Posts for Dimon12321
Dimon12321
He/Him
Experienced Forum User, Published Author, Active player (485)
Joined: 4/5/2014
Posts: 1139
Location: Ukraine
Looking forward to seeing a submission with Level 7 in 4 years. Yes vote
TASing is like making a film: only the best takes are shown in the final movie.
Dimon12321
He/Him
Experienced Forum User, Published Author, Active player (485)
Joined: 4/5/2014
Posts: 1139
Location: Ukraine
Wow, I felt the sense of deja-vu that I've already seen a similar game. Yeah, sounds and voices would really need an upgrade, and the scores counting up on the last mission sound really repulsive. Overall, yes vote
TASing is like making a film: only the best takes are shown in the final movie.
Dimon12321
He/Him
Experienced Forum User, Published Author, Active player (485)
Joined: 4/5/2014
Posts: 1139
Location: Ukraine
BlackWinnerYoshi wrote:
Dimon12321 wrote:
Yoshi, would you please check it?
It synced the first time, including video dumping
Thank you very much!
TASing is like making a film: only the best takes are shown in the final movie.
Dimon12321
He/Him
Experienced Forum User, Published Author, Active player (485)
Joined: 4/5/2014
Posts: 1139
Location: Ukraine
Stage 3 is finished. Half way through! This time, the cursed point was Area 5. Two desyncs =( Link to video User movie #638242302295917953 Yoshi, would you please check it?
TASing is like making a film: only the best takes are shown in the final movie.
Dimon12321
He/Him
Experienced Forum User, Published Author, Active player (485)
Joined: 4/5/2014
Posts: 1139
Location: Ukraine
Darpmosh wrote:
So do you think the reason why this wasn't done at all was because of lack of interest in accuracy? Do you even think that it is possible?
Most likely, that's indeed can be lack of interest. End users usually don't know what is the percentage point, below which a game is considered "inaccurate", so they end up sending good feedback and developers, accordingly, don't see big demand towards accuracy improvement. What's regarding developers, sure they know the accuracy situation better, but at some point, they don't consider further time investments reasonable to push emulation accuracy further. I see 2 obstacles here: - the difficulty - developers know what the problem is, but it may be complex and require a lot of time coding and testing, - the uncertainty - developers doesn't know what the problem is, so they end up re-reading available documentations for 101th time and doing micro-adjustments to the complicated architecture to see if it improves accuracy and not ruins anything else. Sorry for bla-bla-blaing P.S: I agree with CasualPokePlayer
TASing is like making a film: only the best takes are shown in the final movie.
Dimon12321
He/Him
Experienced Forum User, Published Author, Active player (485)
Joined: 4/5/2014
Posts: 1139
Location: Ukraine
Darpmosh wrote:
Hmm, so in other words, accuracy to the original console, huh. So are you alright with an emulator that runs at an abysmal 10FPS in exchange for it being one of the most accurate emulators of a console, esp considering that other consoles don't even come as close to this emulator in terms of accuracy?
Well, I think more experienced people may give a wider feedback on this regard, but from my point of view, the priorities should be in the following order: 1. Moderate emulation accuracy - have approximate emulation regarding games' graphics, sound and logic. In short, have most games playable, completable and going by the flow how a real PS2 game works, without all those crazy glitches which can skip half the game which cannot be done on a real PS2. 2. TAS movie playback stability, so a person would "almost never" have cases of movie desyncs, so he would have to be redo everything again. TASing is a very time-consuming process. Desyncs spoil this =( 3. Tolerable speed while maintaining points 1 and 2. 4. High emulation accuracy If you'll have an emulator being 99.9% accurate to the real PS2, this will be a revolution, since even Dolphin cannot achieve it (at least, in terms of TAS verification on the real console).
TASing is like making a film: only the best takes are shown in the final movie.
Dimon12321
He/Him
Experienced Forum User, Published Author, Active player (485)
Joined: 4/5/2014
Posts: 1139
Location: Ukraine
SJ wrote:
also i think should do a playaround tas next time and save it for next time and also thank you for watching my tases especially the armageddon one!
I have doubts you will obsolete Akheon's playaround movie. With the lack of user feedback nowadays, it's gonna be debatable if your entertainment value will be considered higher than his. On the other hand, the judges may implement extra tiers, like "playaround, volume 2". You'd better ask them out.
TASing is like making a film: only the best takes are shown in the final movie.
Dimon12321
He/Him
Experienced Forum User, Published Author, Active player (485)
Joined: 4/5/2014
Posts: 1139
Location: Ukraine
The Harakiri moment wasn't fast enough and it didn't look that entertaining, so the speed/entertainment trade-off here is questionable. Still, the TAS was pretty watchable and relatively fun. Yes vote.
TASing is like making a film: only the best takes are shown in the final movie.
Dimon12321
He/Him
Experienced Forum User, Published Author, Active player (485)
Joined: 4/5/2014
Posts: 1139
Location: Ukraine
One of the few Japanese NES games I've played in my childhood. What a great discovery the bot has found! That screen scroll that immediately takes to the end of a level. Yes vote. But I'm still wondering what was the point in decreasing your health during the battles, if you could just take damage. It would drain more cosmo points? Also that cosmo draining on the last battle. What it necessary?
TASing is like making a film: only the best takes are shown in the final movie.
Dimon12321
He/Him
Experienced Forum User, Published Author, Active player (485)
Joined: 4/5/2014
Posts: 1139
Location: Ukraine
Looks optimized, but I have 1 question. Isn't it faster to turn around in the air and cover more distance while you fall down being dead?
TASing is like making a film: only the best takes are shown in the final movie.
Dimon12321
He/Him
Experienced Forum User, Published Author, Active player (485)
Joined: 4/5/2014
Posts: 1139
Location: Ukraine
Great work, man! Armageddon's Konquest did a step forward comparing to Deception's Konquest, so the speedrun didn't look boring to me. Yes vote.
TASing is like making a film: only the best takes are shown in the final movie.
Dimon12321
He/Him
Experienced Forum User, Published Author, Active player (485)
Joined: 4/5/2014
Posts: 1139
Location: Ukraine
Looks OK. My favourite section is the night. I found this game a precursor of Zippy Race on NES, but you play as a motorcycle there. Yes
TASing is like making a film: only the best takes are shown in the final movie.
Dimon12321
He/Him
Experienced Forum User, Published Author, Active player (485)
Joined: 4/5/2014
Posts: 1139
Location: Ukraine
Looks spectacular! I noticed you sometimes struggle getting through the doors and walls. In case of lift walls, I assume it's OK since it takes time to load the next level, but in case with doors, the TAS is not optimised because you should clip through each door from the first attempt. When the level is loaded, does your speed get reset? Because you may gain momentum and jump just before a level gets loaded to cover more distance. Why do you stare at the ground while running? Does low FPS (lagging) slow you down? Also, could you comment on Chamber 7? I found the momentum gaining slow. BTW, isn't it faster to let the fireball reach the plug on the roof and activate the platform? Chambers 10 and 11 seems to be slow. In general, there are moments here and there which can be done faster.
TASing is like making a film: only the best takes are shown in the final movie.
Dimon12321
He/Him
Experienced Forum User, Published Author, Active player (485)
Joined: 4/5/2014
Posts: 1139
Location: Ukraine
Did you try performing a different OOB, like shown here? 1:55 Did you try running ahead without hiding from the arrows? Some moments, like 3:56, 4:23, some inventory selections, 5:24, cutscene slow skip and other ones look slow to me. I don't know why you always do a step back before turning around. Killing the Gatling Man looks less optimized than killing Krauzer. It's like by that point you finally got familiar with the control. Some turns and camera angles look a little slow too. I got the feeling like you didn't have enough hands or keyboard input to do them better. Did you use TAS Input? Also, check "Aiming" paragraph in this submission, it might be useful to optimize turns. Reloading the sniper rifle can be improved if you take damage during the reloading animation. I don't remember if it works with CGI cutscenes, but you should try reloading it before fighting Krauzer to see if the cutscene fast-forwarded the animation. What memory card you use for the submission? Did you download it from somewhere or did you use the one that's verified from that submission ("Playing the .dtm file on Dolphin" paragraph)? In short, you first submission is good, but I found it not optimised. Please do an updated version. I don't think it will take long to redo it.
TASing is like making a film: only the best takes are shown in the final movie.
Dimon12321
He/Him
Experienced Forum User, Published Author, Active player (485)
Joined: 4/5/2014
Posts: 1139
Location: Ukraine
BlackWinnerYoshi wrote:
It did work this time, including video dumping, but I won't bother sharing a downloadable encode tho sry
It's alright. Dziekuje bardzo!
TASing is like making a film: only the best takes are shown in the final movie.
Dimon12321
He/Him
Experienced Forum User, Published Author, Active player (485)
Joined: 4/5/2014
Posts: 1139
Location: Ukraine
BlackWinnerYoshi wrote:
Well, it looks like it desyncs in Stage 1, Area 1 for some reason, and the previous movie still works just fine. It also looks like the inputs have been corrupted at some point, so it looks like something broke during the BK2 to DTM conversion?
I was trying to convert it directly from TAStudio, but looks like it records one unnecessary frame because I couldn't jump to frame 0 right away. Here's a new DTM. This time, I exported the movie to .bk2 and dumped the movie User movie #638216767685209812
TASing is like making a film: only the best takes are shown in the final movie.
Dimon12321
He/Him
Experienced Forum User, Published Author, Active player (485)
Joined: 4/5/2014
Posts: 1139
Location: Ukraine
Stage 2 is finished! A lot of progress have been lost due to a broken savestate which led to desyncs. Link to video User movie #638215734154023188 Again, I'm looking for TAS verifiers! BlackWinnerYoshi, FitterSpace would you please check it out?
TASing is like making a film: only the best takes are shown in the final movie.
Dimon12321
He/Him
Experienced Forum User, Published Author, Active player (485)
Joined: 4/5/2014
Posts: 1139
Location: Ukraine
That was fun to watch! A good refresh after [4694] NES Super Orb Bros. "all orbs" by Mizumaririn in 09:17.07 from the last year. Yes vote
TASing is like making a film: only the best takes are shown in the final movie.
Dimon12321
He/Him
Experienced Forum User, Published Author, Active player (485)
Joined: 4/5/2014
Posts: 1139
Location: Ukraine
Spikestuff wrote:
Dimon12321 wrote:
That's cool, but tasvideos does not accept TASes for games that are still in active development.
[5070] Windows Dono's Tale (SAGE '22 Demo) by TheAmazingYucemu & chickennugget in 03:38.75
Wow, didn't notice that =(
TASing is like making a film: only the best takes are shown in the final movie.
Dimon12321
He/Him
Experienced Forum User, Published Author, Active player (485)
Joined: 4/5/2014
Posts: 1139
Location: Ukraine
Super Mario 63 Redux is a game that aims to remake the entirety of Super Mario 63. This is a TAS of the demo version, since the full version is still going to be in development for quite a while.
That's cool, but tasvideos does not accept TASes for games that are still in active development.
TASing is like making a film: only the best takes are shown in the final movie.
Dimon12321
He/Him
Experienced Forum User, Published Author, Active player (485)
Joined: 4/5/2014
Posts: 1139
Location: Ukraine
Looks amazing! I had to watch both this and the elder TAS side-by-side in 2 windows to see your improvements. Yes vote
TASing is like making a film: only the best takes are shown in the final movie.
Dimon12321
He/Him
Experienced Forum User, Published Author, Active player (485)
Joined: 4/5/2014
Posts: 1139
Location: Ukraine
Hello, BikerBones! Welcome to TASVideos. Such a script would be a documentation itself. I'm not aware of someone actually releasing a single script for a GC/Wii game for reference or demonstration. Maybe I missed it, who knows. IIRC, there's at least one submission where Dolphin 5.0 LUA (mentioned in the submission text) was used, so you may give it a try to find it and contact its author. There must be a reason why he/she used that fork. The recent publication of Dragon Warrior 4 for NES provides a library of how a NES game can be brute-forced. I don't Lua at all, but as I understand, there is a core script of how to do basic actions and movements and a lot of scripts for mesh buttons for battles/movement for each game walkthrough chunk (a set of screens of a chapter) until the game is complete. It can be used for reference of what should be done after all useful RAM addresses are found. The only problem is, it was written for Bizhawk emulator, so it might have a different RAM address basis and who knows what else.
TASing is like making a film: only the best takes are shown in the final movie.
Dimon12321
He/Him
Experienced Forum User, Published Author, Active player (485)
Joined: 4/5/2014
Posts: 1139
Location: Ukraine
Wow, this game has probably the best explosion sound of all C64 library. Yes vote
TASing is like making a film: only the best takes are shown in the final movie.
Dimon12321
He/Him
Experienced Forum User, Published Author, Active player (485)
Joined: 4/5/2014
Posts: 1139
Location: Ukraine
Looks like the stability situation for Wii TASing was improved in Dolphin 5.0-17850. Here's that PR discussion, which focuses on various scenarios, including TASing, and a respective Dolphin issue was closed. No Wii TASes have been submitted with this Dolphin version or newer, but this gives hope more complex Wii games would be stable when TASing, as broken savestates is the main reason why TASes go out of sync. Wii also now has MotionPlus support for TAS since 5.0-18878, so the games like Red Steel 2 can now be TASed. There are a handful of Dolphin issues which might be actual even today: https://bugs.dolphin-emu.org/issues/7219 (GC savestates inaccuracy) https://bugs.dolphin-emu.org/issues/9721 (DSP LLE + Memory Card is desync-prone) https://bugs.dolphin-emu.org/issues/9722 (savestate inconsistency with LLE) https://bugs.dolphin-emu.org/issues/9689 (settings changes are not tracked for playback) https://bugs.dolphin-emu.org/issues/10357 (Wii aspect ratio problem for TAS Input)
TASing is like making a film: only the best takes are shown in the final movie.
Dimon12321
He/Him
Experienced Forum User, Published Author, Active player (485)
Joined: 4/5/2014
Posts: 1139
Location: Ukraine
Well, your work is incredible! A working RAM Search and TAStudio (especially, branches) are big advantages over Dolphin. I would switch all my Patreon donations to you (or Bizhawk team), if the core was waterboxed. I came here on this regard. I was building the movie in TAStudio (I work only there) and I decided to replay the whole stage to watch what I have done, and I faced an actual desync. After replaying it from the start-up, doing input manipulations and other assumptions, I found the problem. If I replay the movie from a moment before a specific savestate (indicated as "GreenZone InputLog: Stated", if you go to Settings -> Edit TAStudio colors... in TAStudio window), the movie will bypass that desync and proceed forward, but then it catches another desync, and I hope it was caused by that broken savestate, not by the game complexities savestates wouldn't catch. 1. Is the structure (DTO) of a savestate identical to what Dolphin does? If so, then can my .tasproj file be used to analyze what did that savestate do wrong, detect the problem and implement a better version which keep the movie more stable? I'm not asking you to do that, you have your personal life, passion etc., but still, what do you think? 2. What does "Use Compressed States" checkbox do? Does it only pack/unpack a ready savestate and/or some state data not being saved? 3. I use 5.0-16426 final release. Does 5.0-16793 have better savestates? Nevertheless, thank you, Casual! No objections. I'm sure I would face a similar problem in Dolphin too. UPDATE: 3) So, a savestate in 5.0-16793 release is a little more stable because its components (RAM, video memory, etc.) get saved by a single thread (without multithreading), therefore preventing cases when the components have been saved in the wrong order. However, desyncs still happen rarely and you're ought to periodically rewatch your movie from the very beginning to prevent big progress loses. In Bizhawk's TAStudio, before doing a playback, put some input at the beginning of the movie, or after a savestate you're sure is stable. I had an issue when a desync was spotted only after I invalidated existing savestates in my movie. If you have already done major progress with 5.0-16426, you can download 5.0-16793 release and copy-paste there the file "dolphin-emu-nogui.dll" from your 5.0-16426 folder.
TASing is like making a film: only the best takes are shown in the final movie.