 degree viewing angle or roughly 340x256 pixels at normal viewing distance, where flashing is a change of more than 10% of maximum luminance, at a rate between 3 times and 50 times per second. The upshot out of that is it has to be a big chunk of the screen, look like flickering and not changing images or a blur, and it has to be a pretty subsantial change in brightness with each flash. They also clarify that the bright pulse needs to be close to white or a deep red usually to cause the issue.
https://www.w3.org/TR/UNDERSTANDING-WCAG20/seizure-does-not-violate.html
Ambassador, Experienced Forum User, Moderator, Published Author, Site Developer, Player
(155)
Joined: 3/17/2018
Posts: 358
Location: Holland, MI
Alyosha wrote:
Dimon12321 wrote:
Thank you very much for your effort for GBA emulation and verifications! A sceptic question, sorry. Did anyone else verified a movie which has been previously verified by you? I'm just worried that you may be developing GBAHawk with only your improvised verificator and ROM re-writer in mind.
I don't think so, but all my files are available here: https://github.com/alyosha-tas/GBA_replay_files and I use original carts except for homebrew.
So, anyone can check, but that kind of thing is kind of unavoidable with things this niche.
@Alyosha I experienced a repeated desync in DK King of Swing, do I need to flush the saveram or do some other kind of ram stabilization before running that TAS? I was clearing the saves with the in game menu
Ambassador, Experienced Forum User, Moderator, Published Author, Site Developer, Player
(155)
Joined: 3/17/2018
Posts: 358
Location: Holland, MI
Dimon12321 wrote:
Thank you very much for your effort for GBA emulation and verifications! A sceptic question, sorry. Did anyone else verified a movie which has been previously verified by you? I'm just worried that you may be developing GBAHawk with only your improvised verificator and ROM re-writer in mind.
I've previously cross-verified many of Alyosha's GBC verifications, if we want to double check some of the GBA ones I would just need the GBI style input log and the exact cart to order.
Ambassador, Experienced Forum User, Moderator, Published Author, Site Developer, Player
(155)
Joined: 3/17/2018
Posts: 358
Location: Holland, MI
MUGG wrote:
MUGG wrote:
I found a difference between J1.0 and J1.1.
I figured out the whole story now. You can read about it here: https://tcrf.net/Kirby%27s_Dream_Land#Revisional_Differences
J1.1 is basicly a bugfix for a game-breaking bug that could happen in the Whispy Woods fight. That being said, I don't think this is any useful for speedrunning, but still interesting nonetheless.
Cool!
In the oddities section on that page, I see you uploaded the images for the normal and extra game end screens, but I don't see a difference, did you mean to upload a different image for one of them?
Ambassador, Experienced Forum User, Moderator, Published Author, Site Developer, Player
(155)
Joined: 3/17/2018
Posts: 358
Location: Holland, MI
I was aware to look at it but from all I can tell there's a good chance it's not any faster than the current any% run, the TAS is already one hit KOing most if not all mons in the run after the point where you'd be doing the dupe glitch because guts + quick claw is too powerful, so the only time save you get back is skipping seeing the level up and stat shows if you go all the way up to 100. It could be interesting though for building a theoryTAS of the human route (making decisions and using only the tricks as a human would in a run but with best possible luck in response). It could be applicable to any% though if you could
- maybe skip ever getting quick claw? but it might still be faster to get just for the few fights until you get to using the glitch
- skip using pickup for free rare candies which also takes RNG delays
- in turn maybe fewer catches if you can teach all the needed HMs without zig and/or taillow? But you probably still need a fly friend?
- save time by needing less rng manipulation in fights for para + quick claw activation
Ambassador, Experienced Forum User, Moderator, Published Author, Site Developer, Player
(155)
Joined: 3/17/2018
Posts: 358
Location: Holland, MI
Lobsterzelda wrote:
TiKevin83 wrote:
This is so cool! I had worked on TASing Enters the Race but didn't have an obvious path toward manipulating the RNG needed at the start of the race - if you crash into someone immediately at the race start you can instantly accelerate to the maximum speed, so you need RNG to put another car close enough to you for that to happen faster than the RTA record.
Interesting. Is there a video of this somewhere?
I didn't make any video of the TAS routing, but the video of the RTA strat is here at around the 6-7 second mark. You can see that the green car spawns directly on top of putt-putt, enabling the runner to "crash" into him to instantly gain top speed.
https://www.speedrun.com/puttrace/runs/y926dq2z
Ambassador, Experienced Forum User, Moderator, Published Author, Site Developer, Player
(155)
Joined: 3/17/2018
Posts: 358
Location: Holland, MI
This is so cool! I had worked on TASing Enters the Race but didn't have an obvious path toward manipulating the RNG needed at the start of the race - if you crash into someone immediately at the race start you can instantly accelerate to the maximum speed, so you need RNG to put another car close enough to you for that to happen faster than the RTA record.
