Sonic Generations Hacking (and More!)

All this talk of modelling reminds me of a major part of why I chose to port retro 2D Sonic levels into SG (VERY old stuff, still need to make an update video) rather than pre-existing 3D levels, or even take a stab at designing original stages. It's taught me a hell of a lot about modelling techniques for games whilst still working off of a pre-existing 2D design, keeping it simple. It's also taught me that S1's GHZ just doesn't work perfectly when visualized in 3D due to the layered system, which amusingly enough is ONLY done that way in GHZ, so I've had to take some artistic liberties to achieve a faithful port. I've started over a couple times too, with increasingly speedy iterations, whilst managing to poly-count down by almost 90% (largely the result of forgetting to fiddle with spline interpolation, and also because I'm a silly :v, as well as streamlining techniques for UV unwrapping complex objects (so that grass-edges unfold neatly and uniformly).

As I've been going, I've been compiling some documentation on my techniques, partly to keep myself right about what I'm doing if I have to take a hiatus, but mainly to support another large part of why I'm doing this project, which is to create something of a tutorial; once I've nailed down my best-practices, I'll be able to give back to the community with my modelling, unwrapping and optimization techniques in both video form and text-form.

I completely agree with the difficulty and hesitation in coming up with designs and concept pieces. My suggestion, taking inspiration from what's worked for me in the past with stuff like that, is to use the already-available Wiki pages on themed Sonic levels and take inspiration from already-existing assets, and change them up in combination with a random word, phrase, idea, or wildly different setting that could inspire you for original design ideas.

For example, off the top of my head, "Imagine a design like Sky Rail, which is largely focused on grinding and vertical movement, but set in an underwater facility" What would that look like? What's the story of the level? Are you escaping from the deepest depths? Are you infiltrating? Or are you maybe just exploring the ocean? I like the approach S3&K took, in that every stage felt like it had a purpose to the overall story; the levels were interconnected and transitioned between themselves, the island felt coherent, and the story was evident by the level design alone. I miss that sort of design philosophy; I feel like it's been lost into cutscenes and dialogue.

Anyway, back to the underwater Sky Rail stage concept. Already from that concept, you're probably imagining interesting tubes, coral reefs, expansive underground caverns, maybe a small island section. There's a lot of ways to get inspired, and the easiest way to start is by absorbing, imitating and then iterating. Just remember that the visuals aren't nearly as important as the core theme, gimmicks and settings, so get those nailed down and you can begin to brainstorm about the assets and functionality.

Once you've nailed the stage's purpose and function, you can start searching the web for related images to inspire you for what sort of assets you can make. Just compile a big folder of related images, using as many keywords and synonyms as you can muster. Then it's a case of turning them into fun and functional assets. Drawing viewport references is tough, but you can kind of cheat if you have actual reference images to work from. Typically, you're likely working on your own, so you don't need the reference drawings to be super high quality, just do as much as you personally feel you need to get the correct silhouette going, and then iterate in 3D where needed.

Anyway, I could go on, but you're all pretty competant people, so I'm sure you can figure out where to go from there

 

The problem with making assets is either spreading too thin or not having enough variety.

It still boils back down to someone having a blueprint for everything beforehand.

1 - Concept - concept artwork, badniks, gimmicks, overall aesthetic
2 - Assets - Modelling, texturing, facading, sound and music work
3 - Level design & implementation - Stage design, playtesting, optimization

People who jump into this even as a hobby needs to realize that all three are very distinct fields. #2 and #3 can be done simultaneously for any one project, but #1 needs to be very clear cut or you risk wasting a lot of effort into things that ultimately go unused, or on the flipside find yourself lacking necessary assets to realize the level design.

For porting or remake projects (that aren't direct conversions of e.g. Unleashed data) #1 is already pretty much all there, and bits #2 and #3 can be cannibalized depending on what the original level is from. It's when someone tries to come up with a completely new stage concept then attempts to do all three himself that you see the inevitable burnout or substandard* level mods. My suggestion is, get a team that covers all three bases, then get to work. Otherwise the 'scene' is going to be pretty much stuck in a perpetual bell-the-cat situation till it eventually fizzles out.

*The effort in some projects are admirable, don't get me wrong, but Unleashed and Generations both set a pretty high bar in terms of expectation.

 

I still wonder if it's possible to get Super Sonic playable in bosses, not a model swap, but the real deal.

 

You mean, like this?

 

Yes, just like that. The guy never said how he did it..

 

Someone correct me if I'm wrong (as I am not very well versed in the inner-workings of Generations), but shouldn't his ring count be decreasing if its the 'real deal'? Unless he just hacked it to not do so or something (again, I really don't know how possible that even is).

 

I only recently got posting privileges on a temporary basis and I would just like to say thank you to everyone for making this game better than it already was and thanks for documenting it all here because it's been interesting to read when I get the time.

Edit: I don't see anyone using custom voice over and frankly I don't see why you'd need it but it's really all I can do and it feels wrong not to offer anything I have professional equipment and I'm starting out with a few jobs in indie games/books so if you do need voice I will provide it for free.

 

+1 from Elliot Gindi. I reach each post in this thread and it's really informative and people have some really inspiring stuff to say (such as all the recent 3D modelling motivational talk). Hopefully someday I'll get into the generations scene whenever I get a slighty faster computer :P

 

Doesn't Sonic have an official height by the time of Generations? Because then we should be able to, using stage models, get an accurate scale of how much actual distance is covered, and how long it takes for Sonic to do so. We could easily get speed numbers that way, no?

 

Yeah, I have the same problem. My computer can't run Generations, so I have to miss out on all the hacking fun.

 

Sonic's model is almost exactly 1 unit tall (just measured), and I remember either Zoney or Paraxade saying that the model gets scaled ingame (height wise) by a factor of 1.25 or so.

It's been awhile but I think Sonic's max jump height is around 4.5-5 units tall.

 

Yeah, it's set in the XMLs. IIRC his maximum velocity without boosting is 70. His boost speed is just a multiplier; it's 1.5x his run speed, which comes to 105. Not sure 70/105 what, though.

 

It bothers me how all the theories completely ignore that Sonic probably isn't going his real top speed in the games they're using for reference. If Sonic ran at the speed of sound all the time, we wouldn't be able to control him as a platforming hero. We'd be controlling him as a drag racer.

 

Exactly. According to people like him, everyone in Kanto walks on a grid and have massive heads that shrink when they get into fights. It's intentionally messed with for the sake of the player having a better time. Does music just happen to follow Sonic around and change depending on where he is?

Sonic's greatest speed... In the classic games without the speed caps, if you find some good terrain, he can go a whole lot faster than just holding-right-to-win on a flat plain. Even in Sonic 1, he could go far faster than his own running speed by "rolling around at the speed of sound" in Star Light Zone. Personally, I attribute his characteristic high speed to his rolling and/or his environments. Not his capped running, which is the only form of movement in the newer games. If there isn't another speed limit in the code that I've never noticed, he can keep gaining speed until his inertia craps out at $FFFF.

 

Actually, since the speed values are signed, the fastest he can go to the right is actually $7FFF, and the fastest he can go to the left is $8000. Any faster and he'll switch directions. Furthermore, the speed value is a fixed point value, with 8 bits for the pixel and 8 bits for the subpixel, so it's actually 127.99609375 pixels/frame right and 128.0 pixels/frame left.

 

Huh, I thought when inertia was described as "directionless" that it needn't be signed. At least, that's what the S2 disasm says.