/vr/ - Retro Games

I guess I'll bump this for right now

I think I may redesign one of my levels for a more optimal dragon coin flow.

I literally just realized that I forgot to enable the no-yoshi intro on my level (Capital City of Grinders in the Sky). Can you enable it for me OP?

>>5639912
I'll include that in the next build for sure!

As for levels, in the last thread we discussed the potentisl for a multiperson final level. I actually started creating a plan for the layout in that case, considering some anons were interested. I'd be up for designing the first room to provide a basis for the plan if people want.

I added some more custom music tracks to the game, no longer level specific. Going forward, the baseroms will have a large selection of musics anons can add to the levels.

I also added the No Yoshi Intro to >>5639912's level

My plan was to train a GAN on hundreds of SMW screenshots and recreate the images it generates inside the level editor for a "AI-generated" level, but I got nothing after running it overnight. It'll likely need an order of magnitude more images than the 700 I got.

>>5640832
That's actually super interesting! Would make for quite the level. Have you done any AI related topics in the past?

>>5640837

It's not my own code, I'm using these:

>https://github.com/akanimax/BMSG-GAN (produced flat colors)
>https://github.com/guoyongcs/AEGAN (produced the results I posted)

Playing through romhacks for screenshot material should help. The obvious thing to do would be to record a video and train the GANs on each individual frame, but since the same level features would show up in many images, there's the risk of overfitting in which the GANs just literally copy and paste the features into the "new" images.

>>5640868
Hm, I would suggest taking screenshots rom regular Super Mario World. That, or internal levels within Lunar Magic. There would be a lot more material to use as a result.

>>5638391
I have one level. Its just a long walk then it ends.

How about platform level with no enemies. Or one with enemies but no platforming?

>>5640886
I'm free for anything so long as its designed, and not super shitposty.

>>5640868
>>5640873
Download some longplays of SMW and a few hacks, extract maybe one frame for every 3 seconds of footage in there, and go ham.

If Fire Forest and Four to Zero are not complete by next week, I'll make them unclaimed again. I've heard no word on them since they were claimed, so I think it's fair that after a month of inactivity theyll go unclaimed again.

How are secret levels being tackled? Are they being used at all?

>>5641250
They are allowed and used! Not many anons have decided to make them, although my level included one.

Atm I've given access to all Anons who have requested to join the claim sheet

Bump

I started work on the first room of the final castle. In the case we don't do a Multiperson level, I suppose I'll complete it myself

>>5638391
>hacks
i dunno man the zniggy game looks a lot more interesting

>>5642821
You can do a lot with rom hscks. 4chan is no stranger to collaborative projects either, so why not one more to the list? It certsinly would make for an interesting hack afterall

I am noticing that there are no Level 24 patches compatible with asar...should I use an older patcher or should I just move the final level to a different level number?

>>5643018
I don't think there's anything preventing you from using both asar and xkas on the same rom?

>>5642897
Yeah, no, these projects never fucking work
Even the japanese who are supposed to be super well organized took OVER a decade to finish VIP5 and that hack was surrounded by drama and suicide attempts
If you want to trick anonymous retards into finishing levels you have to make a competition, not a collab, convince them reddit liberals are the enemy who make better levels or some shit. Make up an antagonist and provoke people to be better than each other, incels only work if they feel they are being confronted.

>>5642897
Yeah, no, these projects never fucking work
Even the japanese who are supposed to be super well organized took OVER a decade to finish VIP5 and that hack was surrounded by drama and suicide attempts
If you want to trick anonymous retards into finishing levels you have to make a competition, not a collab, convince them reddit liberals are the enemy who make better levels or some shit. Make up an imaginary antagonist and provoke people one up each other, incels only work when they are being confronted.

>>5643491
To reiterate:
>Let's make a friendly collaboration. Everyone claims a level and send them in whenever they feel like. There's no deadline or peer pressure to do your best work, Once a level is claimed it cannot be stolen by another person. In the end we will assemble everything in a single romhack and there's no given date for the project to be finished either, this could possibly take years.
THIS DOESN'T FUCKING WORK
>Let's make a competition. Everyone sends their levels and they will be ranked and judged in multiple pre-established categories. You have to finish within a deadline and if you don't send in anything you're out. In the end we will assemble everything in a romhack and the project ends within x months.
^^^^^ THIS WORKS

>>5643491
>>5643536
While I get what you're getting at, I don't think that is the way to go about it. This is meant to be a friendly casual thing. The small size (VIP/JUMP were around 100 levels each...this is 36) also makes so it wont drag on. VIP5 only dragged on because they argued over a single level for 2 years (total development took 3 years.) I'm not out to encite anyone at thr moment. The levels submitted are the ones submitted, and whats done after a year is whatll be compiled. Too many variaes with Judging/level moderating too.

>>5643448
Some of the old patch routines in xkas conflict with asar's newer patches, which is my only concern.

>>5643575
So the idea for the final level would involve a lvl24-like mechanic where you get sent to different sublevels depending on non-specified coin and time requirements? I'm honestly not sure that'd work well, especially with how hard the final level should get. Why not simply have 2 exits at the end of each level segment so the player can choose the path he wants?
So, let's say we have Mario go through 2 different sublevels between the first and last ones, we'd get:
>first room: ends with a choice between 2 sublevels
>second room: 2 versions, both with 2 exits
>third room: 4 versions, all with 1 exit to final room
>fourth room: final corridor before final boss
...For a total of 8 small sublevel segments to create. Would that be good?

>>5643763
Ideally I'd rather it work like a traditional bowsers castle with no weird mechanics, but the patch to remove that doesn't work with newer tools. My plan for the castle was to make the player go through three rooms before the finale. Each room however ends with two of its own unqiue room sets, for a total of 14 rooms. Im unsure if thats too many, however. 8-9 seems to be the standard

>>5643802
Well, if you didn't intend to use the lvl24 gimmick, why use level 24 at all? Just use the lvl100+ range, it's fine.

Also, I don't think that's too many rooms, as long as the player doesn't need to go through all of them. So yeah, like a traditional Bowser's Castle. Should the length of these rooms be comparable to those in vanilla?

>>5643891
The whole 024 gimmick slipped my mind until I started work on the first room, which is why I even brought it up. It was always gonna be traditional. As for the length, they should be around the size third a normal level, that way the entire level wouldnt feel bloated, and 11 doors upon replaying wont feel so bad.

Bump. Room halfway done, I'll post images later.

Is anyone using the Yoshi smiling coin bonus thingamajig? I kinda need to have both Yoshi and Podoboos in my level. Since Podoboos upload their graphics dynamically to the same place as Yoshi, they are incompatible. So I made a patch that nukes the dynamic upload and changes the tilemap to use the smiling coin tiles. Would that be okay or am I overstepping my bounds here?

>>5645184
I'd allow it. It's not overly changing core vanilla mechanics, and not adding on completely new sprites into the mix, which is the one thing I did not want to do.

There are 2 doors in Trick House that lead to a test level. Is there a missing sublevel or...?

>>5646024
Oh I forgot to covert them over. Extremely sorry! I changed some of the sublevels to use the ones on the level sheet instead of random ones, but I forgot to switch some of them.

Some progress on my room by the way. The gimmick involves teleporting Grinders (entirely vanilla, based on a weird glitch). Each room should be a tiny bit bigger than this length.

>>5646058
Oops forgot to include the image!

>>5641116

Thanks for the advice, I was able to boost the image count to around 8,200. I also tried making the images greyscale so it wouldn't need to relearn palette swapped backgrounds. Which seems to have worked, but it also made elements of the images harder to make out.

I'll give 30,000 a try, and if that's still not enough data I'll unclaim the level.

>>5646093
Wow, this is incredibly unique! If you're still having issues, I could attempt to help you get the outputs you'd need as I could see this being an extremely interesting level from a design perspective.

Finished the room. Because the grinders were unpredictable, I decided to swap out the gimmick for one involving throw blocks. I think it's quite fun!

>>5646352
I remember a thread or 2 back, you mentioned the idea of using SaGa music for the final boss, like VIP and other such japanese hacks. Did one of these hacks ever use the final boss track from SaGa II (Save the World)?
https://youtu.be/XuzHEDYHsjU
I'm honestly getting kind of burned out on working on my level for the moment, and I feel like porting that song in particular in the meantime. Would that be something you'd like?

>>5646483
I actually picked out two saga songs that I think would be good for porting. This one was already used in a hack iirc. You can port this one or the ones I had in mind if you'd prefer.
https://www.youtube.com/watch?v=Xth1xHOJXBA
https://www.youtube.com/watch?v=M3DD2vDWwOM

Both are on that Midi site btw.

Hot damn, just got my PC running again boys! I'll start working on one.

>>5646602
Sounds good! Be sure to claim your level!

Bump

Why there are no expansion sets for old SMB games like SMW? Just those styles and physics, but new maps by original authors.

FPS games like Doom has a tradition for this, why not SMB?

>>5647363
SMB1 and SMB3 have a lot actually. Perhaps our next project could be for one of those

>>5647382
I didn't talk about fangames, but expansions made by Nintendo in house developers.

>>5647426
Ah, I see what you mean. I suppose the idea would be interesting, but that sorta goes against Nintendo's philosophy.

Getting a bit back on topic, I added some final polishing to my room. I'd say it's fully complete now.

>>5647426
For SMB1, isn't that basically The Lost Levels? There are a few new elements, but otherwise it has the same physics and all.
And SMB3 had the e-reader levels in the GBA remake.

Bump

>>5647363
>Why there are no expansion sets for old SMB games like SMW? Just those styles and physics, but new maps by original authors.
That's what SMB2J is, more or less. It had a couple new features but was almost 100% identical.
If Super Mario Maker kept its original intention of having each style's gameplay faithful to its original iteration, then the levels released by Nintendo would have kind of been that, but then they made all the styles identical to NSMB in gameplay.

I'm thinking of tossing the current World 1 map and create the entire hack's map on the main overworld. Would anons prefer that, or would rather have each map be made on each submap.

Bump...

>>5646093

Well, something's happening.

It doesn't generate enemies, or at least not enemies that are discernible. I'll try bumping up the size to 512x512, tweak the settings and that will be the test to see whether my level idea is viable.

>>5640832
Deepstyling SMW levels seems to work fine.

>>5650860
Oh wow, this looks fantastic~ I'd love to see it in action.

>>5650964
Unfortunately doing this near 1:1 requires the sloping to be able to be flipped without any weird bugs happening.

>>5650973
Technically there is ASM to allow for that, and if not you can always use the cave tileset.

Bump

>>5651679

I'm currently splitting my time between this and Zniggy, so I won't have much progress to report for a while.

crossover when

Is this going to have shell jumps and shit?

>>5652426
I'll allow it but don't make it the focus. I want this hack to stay at a hsrd difficulty, not go into Kaizo often

Bump

I decided to move the overworld to the main map. I'll be posting updates on that in the upcoming days.

Don't die thread ...

Finished planning the overworld out, I just started actually building it.

>>5652426
>>5652541
And here I thought the focus was to keep it in the spirit of the original game, which doesn't require speedrun tricks? Might as well go kaizo if you aren't going to stick with vanilla difficulty.

>>5656820
JUMP and VIP both have shelljumping, but not in every level. I think in both hacks there are like 3 levels total. As a result, I'd allow it, but I don't want it to be a common trend. As you said, this isn't a kaizo hack, and I don't intend to treat it as one.

Design for today. I did not want to place levels for World 2 onward until we have a finalized list, although I am quite a fan of the current design.

I'm making some slow progress on my huge open-ended level, it's starting to take shape. The hardest part is to try and keep backtracking and item babysitting to a minimum. Might post picture this evening.

Also, I'll need some UberASM code to make some enemies graphically compatible with each other (using manual trigger exanimation to load the correct graphics in the parts of the level where certain enemies are). I assume this is okay? It doesn't change vanilla mechanics, it just gets rid of garbled graphics, after all.

>>5658104
Yeah, I'd 100% allow it. I can't wait to see a picture too!

Bump

Did more overworld work, it should be done later today.

I sadly fell a bit ill today, so the map won't be finished today. Sorry folks

Feeling less sick now, so I'll continue.

Also Bump...

The top half is almost done.
I'll claim the resources I added since then:

>Level 23
>Sublevel 2C
>Map16 0C
>ExGFX 88, 89, 8A, 8B (Tileset mixing, sprites, ExAnimation, and vanilla Layer 3 mist tilemap, respectively)
>UberASM code for screen-specific graphics uploading (manual trigger ExAnimation)
>Patch to remap Podoboos to SP2's smiling coin tile, as well as nuke its DMA graphics that make it incompatible with Yoshi
>Replace SP2 globally with the new one w/ Podoboo graphics

I know it's a lot, but I can't help myself, I never was able to fully stick to vanilla...

>>5663497
Forgot to mention, it'll have 3 exits. Exit 1 is a fight against Reznor, Secret Exit 1 is against Big Boo, and 2nd Secret Exit is a secret. For Big Boo to trigger the secret exit, I'd need you to go in the Overworld Editor, then Overworld>Change Boss Sequence Levels, and change the "Boss uses Secret Exit 1 (Big Boo)" to level 23. This is all assuming the level Four to Zero (lvl 13) doesn't need it, but we haven't heard anything about that level since it was claimed, right?

>>5663497
Jesus Christ.

>>5663525
>>5663497
Dude this is insane! I am very excited to play through this. As for Four to Zero, I'm letting the level stay claimed until Monday. I'll unclaim it otherwise.

>>5663497
The Lunar Magic 3.0 square levels is a whole new era of shit that I never even contemplated. Wow

>>5663779
I've been wanting to come back to hacking ever since 3.0 came out, for the sole purpose of trying out this new feature. Never thought I'd end up doing it in a vanilla style, and for a collab no less.
I still need to play around with how items are used and obtained, so item babysitting is reduced (which is 200% worse when you don't know where to go).
Music will probably be Brinstar Overgrown with Vegetation Area, unless I find something better in the meantime.

>>5663779
LM3.0 added a lot of interesting things, which is exactly why I relaunched this project when I did. All these new features could be used for a great project!

Bump

>>5663525
I'm considering on the OW maybe making one of your exits loop back around to an older OW level to save some space. Would you be ok with that?

>>5667397
Making a shortcut to an earlier spot on the OW can be nice, but it feels kinda unrewarding by itself, imo. Though it could work if that exit was much simpler or straightforward to reach than the others... Kinda like Chocolate Island 3, except it's not completely useless.

Yeah, do it, I know what to do on my end to make it work optimally. It should go well with some improvements I've been thinking about as well. The level may look impressive at first glance, but I'm not terribly happy with all the back-and-forth needed for all 3 exits.

>>5667604
Sounds good, thanks!

Sadly my computer crashed right before the final version saved, But I took a picture of the full overworld. Maybe it's for the best, as I am unhappy with how the second half turned out.

So as mentioned, Four to Zero and Fire Forest have not updated me in a month now. As a result, I am making their claims clear. Additionally, I'm going to change my switch palace level as I am not super super proud of it.

Working on some ideas, they may not stick though

>>5672047
What does the ON/OFF switch control here? Layer 2 moving up and down?

>>5672540
That would be correct.

Some more images of my progress. I decided to change it from an on-off level to a smash level.

Bump