Posts for CasualPokePlayer

Posted: 9/2/2022 2:15 AM

Quote

CasualPokePlayer

Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player (609)

Joined: 2/26/2020
Posts: 698
Location: California

Spikestuff wrote:

Following the string of links that lead to the GitHub page to another GitHub page implies/showcases that the once made for GBJAM is now following a more active development for the passion project, meaning that the version you've submitted is (and it says on the itch.io page) incomplete, missing features, and not fleshed out. Maybe wait when the development to this demake is actually complete instead of running through an unfinished prototype.

In fairness, the "new" version repo was last updated 3 months ago, and seems to not actually have anything meaningful present. Literally the only code I can see is some hello world code and generic C lib functions. The repo also was start 2 years ago (which is also around the last time the "old" version's repo was updated) and had a large period of inactivity (nothing between Sept 22, 2020 and May 10, 2022), and now seems to have yet another period of inactivity. I'd consider it unlikely this person actually "completes" this project given this, and the "old" version is probably the best you'll ever have for this project.

Posted: 8/28/2022 1:09 AM
(Edited: 8/28/2022 1:11 AM)

Quote

CasualPokePlayer

Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player (609)

Joined: 2/26/2020
Posts: 698
Location: California

Now with the source we now know the emulator is very much nondeterministic (essentially Dolphin dual core on steroids) and would require a lot of modifications to make it deterministic. So we probably shouldn't expect a TAS capable emulator out of this any time. Also when such comes, you could expect the performance to tank a bit.

Posted: 8/27/2022 8:43 AM

Quote

CasualPokePlayer

Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player (609)

Joined: 2/26/2020
Posts: 698
Location: California

Spikestuff wrote:

To my knowledge, a 9k-Intel CPU and an NVIDIA 20 series will sync this movie. Anything with a combo of an AMD CPU/GPU will not sync.

I don't think it's AMD CPU/GPU specific. Two AMD-less machines do not sync this movie either (also regardless of GPUs used, both iGPU and dGPU desync same way). They have old Intel CPUs (Haswell and Broadwell) so I suspect it's older CPUs not syncing well (something with newer CPUs having something older CPUs don't have).

Posted: 8/18/2022 2:51 PM

Quote

CasualPokePlayer

Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player (609)

Joined: 2/26/2020
Posts: 698
Location: California

You're arguing about a judgement which is obsolete due to various rule changes. A separate branch would be unlikely at the time given it didn't have the entertainment for Moons in the first place. And while I too don't agree with the statement with Vault, it is accurate to how Vault operated. The newer rules are probably flexible enough to give it a separate branch and at least allow publication.

Posted: 8/10/2022 5:12 AM

Quote

CasualPokePlayer

Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player (609)

Joined: 2/26/2020
Posts: 698
Location: California

Most of the performance updates don't apply to us, due to our use of interpreter rather than the recompiler. It'll be updated eventually, but don't expect massive performance gains from an update.

Posted: 7/30/2022 9:19 PM
(Edited: 7/31/2022 1:34 AM)

Quote

CasualPokePlayer

Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player (609)

Joined: 2/26/2020
Posts: 698
Location: California

BizHawk is not supported for MacOS. This is largely due to our use of 64 bit WinForms, which isn't supported on MacOS. What exists are builds of very old versions of BizHawk using a hacky reimplementation of WinForms. The latest builds seem to be down for now, but should come back soonish (EDIT: They're back). Be aware too that if you're on an M1 Mac (i.e. ARM based ones), you will need Rosetta. This is on top of other dependencies listed in that thread. (You're also probably best off making a Window's VM if anything)

Posted: 7/27/2022 12:12 AM

Quote

CasualPokePlayer

Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player (609)

Joined: 2/26/2020
Posts: 698
Location: California

A look at the IOS_FMT seems to say it happens *before* the game is actually loaded. Or more so when the game is being hashed. Not sure why that would have issues though, I might just disable panic handlers here so that doesn't pop up while the game is being hashed.

Posted: 7/26/2022 7:55 PM

Quote

CasualPokePlayer

Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player (609)

Joined: 2/26/2020
Posts: 698
Location: California

IOS is only a Wii thing? GameCube doesn't have that. Why it's appearing I don't know, it would have to guess it's getting blocked. Only the second disc name is actually stored in the dtm, you can see that here: https://tasvideos.org/EmulatorResources/DTM There shouldn't be any reason why XMLs don't work with non-iso disc formats. In testing the IR does work "fine", however it breaks badly when in conjunction with accelerometer data, as games are trying to figure out the position of the wiimote with both, and the accelerometer values with the IR values generally end up making no sense. It needs to be reworked so the position of the wiimote is inputted (along with its angle and such), and IR/accelerometer values derive from that.

Posted: 7/26/2022 6:10 PM
(Edited: 7/26/2022 6:32 PM)

Quote

CasualPokePlayer

Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player (609)

Joined: 2/26/2020
Posts: 698
Location: California

The first just indicates it failed to write a file. Make sure you are not placing the BizHawk folder into Program Files or w/e where it would not have write permissions. Maybe reboot computer? idk, might also just be AV software just screwing Dolphin over? The second is telling you the issue, the disc name is too big for the DTM dump. There is a 40 character limit for the file name (including extension and dot) for the disc file for DTMs. Rename the file accordingly. I don't know what the third thing is, I have to assume it's throwing here? https://github.com/TASEmulators/BizHawk/blob/bd58bde07c887f0d76c96d41527a9ecca9e83c00/src/BizHawk.Client.EmuHawk/movie/PlayMovie.cs#L219

Posted: 7/26/2022 12:11 AM
(Edited: 7/26/2022 12:12 AM)

Quote

CasualPokePlayer

Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player (609)

Joined: 2/26/2020
Posts: 698
Location: California

InputEvelution wrote:

CasualPokePlayer wrote:
No SaveRAM support, save data created by dolphin is wiped on core boot (if you really want to preserve save data, use savestates)
This aspect concerns me a bit in relation to TASes that rely on content being unlocked. Does a half-hour TAS really need to include the 5 hours of input required to unlock the game mode inside the same movie file? What does this do to entertainment and acceptability (especially if that same 5 hours can be found in another already published TAS))?

It's more so a technical problem due to Dolphin tying save data to file I/O and that's not simple to hack around. Along with such TASes being so rare it's on the bottom of the list in priority.

Posted: 7/13/2022 8:24 PM
(Edited: 7/13/2022 8:26 PM)

Quote

CasualPokePlayer

Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player (609)

Joined: 2/26/2020
Posts: 698
Location: California

Probably just TMD data is missing for it. Should be an easy fix.

Posted: 7/13/2022 10:02 AM

Quote

CasualPokePlayer

Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player (609)

Joined: 2/26/2020
Posts: 698
Location: California

phoenix1291 wrote:

CasualPokePlayer wrote:
RVZ is a loseless compression format, that's not going to be an issue.
My question was more regarding the RZV to ISO format.

As the compression is loseless, converting to ISO will result in the original ISO that was used to convert to RZV.

phoenix1291 wrote:

CasualPokePlayer wrote:
The "bug" you see in MediaInfo is more due to the A/V stuff not splitting on framerate change (on the first few frames the VI is set to "120 FPS", and shortly it'll be set to its proper FPS). The framerate in the file is whatever the framerate was at the beginning of capture.
Ok, so that's also why even with a native 16/9 GameCube or Wii game the first frame is 4:3?

That is unrelated.

Posted: 7/13/2022 5:01 AM
(Edited: 7/26/2022 12:06 AM)

Quote

CasualPokePlayer

Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player (609)

Joined: 2/26/2020
Posts: 698
Location: California

The only difference I'm seeing here is one has "Mode" (I assume that's permissions?) empty. I assume that just means no permissions are assigned it. Seems to not matter for Windows, I assume it's only a possible problem on Linux. EDIT: This got fixed a while ago.

Posted: 7/13/2022 4:52 AM
(Edited: 7/13/2022 4:54 AM)

Quote

CasualPokePlayer

Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player (609)

Joined: 2/26/2020
Posts: 698
Location: California

ikuyo wrote:

I attempted to download [4594] Linux Iconoclasts by Serena in 22:47.36's ltm file and while the download finished, the zipped ltm file (serena-iconoclasts.ltm) was not recognized as a valid ltm file for libTAS or even as a .tar.gz file (it appeared to be either corrupted or empty). Downloading the zipped submission from #7292: Serena's Linux Iconoclasts in 22:47.36 gave me the file she uploaded (iconoV2.ltm) which works as expected.

It seems to be a valid .tar.gz file as far as 7zip sees. The SHA-1 also matches the submission file. Might be something weird during that one time you tried? Try downloading the file again.

Posted: 7/12/2022 2:12 AM

Quote

CasualPokePlayer

Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player (609)

Joined: 2/26/2020
Posts: 698
Location: California

RVZ is a loseless compression format, that's not going to be an issue. The "bug" you see in MediaInfo is more due to the A/V stuff not splitting on framerate change (on the first few frames the VI is set to "120 FPS", and shortly it'll be set to its proper FPS). The framerate in the file is whatever the framerate was at the beginning of capture. In practice, this isn't so bad as video players seem to handle the videos OK anyways, and a framerate change from PAL will accompany a resolution change which will have the desired result of splitting the video.

Posted: 7/12/2022 12:38 AM
(Edited: 7/12/2022 12:41 AM)

Quote

CasualPokePlayer

Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player (609)

Joined: 2/26/2020
Posts: 698
Location: California

The core is not waterboxed (if it were, you'd deal with far more slowdown by dealing with software renderer / Interpreter), also consider some people having desyncs in the first place meaning it's not waterboxed. The slowdown is mostly overhead BizHawk has, along with the design being antithetical to how Dolphin works.

Posted: 7/12/2022 12:07 AM

Quote

CasualPokePlayer

Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player (609)

Joined: 2/26/2020
Posts: 698
Location: California

phoenix1291 wrote:

However increasing the video memory seems to help it crash less often.

What do you mean by this?

Posted: 7/7/2022 7:28 AM
(Edited: 8/2/2022 2:37 AM)

Quote

CasualPokePlayer

Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player (609)

Joined: 2/26/2020
Posts: 698
Location: California

Doing some more research into this game, there seems to be some extra code ran in case a loop occurs, here's the subroutine responsible for that

Language: asm
8503: DD 7E 06     LD A, (IX+06h)             AF:0054 BC:0000 DE:0003 HL:72E6 IX:7291 IY:729E SP:73BA Cy:124485784 cnP-H-Zse
8506: DD BE 02     CP (IX+02h)                AF:0154 BC:0000 DE:0003 HL:72E6 IX:7291 IY:729E SP:73BA Cy:124485803 cnP-H-Zse
8509: C8           RET Z                      AF:019B BC:0000 DE:0003 HL:72E6 IX:7291 IY:729E SP:73BA Cy:124485822 CNp3H-zSe
850A: 3A 8E 72     LD A, (728Eh)              AF:019B BC:0000 DE:0003 HL:72E6 IX:7291 IY:729E SP:73BA Cy:124485827 CNp3H-zSe
850D: FE 02        CP 02h                     AF:019B BC:0000 DE:0003 HL:72E6 IX:7291 IY:729E SP:73BA Cy:124485840 CNp3H-zSe
850F: 28 26        JR Z, +26h                 AF:0193 BC:0000 DE:0003 HL:72E6 IX:7291 IY:729E SP:73BA Cy:124485847 CNp-H-zSe
8511: DD 7E 02     LD A, (IX+02h)             AF:0193 BC:0000 DE:0003 HL:72E6 IX:7291 IY:729E SP:73BA Cy:124485854 CNp-H-zSe
8514: DD 77 06     LD (IX+06h), A             AF:0993 BC:0000 DE:0003 HL:72E6 IX:7291 IY:729E SP:73BA Cy:124485873 CNp-H-zSe
8517: 3D           DEC A                      AF:0993 BC:0000 DE:0003 HL:72E6 IX:7291 IY:729E SP:73BA Cy:124485892 CNp-H-zSe
8518: 28 1D        JR Z, +1Dh                 AF:080B BC:0000 DE:0003 HL:72E6 IX:7291 IY:729E SP:73BA Cy:124485896 CNp3h-zse
851A: 3A 84 72     LD A, (7284h)              AF:080B BC:0000 DE:0003 HL:72E6 IX:7291 IY:729E SP:73BA Cy:124485903 CNp3h-zse
851D: 06 03        LD B, 03h                  AF:040B BC:0000 DE:0003 HL:72E6 IX:7291 IY:729E SP:73BA Cy:124485916 CNp3h-zse
851F: FE 01        CP 01h                     AF:040B BC:0300 DE:0003 HL:72E6 IX:7291 IY:729E SP:73BA Cy:124485923 CNp3h-zse
8521: 28 08        JR Z, +08h                 AF:0402 BC:0300 DE:0003 HL:72E6 IX:7291 IY:729E SP:73BA Cy:124485930 cNp-h-zse
8523: 06 04        LD B, 04h                  AF:0402 BC:0300 DE:0003 HL:72E6 IX:7291 IY:729E SP:73BA Cy:124485937 cNp-h-zse
8525: FE 04        CP 04h                     AF:0402 BC:0400 DE:0003 HL:72E6 IX:7291 IY:729E SP:73BA Cy:124485944 cNp-h-zse
8527: 38 02        JR C, +02h                 AF:0442 BC:0400 DE:0003 HL:72E6 IX:7291 IY:729E SP:73BA Cy:124485951 cNp-h-Zse
8529: 06 05        LD B, 05h                  AF:0442 BC:0400 DE:0003 HL:72E6 IX:7291 IY:729E SP:73BA Cy:124485958 cNp-h-Zse
852B: DD 7E 03     LD A, (IX+03h)             AF:0442 BC:0500 DE:0003 HL:72E6 IX:7291 IY:729E SP:73BA Cy:124485965 cNp-h-Zse
852E: B8           CP B                       AF:0442 BC:0500 DE:0003 HL:72E6 IX:7291 IY:729E SP:73BA Cy:124485984 cNp-h-Zse
852F: 30 06        JR NC, +06h                AF:0493 BC:0500 DE:0003 HL:72E6 IX:7291 IY:729E SP:73BA Cy:124485988 CNp-H-zSe
8531: DD 34 03     INC (IX+03h)               AF:0493 BC:0500 DE:0003 HL:72E6 IX:7291 IY:729E SP:73BA Cy:124485995 CNp-H-zSe
8534: DD 34 04     INC (IX+04h)               AF:0401 BC:0500 DE:0003 HL:72E6 IX:7291 IY:729E SP:73BA Cy:124486018 Cnp-h-zse
8537: AF           XOR A                      AF:0401 BC:0500 DE:0003 HL:72E6 IX:7291 IY:729E SP:73BA Cy:124486041 Cnp-h-zse
8538: C9           RET                        AF:0044 BC:0500 DE:0003 HL:72E6 IX:7291 IY:729E SP:73BA Cy:124486045 cnP-h-Zse

Most of this is fluff, the main prize is at the end where some counters are maybe incremented. 0x7293 contains the difficulty counter (this can be seen easily as you are given the choice of starting at 1 to 4 for difficulty before the game begins). Another prize is the final check for aborting doing these increments: the difficulty increase is aborted if the difficulty counter is >= 5, thereby capping the difficulty to 5. I suppose this means that a single "loop" in terms of game logic is 8 rounds. Assuming the TAS starts on difficulty 4, it would need to do a second loop for the "hardest" difficulty, after which the game will forever loop with the same difficulty. I suppose given this, there are three options a TAS could go for: 1. Do all unique levels (so 4 rounds). 2. Do 1 full loop (so 8 rounds). 3. Play until difficulty stops increasing (so 16 rounds). 1 is enough for publication here, 3 would obsolete, not sure about 2 (this is an odd case not clearly defined by the rules). This current submission doesn't fit these however as it opts to do a "second loop" and doing a 5th round as a result (in reality, this is a partial loop of course). So it could be trimmed to the 4th round, or opt to do the 2nd/3rd options here.

Posted: 7/7/2022 5:36 AM
(Edited: 8/2/2022 2:37 AM)

Quote

CasualPokePlayer

Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player (609)

Joined: 2/26/2020
Posts: 698
Location: California

Looking at the code, seems that it uses a 8 round loop, going 1 -> 2 -> 1 -> 3 -> 2 -> 1 -> 3 -> 2. The lookup table used for this loop can be found in the ROM file at offset 0x4FB (and changing it will subsequently change level order).

Posted: 7/6/2022 7:20 PM

Quote

CasualPokePlayer

Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player (609)

Joined: 2/26/2020
Posts: 698
Location: California

InfamousKnight wrote:

Gtk-Message: 15:12:35.192: Failed to load module "canberra-gtk-module" Error: Trying to access Windows-1252 fonts but they are not loaded. Games may not show fonts correctly, or crash. Killed

Ok sounds like there's some weird issue here, I did fix the font issue a while ago. You are using the latest build right? Is there a Sys directory next to the BizHawk.exe?

Posted: 7/6/2022 7:07 PM

Quote

CasualPokePlayer

Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player (609)

Joined: 2/26/2020
Posts: 698
Location: California

InfamousKnight wrote:

CasualPokePlayer wrote:
InfamousKnight wrote:
Need for speed carbon still has a black screen for some reason.
Can you try enabling Panic Handlers and seeing what gets outputted in the laststderr file?
How do I do that?

Scroll down in the Main Settings

Posted: 7/6/2022 6:30 PM

Quote

CasualPokePlayer

Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player (609)

Joined: 2/26/2020
Posts: 698
Location: California

InfamousKnight wrote:

Need for speed carbon still has a black screen for some reason.

Can you try enabling Panic Handlers and seeing what gets outputted in the laststderr file?

Posted: 7/5/2022 6:03 PM
(Edited: 8/2/2022 2:37 AM)

Quote

CasualPokePlayer

Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player (609)

Joined: 2/26/2020
Posts: 698
Location: California

EZGames69 wrote:

What does the time save look like?

As I stated, nearly 1 second (it's 57 frames). Although how much would be lost/gained due to different rng in the last level I can't tell without tasing that myself.

Posted: 7/5/2022 6:21 AM
(Edited: 8/2/2022 2:37 AM)

Quote

CasualPokePlayer

Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player (609)

Joined: 2/26/2020
Posts: 698
Location: California

On the second level, you seem to wait for the enemy to pass before trying to put the final key in its lock. At a glance this makes sense, although raises some questions of RNG manip / overall planning. However, those questions are a bit moot here, as I tried to rush the key in the hole instead of waiting for the enemey, and that seems to work? This itself saves nearly a second, so I'm not sure why this wait is done? Is this some sort of RNG manipulation (saving enough time to justify the initial nearly 1 second timeloss)? Or was this just a mistake? The former seems a bit unlikely here, as the TAS seems to remain "synced" until the final level anyway?

Posted: 7/4/2022 1:21 AM
(Edited: 7/4/2022 8:22 AM)

Quote

CasualPokePlayer

Emulator Coder, Experienced Forum User, Judge, Published Author, Experienced player (609)

Joined: 2/26/2020
Posts: 698
Location: California

phoenix1291 wrote:

Taxi 3 crashes continuously.

Seems like this is a case of the game having a "too big" resolution. Should be as simple as doubling the amount of space I'm giving for the core to dump video data into.

phoenix1291 wrote:

Don Bluth Presents Dragon's Lair Trilogy (USA) (Wbsf, Wii) constantly crash.

Probably same issue, probably fixed in latest build?