Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player
(609)
Joined: 2/26/2020
Posts: 698
Location: California
SmashManiac wrote:
This is an easy Yes vote for me! I never would have considered waiting on the GBC BIOS screen to manipulate RNG! I would be really interested to know why this works in more details.
Using the GBC BIOS to manip RNG has been done before. It's rather simple, changing palette (or other actions), at certain points makes the BIOS exit at different times, which is explained here. Even more changing the palette resets a delay counter, which makes it so you may delay the end of the BIOS for as long as you want. It's likely seeding gets done with rDIV or some other hardware timer which would be affected by palette changes and these delays.
Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player
(609)
Joined: 2/26/2020
Posts: 698
Location: California
MESHUGGAH, this is an official warning. Drop this. Questioning why this "abomination" was accepted in the manner you did was simply rude. You may not have intended this, but your "Amazing superplay" comment could be seen as sarcastic too. Don't escalate this further.
EZGames69 too, please don't escalate this either, the moderation team will handle this. Sarcastic remarks back aren't helping here.
Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player
(609)
Joined: 2/26/2020
Posts: 698
Location: California
Emu83 is a brand new TI83 emulator developed by myself, written in C. It features much faster emulation than TI83Hawk, around the same accuracy, proper savestates, and the ability to load in all link files at once via the multidisk bundler, allowing for proper TASing of most games for the TI83. This will be available as a BizHawk core in 2.8, although for now you can use the core in development builds. Report any issues here or the GitHub issue tracker.
Guide on using this core by TiKevin83:
Link to video
Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player
(609)
Joined: 2/26/2020
Posts: 698
Location: California
MUGG wrote:
TiKevin83 wrote:
To make sure, you're using console mode GBA?
I'm using mode GBC mode. Oh, that might explain why there was such a time gap between the two movies.
The boot-up on GBA takes a bit longer.
Should the GBC mode be used because it is fastest?
Or should the GBA mode be used to be able to play it back on console?
Using one console or the other for purely for a faster boot time is not considered a gameplay improvement here. You wouldn't be penalized for using GBA mode. Up to you for which mode you want to use.
Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player
(609)
Joined: 2/26/2020
Posts: 698
Location: California
Darkman425 wrote:
I'm doing research about the current game end glitch for Final Fantasy Legend 2 and I have a question about what would be an acceptable ending. The current RTA world record by catklaw that I found just prints the "The End" screen from after the post credits scene. The glitch that was used can also flip the flag for the final boss to be considered beaten and through other effects just requires going to the right NPC to trigger the credits properly. This of course takes a little more time to complete due to more actions to reach the credits roll but is what I consider a more proper ending. Would it be acceptable if I took the slower path for the sake of actually rolling the game's credits or would having the game print the "The End" screen be considered the faster acceptable ending?
A more "complete" ending would reduce questions of legitimacy, and it's perfectly fine as a personal choice. It wouldn't be penalized against you.
Darkman425 wrote:
I also plan on using GBC mode since the Game Boy Player 1 runs games faster than Game Boy hardware. The different platforms would likely also affect the starting game RNG if it's anything similar to how it's initialized in The Final Fantasy Legend, the first game in the series. That may also affect what the resulting end time come to but is more likely to be accurate to hardware since Super Game Boy initial RAM addresses aren't set correctly for The Final Fantasy Legend in emulators and will probably also apply to Final Fantasy Legend 2.
As a note in reality initial RAM on hardware is essentially "random with patterns" with said patterns varying with different consoles/environments/etc. It's not really a good measure of accuracy nor legitimacy in the first place as emulators are forced to be deterministic (which initial RAM is not on hardware) since that is needed for TASing (and even for the sake of console verification RAM is set to the same pattern as emulators with custom ROMs before switching over to the actual game). Don't worry about legitimacy issues here, it's the last thing on the list that can cause that kind of issue.
Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player
(609)
Joined: 2/26/2020
Posts: 698
Location: California
KidKoopa wrote:
CasualPokePlayer wrote:
Typically the emulator will just provide it for you. You can also just go google SHA1 file checksum online and get a site which gives the SHA1 of a file.
That said, typically this isn't really needed for a submission, it's more needed when the movie format doesn't provide such info (rare, especially with BizHawk used for most tasing and it providing this functionality in the movie format and throws warnings if it detects mismatches) or it's hard to get such info (binary movie formats, e.g. Dolphin's movie format) or the checksum contained is just garbage (VBA with CRC8, which is very unreliable).
I used Bizhawk 2.7, and probably will continue to use Bizhawk for any TAS I make in the future. So does this mean the b2k file I submit will provide judges the checksum for me?
Yes, BizHawk provides a checksum within it's movie file right in the header file (which is easily obtainable as the bk2 is just a zip file in reality).
Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player
(609)
Joined: 2/26/2020
Posts: 698
Location: California
KidKoopa wrote:
I finished making my first TAS today, and I have everything I need to submit it, except for a checksum. The submission instructions says to provide a checksum of your ROM (SHA1 recommended). I have no idea what this is or how to find it. Can someone give me easy instructions on how to find it?
Typically the emulator will just provide it for you. You can also just go google SHA1 file checksum online and get a site which gives the SHA1 of a file.
That said, typically this isn't really needed for a submission, it's more needed when the movie format doesn't provide such info (rare, especially with BizHawk used for most tasing and it providing this functionality in the movie format and throwing warnings if it detects mismatches) or it's hard to get such info (binary movie formats, e.g. Dolphin's movie format) or the checksum contained is just garbage (VBA with CRC8, which is very unreliable).
Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player
(609)
Joined: 2/26/2020
Posts: 698
Location: California
phoenix1291 wrote:
CasualPokePlayer wrote:
This is known, and is also why the appveyor links are removed on the readme. We'll be getting something up and running sometime eventually, but for now you'll have to live with gitlab/building yourself/asking devs for builds.
Nymashock anyways I've been sharing builds on TASVideos Discord due to this problem, I'll send builds via private messages too in case you aren't in the discord and don't want to join the discord.
Thanks, I took the version you shared on Discord.
That version is this buid, right?
You can send me your buids too, if you don't mind, I'm almost never on Discord.
Yes that would be the build. I expect the PR to be merged soon-ish anyways so gitlab will be an option soon enough anyways.
Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player
(609)
Joined: 2/26/2020
Posts: 698
Location: California
This is known, and is also why the appveyor links are removed on the readme. We'll be getting something up and running sometime eventually, but for now you'll have to live with gitlab/building yourself/asking devs for builds.
Nymashock anyways I've been sharing builds on TASVideos Discord due to this problem, I'll send builds via private messages too in case you aren't in the discord and don't want to join the discord.
Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player
(609)
Joined: 2/26/2020
Posts: 698
Location: California
Lua scripts generally have to use emulator implemented functions (such as reading memory), which will obviously vary from emulator to emulator. Half the time these functions share the same name (somewhat out of coincidence, somewhat due to simplicity of naming some functions), the other half other of the time there will be an equivalent function under a different name. Or maybe there's just no equivalent and you're screwed over. But there isn't any way to automatically "convert" scripts besides some giant regex (which even is not guaranteed to work). BizHawk's lua functions are fairly well documented, so it shouldn't be too hard to convert scripts.
Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player
(609)
Joined: 2/26/2020
Posts: 698
Location: California
Different emulators have different levels of accuracy, usually resulting in different lag or maybe different RNG. Importing a movie has no guarantee on sync, it just gives you a movie file which can then be manually resynced.
Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player
(609)
Joined: 2/26/2020
Posts: 698
Location: California
Fortranm wrote:
Out of curiosity, is there anything particularly special about Butterfree in this game?
You can get it early in the game and it is garbage everywhere. It's used more as a fun in-community meme rather than anything objective. If you want the worst of the worst you could do the entire thing with garbage unevolved Pokemon, or if you want an early fully evolved Pokemon Beedrill would be even worse due to its movepool.
Also Drowzee has an evolution so I don't get that part.
Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player
(609)
Joined: 2/26/2020
Posts: 698
Location: California
Putting a core into Bizhawk (either by porting or making the core) is not a trivial task. The dev team is also just a bunch of volunteers doing what they want. Don't expect any core requests to be fulfilled any time soon. Maybe eventually they'll be done, but that will take time and someone with the know-how wanting it.
(and of course always the option of doing it yourself and PR'ing it)
Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player
(609)
Joined: 2/26/2020
Posts: 698
Location: California
ThunderAxe31 wrote:
So, I've started TASing this game. I've picked up Blue Rescue Team because it loads dungeon floors much faster than Red Rescue Team, otherwise it's basically the same. Oh and the fact that the game screen is bigger makes it a little nicer. The stylus inputs don't seem to save much time for menuing.
Here is my current WIP movie file: User movie #75972551996754762
And here is the Lua script I've made for bruteforcing floor layouts: User movie #76000472267214701
I've picked Totodile as player and Torchic as partner, as advised by AntyMew. So far I've only beaten the first dungeon. As you can see from the movie, I managed to spawn close to the stairs for all 3 floors of Tiny Woods, thanks to major RNG manipulation.
The problem is that I couldn't use my Lua script for the second dungeon. For the script to work, it needs to overwrite the floor RNG value before that the game uses it for generating the relative floor. However, for the second floor, the game reads the value in the same frame that it has updated it. This is a problem, because BizHawk can write RAM values only at the beginning of a frame. Freezing the RAM addrasse with the cheat functionality doesn't help, because that does actually just overwrite the RAM value at the beginning of every frame, instead than actually preventing it from changing! So right now it looks like my script is unusable, unless BizHawk gets updated with a proper cheat functionality. Or if anyone has any alternative, it would really help!
BizHawk can write to RAM midframe by abusing memory callbacks. Although the DS core doesn't have that yet, so you don't really have many options other than trying random crap to coerce those two actions on separate frames.
Also "BizHawk gets updated with a proper cheat functionality" isn't quite how it works. Most cores just call a function which will run a frame, which can't be interrupted to do writes to memory unless via memory callbacks, ultimately core dependent. This is what you get with a generic cheat framework, anything better requires adding in memory callbacks into cores and using those (not exactly fun) or doing the cheats inside the core (which is ew).
Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player
(609)
Joined: 2/26/2020
Posts: 698
Location: California
Already a request for that made on the github: https://github.com/TASEmulators/BizHawk/issues/2419
Also, the scrolling behavior you describe could possibly be just due to how big savestates get for newer systems and the default settings being rather low for them. I suggest going into TAStudio's savestate settings (Metadata->Savestate History Settings...) and quadrupling every "buffer size" setting.
Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player
(609)
Joined: 2/26/2020
Posts: 698
Location: California
MrTASer wrote:
I want to TAS arcade games in BizHawk for it is my most user friendly emulator but I can't find a perfect way to even open the arcade games in it...
Please tell me a way in which I can open and make some TASes.
Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player
(609)
Joined: 2/26/2020
Posts: 698
Location: California
I did go through the movie for a while, and I couldn't find any sort of improvements. If these potential hidden improvements would require far more effort, then it's not rejection worthy at all.
Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player
(609)
Joined: 2/26/2020
Posts: 698
Location: California
DeSmuME-VS2019-x64-Release is indicating you got a dev build of DeSmuMe. This itself doesn't actually provide useful information on the version used, the exact revision would likely be needed to have a chance at syncing this movie.
Although luckily it seems that this movie happens to sync fine on the current latest revision (3ba38212281096f8417a0cce24396200923a2810).
(Also this dev build usage is why the movie parser is claiming this movie starts from a savestate, due to movie format changes after the latest stable build; which on a sidenote I don't know what you are talking about a .dsv file as the movie seems to expecting no save data)
Still, this movie does not seem to sync old stable builds... and dev builds are not really acceptable for submission barring some certain cases. Don't know if said cases apply here.
The movie itself is well done and quite entertaining, good job on that!
