AI Recreates the CornellVerse (with John Lions)

[John Lions]

I've played TEW for over half of my life at this point. I've put in countless hours poorly trying to create my own graphics for my personal games. I fell in love with Australia in the CVerse partly because all of the renders were consistent (thank you @sebsplex for your phenomenal work, I would have thousands less hours on TEW if it wasn't for you). ChatGPT+ and DALL-E 3 has given me, along with many of you, the ability to create a ton of graphics for your own personal TEW games, very quickly. I've become addicted to it, and after seeing the amazing work @WalterSobchak has been putting together, I'm inspired to re-render the CornellVerse, one graphic at a time.

For the renders, I have always liked the more cartoon-ish style of sebsplex's renders, rather than the more realistic 3d renders. I'm going to experiment while I try and figure out a style I like, but I'm leaning toward a more illustrated/comic book style of render for the wrestlers. I'm excited because, once I get a render I like for a wrestler, I can save the description and create other graphics around them.

Everything I do is primarily utilizing ChatGPT+ and DALL-E 3 to do the majority of the work. I just toss the image into GIMP and make minor edits to great results. If anyone is interested in the prompts I used for some of the graphics, please don't hesitate to ask.

I'm going to try and create a simple API in ChatGPT to save descriptions associated with wrestlers, and I'll track the success (or lack thereof) to keep things straight. I'm optimistic that if I iterate enough times I can get it to create consistent renders.

If you notice a typo or some word salad in any of the graphics, let me know and I'll see if I can fix it. I try to catch everything but I'm not always successful.

_______________________________________________________

AI CREATES THE CORNELLVERSE (Last Update 12/1 - v0.2):

DOWNLOAD LINK:

https://www.mediafire.com/file/hpqe0l14sz82ci7/AIVerse_0.2.zip/file

LAST UPDATED - 12/01/2023

Contents:

Logos - 100% [Custom promotions - WCW, PPW]
Banners - 100% [Custom promotions - WCW, PPW]
Belts - 5SSW, 21CW, ACPW, APW*, BCG, CGC, CZCW, DIW, FCW, IPW, KCCW, MAW, QAW, QPW, RAW, SWF, TCW, USPW, VWA, WINNOW, ZEN [Custom promotions - WCW, PPW]
Events/TV - 5SSW, 21CW, ACPW, APW, BCG, CGC, CZCW, FCW, IPW, KCCW, MAW, QAW, QPW, SWF, TCW, USPW, VWA, WINNOW, ZEN - [Custom promotions - WCW, PPW]
Broadcasters - NONE
Stables/Dojos - Golden Army - [A ton of custom stables]
People - A ton of random custom ones if people are interested, I'm likely redoing these all last after finishing the above

FIXES FOR NEXT UPDATE:

 

_______________________________________________________

My goal is to recreate every single graphic in the base CornellVerse, as well as creating graphics for events, stables, etc that are blank in the current data. I want to try and keep the original image and design of the promotions, titles, etc as best as I can.

I will be updating this download link weekly. If you have any promotions you want me to work through first, feel free to drop a comment in this thread.

All of these graphics are created with AI, with me editing the vast majority of them. I am trying to balance quantity with quality, so if you notice anything that looks off and fixable (I forgot to cut out some pixels, the text is uneven to a bad degree, etc), TELL ME SO I CAN FIX IT!

If you use GIMP, I included in the download link the XCF files that have all full size logos, titles, events, etc. Use them however you want! Hell, if you see a crappy cut of an event and want to take another swing at, be my guest!

All files that have an image in the default data should have the exact same name as the CVerse file, so you should just be able to replace them (any or all). If there are any where this isn't the case, let me know. For the graphics that aren't in the base data (FCW and ACPW events, for example), I'm not really sure of an easier way to share it - you'll have to go into the editor and set all of those graphics.

_______________________________________________________

CURRENT QUEUE:

AI Rerenders The CornellVerse Progress:

EVENTS:

MAW, DIW, RAW, ?

BELTS:

MAW, Finish APW, ?

_______________________________________________________

USAGE:

Everything I make is free to use for anything. No credit required, though I wouldn't complain an @ if you use my graphics anywhere so I can see how they turned out, or updating in the thread here if you use them for your personal game. 

The exception is anything created for someone else. If the asset was explicitly created for someone else (lavelluk's alt CV22 mod graphics, or HiPlus's Sydney Zone Wrestling dynasty), then if you wish to use these graphics for anything other than a personal game, get permission from the person who I made them for.

_______________________________________________________
REQUESTS?:

I have no plans on doing formal requests, at least until I complete my goal of re-rendering every single graphic in the default CornellVerse. HOWEVER, that does not mean you can't just post in here and ask for something. Do you want me to do a European promotion next? Sure! Do you have a long running personal game and want me to make a new banner/logo for it? Probably, something like that shouldn't take me too long. Do you want me to make 1000 renders for your mod you just thought up? We'll see... just be reasonable with your asks and I'll probably be happy to help.

One exception - content creators, (dynasties, youtube content, etc), I will happily prioritize your requests (within reason). The dynasty boards are what make TEW and GDS great and I want to do everything I can to make the all of your content that much better.

_______________________________________________________

TUTORIALS:

The Absolute Basics of GIMP
Editing Text on a Logo in GIMP (Adding, Removing, Shrinking, Creating)
More Advanced Logo Editing in GIMP
 

_______________________________________________________
LOGOS:

CORNELLVERSE:

CUSTOM PROMOTIONS:

SYDNEY ZONE WRESTLING [ @HiPlus ]

 

ALT CVERSE22 [ @lavelleuk ]

_______________________________________________________
BANNERS:

CORNELLVERSE:

CUSTOM PROMOTIONS:

SYDNEY ZONE WRESTLING [ @HiPlus ]

 

ALT CVERSE22 [ @lavelleuk ]



 

_______________________________________________________
EVENTS/TV:

CORNELLVERSE:

5SSW:

21CW:

ACPW:

APW:

BCG:

CGC:

CZCW:

FCW:

IPW:

KCCW:

MAW:

QAW:

QPW:

SWF:

TCW:

USPW:

VWA:

WINNOW:

ZEN:

CUSTOM PROMOTIONS:

WCW:



PPW:

SYDNEY ZONE WRESTLING [ @HiPlus ]

ALT CVERSE 22 [ @lavelleuk ]

RockStar Wrestling: 



Phoenix Championship Wrestling:



Scandinavian Pro Wrestling:

_______________________________________________________

 TITLES/TROPHIES:

CORNELLVERSE:

5SSW:

21CW:

ACPW:

APW:

BCG:

CGC:

CZCW:

DIW:

FCW:

IPW:

KCCW:

MAW:

QAW:

QPW:

RAW:

SWF:

TCW:

USPW:

VWA:

WINNOW:

ZEN:

RETIRED:

 

CUSTOM PROMOTIONS:

WCW:




PPW:

ALT CVERSE22 [ @lavelleuk ]

SPW (CV2022):

  

 

DOJOS/STABLES:

CORNELLVERSE:

CUSTOM PROMOTIONS:

ALT CVERSE22 [ @lavelleuk ]:

This is the most active I've seen the boards in a long time, and I'm happy to help everyone become self-serving in TEW graphics for their dynasties, personal games, mods, etc.

If anyone has feedback/advice/etc., especially on the worker renders, please feel free! I'm not claiming to be skilled and nothing I use these tools to make is perfect by any means, and I'm always looking for ways to improve.

 

Edited December 2, 2023 by John Lions
12/1 added up to KCCW, updated download link

[John Lions]

_______________________________________________________

RENDERS [BOLD are ones I am thinking of revisiting]:

CORNELLVERSE:

MEN: 

Cameron Saint / Charon (Skyscraper alt) / Captain Wrestling / Adrian McGhee / Brian Rampage

Paul Reinhold / Bryn Archer / Brown Snake / East Coast Panther / Septimus Stubbs

Bounce Buchanan / Flip Buchanan / Bounce / Flip / Hack The Hunter

Davis Wayne Newton / El Hijo Del Neutron / Fenix (Edwin Velasquez masked) / Halloween Knight / HalloWarped

Cyanide / Gareth Case / Leon Nameth / Lerone Slim / Lucas Pobb

Maniac Monster / Underworld / Peril / Pumpkin Jack / Milton Hittlespitz

Suguru Emoto / Reggie Hammer

WOMEN: [* indicates they are from the CVerse Women's Revolution mod]


Spectra (Harper Reid alt)* / Harper Reid* / Christy Higgins / Bonnie Bogan* / Maneater*

Sarah Rioli* / Sonnaz Shah / Elsa Calvo / Yoko Ikina / Chitose Ariwara



Kalliope (Woodchuck) / Golden Delicious / Koala Mask* / Kaia Marshall* / Rei Chikanatsu

Bizarro / Megumi Nakajima (casual) / Megumi Nakajima (wrestler) / Yu Hashimoto / Maki Yoshifumi

Tempest (fka Matilda Jones)* / Tempest (Cosmic Titans alt)* / Tremolo Riff (Cosmic Titans)* / Riana Mete (Tremolo Riff alt)* / Momoe Hamuera*

Thea Davis / Tatum Richards / Kaori Takenouchi / Aki Kamio / Asa Kurata*

Rekha / Melissa Abernathy / Karen Bilous (casual) / Karen Bilous (wrestling) / Stephanie Drucker

Spider Isako / Rika Tsujimura / Kichi Mochizuki / Nami Genda/ Asami Okubo

Steph Blake / Lori / Ceri Dordevich / Super Ceri / Emma Chase

Crusher Ichihara / Jaime Quine / Talia Quinzel / Selina Svelte

 

CUSTOM RENDERS:

MEN: 

WOMEN:

OLD RENDERS:

_______________________________________________________

FULL SIZE GRAPHICS (POSTERS/MERCHANDISE/ETC):

_______________________________________________

Edited November 6, 2023 by John Lions
Changed post to have just renders (hit size limit)

[ShikariFC]

The belts in particular look great! 

[HurkaDurk69]

These are so good, this might be the final push that makes me give chatgpt my phone number lol

[pewaxe]

Personally, I'm a fan of the more illustrated style you're going for. I'm using a similar style, albeit made with StableDiffusion running locally, not DALL-E. I've found that illustrated pictures mesh better, even when the styles are different. The different 3D-rendered styles used in the CVerse clash too much for my liking. Here's a sample of some I've cut out for a custom universe I've been playing as examples. The different styles are noticeable if you look closely, but to me at least, the differences don't jump out like they do with renders.

As a bonus from both an artistic and practical standpoint, you can abandon the classic blue background, leaving the AI-generated background, adding character to the picture while also saving yourself the effort of cutting out the person and pasting them on the blue, although I realize many people on this board are married to that background.

[John Lions]

Thanks for the kind words, everyone. The more I get a hang of what I'm looking for in this style, the better the results I'm getting. I'm starting with my women's division, because there were relatively few women's renders from the GDSverse, so most of the renders I'm currently using are frankenstein creations of a handful of other existing renders...

@pewaxe I feel the exact same way, they aren't the exact same style and if I look really closely I can tell, but at a glance they all look fine as long as I get the proportions right. Call me insane, but I still prefer cutting out on the classic Cverse blue, it has nostalgia value for me I guess.

Here's what I have so far. The right-most "old" render is the original alt I made for my personal game.

Christy Higgins -

Old:  New: 

Koala Mask - 

Old:  New: 

Kaia Marshall - 

Old:  New: 

Yu Hashimoto - 

Old:   New: 

Maki Yoshifumi - 

Old:  New: 

Sarah Rioli -

Old:  New: 

Bonnie Bogan - 

Old:   New: 

 

Edited October 28, 2023 by John Lions

[John Lions]

A couple more renders and resizes tonight. I'm going to put together a GIMP tutorial tomorrow for making simple edits on logo text, and will probably take a break from the renders once I finish out the rest of the women's division.

Resizes/Re-renders:


Christy Higgins (resize), Bonnie Bogan (resize), Yu Hashimoto (re-render).

Maneater - 

Old:    New: 

Megumi Nakajima - 

Old:   New; 

Bizarro - 

Old:   New: 

Harper Reid (+ Spectra alt) - 

Old:  New: 

[StatboyVT]

5 hours ago, John Lions said:

A couple more renders and resizes tonight. I'm going to put together a GIMP tutorial tomorrow for making simple edits on logo text, and will probably take a break from the renders once I finish out the rest of the women's division.
 

Great work! And that tutorial will be greatly appreciated.

[John Lions]

More renders. I'm planning on putting together the GIMP logo guide sometime tomorrow, I want to finish my women's division first now that I started. I will also update the first post with all the women's renders together.

I know it's a fruitless effort to get all of the styles exactly the same. I'm trying to get them close enough that they all look similar at a glance, including trying to get proportions right across every render. If you guys could take the time, could you let me know if any of the renders stand out in a negative way from the bunch? You can be as harsh and critical as you need to, though I may disagree... There's a few I think need to be resized (Harper Reid) and some of the Japanese wrestlers feel a bit out of place still, but I don't know if it's TOO much yet. I'm trying to make sure I keep every render from being too realistic, so I have a bit more flexibility when it comes to art style.

Looking at them together right now, I think I need to try and redo Elsa Calvo a bit les cartoon-y (was tough, content filter was fighting this one hard for some reason). Incredibly impressed with how the ones for Momoe Hamuera, Riana Mete/especially Tremolo Riff turned out, the full sized version is tremendous (just need to change one letter on the jacket):



Elsa Calvo -

Old:    New: 

Golden Delicious - 

Old:   New: 

Sherry Flyman [regen, Golden Delicious's niece] -

Old:   New: 

Logan Flyman [creation of mine - Golden Delicious's son] -

Old:  New: 

Kalliope [formerly Woodchuck] -

Old:  New: 

Karen Bilous -

Old:   New: 

Momoe Hamuera -

Old:  New: 

Tremolo Riff (and alter ego the World Champion surfer Riana Mete) -

Old:  New: 

[John Lions]

BEGINNER'S TUTORIAL TO GIMP
For making your graphical needs 10x easier to meet.

 

WHAT IS GIMP?


https://www.gimp.org/

GIMP is a free, open source image editing program. It can't edit images to the extent Photoshop can, but I think I speak for most of us when I say we don't all have the money to shell out half a grand a year in subscriptions to help you edit logos and other things for a wrestling sim game! Even though all of this tutorial will be framed around editing images in GIMP, a lot of these tools and concepts should work similarly in other programs such as Photoshop.

THE BASICS



There's a lot to any image editing program, especially if you've never opened one before. I will try and cover the absolute basics - if you are a complete beginner and there is something I am missing or you are confused about, PLEASE ASK/COMMENT! I can always add more to this, and I am planning on continuously updating a guide/FAQ like this (and one for generating content you want through ChatGPT/DALL-E 3).

I highlighted the 5 areas I actively use when using GIMP, and I will go into more detail onto each.

___________________________________________

1) The Menu Bar


Most of this is self-explanatory - here's some details on the absolute bare minimum you need to do some simple editing in GIMP.

File -> create new image, save an image, etc.

One important distinction - GIMP "saves" as file type .xcf, which is a file type that will save all the layers that you worked on. If you want to save the image to be shared on the forums or use in TEW, you need to choose "Export As" and type the file type (ex: .jpg)

Layer -> scale, rotate, etc to just the layer you have selected.

Everything that we do will involve manipulating layers so that the final image looks as close to how we want it as possible. The most frequently used by me are:
Transform -> Arbitrary Rotation - allows you to free rotate the layer by clicking (great for rotating text after removing an extra letter)
Scale Layer - self explanatory - scales the layer, so that you can reduce the size for a 150x150 graphic in game, shrink an individual letter that you had to move around to fit, etc.

There are other options under the "Layers" tool bar button, but I find it easier to do some of those through the main Layers toolbar. This might be framed as a "guide" but I am a very casual image editor so this is just how I do things - feel free to play around and find out what works best for you.

___________________________________________

2) The Toolbar

These determine what happens when you click on the image. I (poorly) numbered the tools that are simple to understand, and the ones I use to edit images the most.

1) Move Tool - Self explanatory, with the layer selected, you and click and drag or use the arrow keys to move everything on that layer.

2) Shape Select Tools - You can click and hold to change to rectangle or circle, but doing so will select that area of an image. Any action you do, including cutting, pasting, painting, erasing will only affect the area within the selection.

3) Free Select Tool - The one I use more often. You can click anywhere on the screen to start selecting, or hold to just drag your mouse, to create an area to select.

4) Fuzzy Select/Select by Color - one click on any solid area and it will do it's best to highlight only that color.

Most of what we do will involve using 2) and 3) to select an area of an image, so very important to find a way that works best and get comfortable with it.

After you make a selection, if you were to make another selection with any tool, it will replace the last one. If you hold SHIFT, it will add a "+" sign to the tool, and any selection you do will add to the current one. Similarly, if you hold CONTROL, it will add a "-" sign, and any selection will remove from the existing one.

5) Bucket Fill/Gradient - Fill a layer (or selected area of a layer) with the top color on the Colors toolbar. Gradient is a bit more complicated, but you can do the same thing with a gradient, using both colors on the Colors toolbar.

6) Paintbrush/Pencil/... - Turns your mouse into a brush, and will color on the layer you are on the primary Color on the toolbar.

7) Eraser - Turns your mouse into a brush, and will ERASE the layer you are on.

9) Color Picker - Click on the image to change your primary color to the pixel you clicked from the image.

These tools are there to help you add or remove details of an image, be in a logo/event/stable/render/whatever. All are pretty straighforward.

8,) Clone Tool - Turns your mouse into a brush. The first click you do, while holding CONTROL, sets the reference for what you are going to "clone". Any other left clicks will copy the area from where you first clicked with control - dragging the brush will drag the perspective. You can clone from outside of a selection to inside one. This tool is very useful for cleaning up a background that has a unique texture, one of the only ways I've found to do it.

___________________________________________

3) Colors toolbar

Self explanatory. The top left box is the primary color, the color that it will be when you click to fill or paint. The bottom right color is the secondary color, and you can hit the angled arrows to swap them.

4) Tool Options

Whenever you click on a tool, you will get distinct options for each tool. Each tool has its own list of options which for the most part you don't need to change.

I posted the screenshot of what one of the options looks like for brushes, as it's one you need to know.

You can click on the picture of the brush to change the type of brush you want. I stick with the default, but you could find one that works better for the specific editing you may be doing.

The brush size is very important, as this affects any tool you use that uses a brush (paint brush, clone, eraser). You can manually click to type in the size you want, but you can use the "[" and "]" keys to increase or decrease the size, either clicking or holding.

The Opacity refers to the transparency of the brush when you use it. 100% is default, but if you wanted to add a lighter touch (with eraser, clone brush, etc), you can reduce this to try and blend rough edges together or remove background noise.

There are a lot of other settings, and useful ones at that. Probably some I've never even tried that would make this whole process a lot easier.

5) Layers

(look at the crappy red #s I drew on in the Snipping tool)

You need to understand the Layers dock to do anything beyond incredibly simple edits. When you paste an image from DALL-E or something else into GIMP, it is just 1 "layer" and you will likely need to manipulate layers to make the small changes you need to make your AI image 10x better.

All layers are "stacked" top to bottom, so if one layer has pixels in the same area, it will sit "on top" of all the ones below it. You can click and drag layers in that dock, and right click to do any actions specific to the layer you selected.

1) The "Mode" and "Opacity" determine how the layer you have selected affects the ones below it. By default, it will just "sit on top", but you can change it for some minor edits. Might be difficult to understand it's uses without an example: 

Let's say you have a render where the wrestler has BLACK sunglasses, but you want them to be RED. You can put that render into GIMP, create a new layer on top of the render. Fill that render with a solid RED color, then play around with the opacity and mode until you get the glasses colored how you want while still keeping the original texture. You won't use these very often, but they are good to know how if you need a specific change.

2) Toggle Visibility

Clicking the eye will toggle whether the layer is visible on the screen or not. Useful for checking what's on each layer, how things look with and without the change you made, etc. Whatever layers are visible is the image that will export when you export to JPG, so you can easily create alts or variant versions of a logo or graphic in one file.

3) Layers Options

First, let's talk about that first object in the layers dock, "Floating Selection (Pasted Layer #2)" WHENEVER YOU PASTE SOMETHING FROM SOMEWHERE ELSE (LIKE DALL-E), IT WILL APPEAR IN YOUR GIMP IMAGE LIKE THAT. Before you do ANYTHING, right click on that Floating Selection and choose "To New Layer".

When you right click on a layer, these are the options I use, that I find important:

1.) New Layer - self explanatory, creates a new transparent layer ABOVE the one you right clicked on.

2.) Duplicate Layer - self explanatory

3.) Merge Down - combines the layer you selected with the one directly above it. Very important for working with logos - you will likely have multiple layers of letters, and merging all of those layers together will let you move and manipulate a whole word, or couple of words.

4.) Delete Layer - self explanatory

5.) Scale Layer - Opens a window where you can scale the content on the layer you posted. VERY important - sometimes when copying letters and moving them around, you need to scale them to match (if a logo letter shrinks going from left to right, or grows, for example), and this is the best way to do that.

6.) Add Alpha Channel - Niche, but by default a jpg or other shareable image type will not have an "Alpha Channel". What that means is, if you erase content on that layer, it will turn to your secondary color, not transparent. You may never need this, but good to know if you are trying to create transparent renders or anything of the like.

7.) Merge Visible Layers - Self explanatory. Combines all layers into one, which can be useful if you want to copy the whole image to another GIMP image you have open as one layer.

___________________________________________

Those are the absolute basics, and with just that I can edit most logos that aren't completely FUBAR that come out of an AI generator like DALL-E or ChatGPT. If you have used this stuff before, it all should be either stuff you know or refreshers, but I hope if you've never even edited images before that this can help you get a place to start from. I didn't really go into this with a planned outline and turned into more a stream of consciousness, so please tell me if this is useful and worth continuing.

 

I need your help with the next steps: I think the best way for me to do this, rather than try and come up with examples on my own, is for you guys to post your logos, events, etc, and I can walk through how to edit it. Post some logos below that need some minor edits, and I can walk through ways to do it with those examples - starting with some easier fixes to the more complex ones.

Additionally, I'd like some feedback from you guys - do you find this kind of content helpful? I think AI-created art is here to stay, and I think it can be a very powerful tool for any TEW player to create their own graphics for their own games, without a massive level of time commitment or skill. I have been experimenting with creating more precise prompts with Chat GPT integration to get faster results, and I want to be as transparent as possible so that all of us can keep getting better at this together.

 

 

Edited October 29, 2023 by John Lions

[John Lions]

I hyperlinked the first guide to GIMP to the first page. As I put together more content, I'll add it there as well. Again, please feel free to drop some AI created logos that just need some minor fixes and I can use them as an example to create the guide - otherwise, you'll have to wait until I get tired of doing renders and go back to working on logos...

I think I found a more consistent rendering style and I can maximize ChatGPT to reduce the number of back and forth. I re-did a number of the renders I originally did, as well as re-centered the ones that still fit in with the style. I'm trying to narrow the range between the most and least realistic render of the bunch and keep them as consistent as possible so, at a glance, they all look to be in as close to the same style as possible. I will only include the original renders/inspiration if it's a new worker to the post, and I'll include every render I do from here on out to show the progress.

______________________________________________________

Rei Chikanatsu:

Old;  v1:  New (v2): 

Tatum Richards -

Old:  v1:  New (v2): 

Tempest (fka Matilda Jones) - 

Old:  v1: 

Thea Davis - 

Old:  v1:   New (v2):  

Yoko Ikina - 

Old:  New: 

Megumi Nakajima - 

v1:  New: 

Maki Yoshifumi - 

v1:  New: 

Yu Hashimoto:

v1:  New: 

Sarah Rioli - 

v1:  New: 

Maneater - 

v1:  New: 

Golden Delicious - 

v1:   New: 

Kalliope -

v1:  New: 

Elsa Calvo - 

v1:  New: 

Bizarro -

v1:  New: 

[GnEric]

Awesome stuff, I'm loving that Megumi she is just SO happy.

If you need logos for tutorials you can use this one that I made for my custom promotion. It came out really well and I've made a few tweaks myself but the "E" in "REA" feels a bit off, like the top section is too thick and the the E is a bit too wide.

 

Or the lettering on this one could use a quick fix, the L is kind of squished.

Hope those are useful!

[John Lions]

Thanks for the feedback! Your two examples should be doable, I'll mess around with them sometime tomorrow.

Anyone have any examples of misplaced letters, or wrong words, or letters that got mashed together? Those are the more common botches in my experience, so I'd probably start with the easy fixes first.

One last thing to add for now - I was trying to get alt for Thea Davis, and I ended up getting this image. 10000x better than what I ever expected, even if the lighting might be a bit off for 150x150 render:

[John Lions]

I decided to start with my women's division, because I had pretty limited resources to create my own alts for female workers in the same style as the ones in the main database. That means I could finally put some real images in my mind of what a lot of these workers look like.

Some of these have varying levels of realism, and I might redo some of the ones on the fringes either side. Thankfully, for those of you who aren't familiar with the Women's expansion workers, most of these workers are from the base data!

I decided to go and do my developmental roster as well. I did renders for my first staff members (Tatum kind of counts since she primarily refs), and I think I'll finish my dev company before moving on. Finally, I'm going to try rendering some more male wrestlers...

THE LADIES OF PPW:

Chitose Ariwara

Old:  New: 

Asakura (fka Asa Kurata)

Old:  New: 

Asami Okubo

Old;  New; 

Nami Genda:

Old:  New: 

Jeannie Dreamboat

Old;  New: 

Melissa Abernathy

Old:  New: 

Steph Blake (retired, now an announcer)

Old:  New: 

THE BLOCK

Aki Kamio

Old:  New: 

KAZUMI (random gen worker)

Old:  New: 

Kaori Takenouchi

Old:  New: 

THE SPIDER'S WEB

Rika Tsujimura

Old: New; 

Spider Isako

Old:  New: 

Kichi Mochizuki

Old:  New: 

[hollywoodshinebox]

This is some artist shit, I think I still prefer the 3d, but love these.

Are you taking requests? Because I would love some TCW belts.

[John Lions]

@hollywoodshinebox  I'm not really doing "formal" requests, but if you have anything you'd be interested in seeing, you can just drop it in this thread and I'll get to it when I get to it. I want to eventually re-do the entire Cverse, so you guys would just be adjusting my priorities. I'll take a look at the TCW belts this week.

Thanks again for the praise and feedback. Sharing your method for image editing helped me immensely - I didn't even think to add guidelines in, but as soon as I started roughly adhering to them, my renders look way more consistent together. Appreciate you sharing.

___________________________________________

I think after I do some basic GIMP editing guides, I'll spend some time working on belts and do some tutorials for those. Significantly less work in GIMP or an image editor, but I'll outline how I get the style I want from ChatGPT and maybe have some tips for someone using any other program, or keywords to get the right output (or close enough to).

I'll have the GIMP tutorial for editing logos started later today. For now, here are some of my favorite full size renders from the most recent batches - some of the 150x150s don't do the pictures justice. I have every full-size render saved for all the re-renders I've done, so if you're interested about one feel free to drop it in the thread and I can pull it out. Full-size they may look a little choppy not on a white background -

Aki Kamio

 

Chitose Ariwara

 

Spider Isako

 

 

Edited October 30, 2023 by John Lions

[karlstern]

This is beyond fantastic. 10 out of 10. I can't wait until I see more of the male wrestlers. But so far (as with all the AI stuff in other threads) this is great, great stuff.

[Fleisch]

I love this new style. Cant wait to see what else you come up with.

[John Lions]

TUTORIAL 1: SIMPLE TEXT CLEAN UP ON LOGOS/EVENTS/STABLES/ETC

My good friend @WalterSobchak sent me a TCW logo he was working on that is a perfect example for some of the most common issues with AI-generated logos.
 

"Is this logo savable?"

From my very little experience, most logos are pretty simple to edit once you know how. We have a massive advantage working with the full size version of the logo, as even if your edit is sloppy, shrinking to 150x150 can turn an ugly spot into one wrong pixel. The only times I've found a logo to be unsavable are when:

1) The logo is missing a letter in the font you need it (some exceptions for easy to recreate letters and smaller text)

2) The text is on top of a complex background

There's more nuance to it than that, but those are the most common reasons I can't save a logo design I liked. Thankfully, most of the outputs will include every (or close to every) letter you need, and the text is usually on a solid background.

____________________________________

Looking at the above logo for TCW Threatening Behavior, I see three things that we need to do.

1) Remove Letters - we need to remove the letters that are in the wrong spot on the logo

2) Move Letters - we need to move letters to the correct spot on the logo

3) Add Letter(s) - We need to add a letter to the logo, that already exists on the logo in the same font (on this graphic, it's the letter "E" in Threatening - we need to add a second one)

4) Create Letter(s) - We need to create a letter in a font that doesn't currently exist (on this graphic, it's the letter "O" in Behavior")

5) Adjust Size of Letter(s) - We need to adjust the size of some of the letters in the word "Threatening" so it fits the same width when we add 1 extra letter in.

This may seem like a lot, but these are all very easy once you know how to do them. For context, fixing this logo took me around 5-10 minutes, and once you get more comfortable using GIMP, you can do it that fast as well.

Reminder that these are all just ways that I found easiest to do this - there might be other ways to do this even easier and faster, and if you are reading and know something feel free to comment! This is just as much a guide to new GIMP users and image editors as a way for you guys to help me (and by extent the rest of us) get better at self-serving graphics for our niche video games + beyond.

Let's walk through each of these steps:

____________________________________

INITIAL SET UP

First things first - we need to get the above image into GIMP. There are a few different ways you can do this. You can copy and paste directly into GIMP with no images open, and it will create it, or you can create a new blank image. The standard size for logos that come out of DALL-E/ChatGPT are 1024 pixels x 1024 pixels, but if you get a portrait or landcape picture the largest it would be is 1792 x 1792.

Let's start by pasting the above TCW logo into GIMP.

 

 

Before I do ANYTHING, I am going to make a new layer beneath the TCW logo. To save time, I'm going to make it the same color as the background of the text I want to change.

• Right click on the layer with the Logo, and choose "Add Alpha Channel" (if needed)
• Right click on Layer -> "New Layer" and hit OK
• Drag the new layer under the Logo
• Click the Eye Dropper tool and click the background behind the letters
• Click the Fill Bucket tool, and click anywhere on the image (MAKE SURE YOU HAVE THE NEW LAYER SELECTED.

 

REMOVING LETTERS

First, let's learn how to remove letters. Looking at the above image, there is one letter that needs to be removed: The second "R" in Behavior is not needed.

We can use the R to help us later, so I won't be deleting it entirely. But this is how you do it:

• Make sure you have the "Logo" layer chosen in the Layers tab.
• Pick whatever select tool you like the most (free select lasso or rectangle work best). I will be using the rectangle, since it looks pretty painless to do here.
• Select the letter you want to remove.
• Press "DELETE" on your keyboard, and it will delete the letter. (I will be doing something different, which I will get into later).

Follow those steps and you will have successfully removed the letter that doesn't belong. If the background doesn't match exactly with the new background you put in, there are some creative ways to get there, but we'll focus on the simple way for now.

 

MOVING LETTERS

Next, lets get to the fix you will likely end up doing the most - moving letters to where they belong in a word. Looking at the above image, we have two that we have to do - swapping the R and E in Threatening, and swapping the A and H in Behavior. I'll walk through the steps for "Threatening", as the steps will be exactly the same for Behavior:

• Make sure you have the "Logo" layer chosen in the Layers tab.
• Pick whatever select tool you like the most (free select lasso or rectangle work best). I will be using the rectangle, since it looks pretty painless to do here.
• Select the first letter you need to move.
• Press "CONTROL-X" on your keyboard. This will cut the letter out - making a copy and deleting it from the layer it's on.
• Press "CONTROL-V" on your keyboard. This will paste the letter you just cut out to a NEW LAYER - you will need to right click on this floating selection and send it to a New Layer before you can do anything else!
• Use the Move tool to move the letter to where it needs to be.

Always make sure that when you are cutting (ctrl-x), you have the layer selected that you are going to be cutting from.

For this logo, I need to cut out two letters (R and E, A and H), add them to new layers, then move them to the correct position. You don't have to name your layers every time, but it can be helpful when you are learning to start.

ADDING LETTERS

Sometimes, the logo you have is missing a letter in the word, but that letter exists on your logo in the same font. Don't worry, we can fix this too. In the above example, we need to add an extra "E" to "Threatening". The steps for this will be a bit more complicated and use some of the skills shown in the above 2 fixes, so I will be short handing some of the steps.

• Make sure you have a layer with the letter you need to add. In this example, we need to add an "E" and we already have an E in the "Threatening" font and style, so it saves us some work.
• (If you did not have that, you would do the steps from the previous fix to create a new layer with just the letter you want to add)
• Right click and Duplicate the layer with the letter you want to add.
• Move the new letter to the exact place on the logo you want it.

This is where things get more complicated. What we need to do now it move "THREAT" before that new E, but all of the letters for "THREAT" are on different layers! the next step we need to do is combine the string of letters we need to move into 1 layer, above the logo.

• Move ALL LAYERS that make up the string you need to move directly above each other. This might sound a bit confusing, but makes more sense when you see the next step. In this case, I need to move the first "E" and the "R" layer down above the base layer (which is where the rest of the letters in "THREAT" are.
• Click the eye next to the letter you are adding's layer so that you don't see it for the time being. This is just so you can get a clear picture of the string you want to snip + move.

• Starting with the top-most layer, right click and choose MERGE DOWN. We want one layer where all of the letters we want to move are located. In this example, I am merging E and R down into Logo.
• Use the selection tool of your choice to select "THREAT" from the merged layer. I am using the free lasso, as I can click and point with more precision, and it's not exactly a rectangle. Try and select as close to the letters as possible, to avoid cutting out parts of the background that are more than just black (that would be much harder to fix after!)
• Cut (ctrl-x) your selection and paste (ctrl-v) to a new layer (right click the floating selection and make it a new layer after pasting).

• Click the "eye" next to the letter you are adding (E) to show it.
• Move the old text layer accordingly to make it all look consistent.

It's not 100% perfect and lined up, but when it's shrunk to 150x150 any minor inconsistencies become fractions of pixels. Sometimes, this is all you will have to do, but adding letters often upsets the spacing between other words or the logo itself. The "Threatening" in the logo above may look passable, but if you look at it closely, you can see that it's now uneven in the center of the logo, and the T stick out over the background graphic of the logo. I'll go back to that last, but first let us make ourselves an "O" using a bit of creativity...

CREATING LETTERS

Sometimes, the logo you have is missing a letter in the word, and that letter is no where in the logo at all! If you know the exact font and style, you could use a text tool to type it and recreate it. I don't, and dealing with text in GIMP is a horrible experience, so we are going to use some cheeky layers and borrow parts from other letters to build the letter we need. This doesn't work all the time, especially if the text itself has a unique pattern to it, but it can work under certain circumstances. For this example, we need to create a letter "O" in Behavior.

• Identify letters in the same font you can use to "Frankenstein" the letter O. This is all practice and judgement calls, as there's a ton of ways to do it. For this example, I think if I take the previously-deleted "R" and take the top right corner of that letter, I can duplicate it to create a Frankenstein'd "O".
• Cut, flip, rotate, etc as many different parts as you need to create the letter you want. This is very vague, and can probably be a whole guide on it's own, but at the end of the day it's a judgement call on the "how" - pick whatever makes the most sense in your head. I will include the exact steps I used to recreate the "O" as an example of how to do it:
  • Remember that "R" we were supposed to delete, that I cut and saved for later? Unhide that layer and bring it back above the "Logo" layer.
  • Use a selection tool to select the top half of "R" - right before the right side starts to curve in on the right, and right before the end of the inner circle on the left (see screenshot below).
  • COPY (CTRL-C) the selection (DO NOT CUT!), paste as new layer.
  • We want to flip this horizontally, so choose Layer -> Transform -> Flip Horizontally.
  • Move the floating top corner of our O to the top left, so the top of the letter now looks like an "O". Merge down.
  • Select the now top-half of our frankenstein'd "O" (see second screenshot).
  • Copy the selection, paste as new layer.
  • We want to flip this vertically, so choose Layer -> Transform -> Flip Vertically.
  • Move the bottom part of the O around until you are content with the orientation.
  • Last, clean up the bad pixels. In this case, inside the O and around the letter there are some mess, you can just identify the layer they are on, select that area, and DELETE it. There are a ton of other ways to do that same thing as well.1) 

Now we're cooking. One last change to make before we can feel content with our new logo:

ADJUSTING THE SIZE OF LETTERS

This might not always work, and might not always look the cleanest, but this is the easiest way I've found to "squeeze" the letters back together once you fit one or two more in. After adding the E in "Threatening" in our example, we need to make sure the word is evenly aligned where it was on the first logo. Thankfully, some of the work we did before will make these steps a bit easier.

• Identify the string of letters you need to shrink down, and get it all on the same layer. In this case, we have that already with "THREAT".
• With that layer selected, right click on the layer dock and choose the option "Scale Layer".

• Click on the pixel-lock icon to the right of the boxes with width and height. By default, when you scale a layer it will keep the width and height in proportion of each other, but we don't want that.
• Reduce the width, taking out the width of the letter you added. Honestly, there's easy ways to get that number, but I usually just eyeball it. For the sake of the guide, I'll walk through how to get an exact ballpark. We know we added the "E", so you can click on the "E" layer, right click and hit "scale layer" to see exactly what the width of the E is right now. On my file, it's 28 pixels, so I am going to reduce the width of "THREAT" by 28 pixels, from 201 to 173.
• Keep trying until it's the right size - when you have it right, move it back into position.

Before and After:

 

You can see that it's not perfect - "ENING" is slightly wider, and you could go through the effort of doing the same steps to shrink before and after the E by half as much size... but do you need to go that far to make it perfect? I can see cases where you would, but most of the time I just need it "good enough" for the 150x150.

 

FINAL RESULT:

 

______________________________________________________________

Please let me know if you find this useful. It was a bit of effort to put together, but I'm happy to continue to do more of these to cover most of the examples of errors and how you can fix them on your own. This might look daunting but again I emphasize once you have done this a few times, it will take no time at all. Feel free to walk through these steps on your own logos, or use this one as an example.

Thanks again @WalterSobchak for providing a perfect first image to edit, and @GnEric I will see what I can do with yours later this week!

Edited October 30, 2023 by John Lions
"final results" were the wrong version

[WalterSobchak]

This is awesome! Great work JL!

[John Lions]

Will likely dump the rest of the renders tonight, but I wanted to share this one. I'm a bit concerned they are a bit "too" cartoony, but man I love how the Buchanan twins turned out:

 

Bounce and Flip in the default data (with some alts I've made)

The Buchanans

 

The Blue Cannons ("super hero" alter ego) (w/"Super Ceri" Ceri Dordevich)

Edited October 30, 2023 by John Lions

[GnEric]

Excellent guide! I've used GIMP a lot over the years for various minor things but never for edits like this. Seeing the step by step process is really helpful. Thanks for putting this together.

[John Lions]

I've realized at a point with the renders I've been doing, I was comparing them to the last render I had made rather than them all together. Looking back, I realized the style I've been trying for got out of whack, so I reorganized every since one I did into categories, sorted on a scale of "Too Realistic" to "Not Realistic Enough". Right now, I'm thinking of redoing the ones that fall outside by "bounds" and then make sure I have the most and least "realistic" of the renders to compare and make sure anything new falls within that spectrum.

If I was actually an artist making these, they'd obviously all be the same style, but instead I am relying on a robot and my own eye and input.

Take a look at how I broke them up below and let me know what you think. Does my method seem sound, or is there anything I'm missing?

TOO REALISTIC (planning to re-do)

__________________________________________________________

MOST REALISTIC (keep, but are borderline)

SWEET SPOT

 

LEAST REALISTIC (keep, but are borderline)

__________________________________________________________

TOO HAND DRAWN (planning to re-do)

WRONG STYLE (planning to re-do)

WRONG LIGHTING (planning to re-do)

____________________________________________________________

Unfortunate for the Buchanans, but we'll see them again... I can see where I "lost the plot", and think I'll have a much better time using this as a guide. I want to always make sure the two renders at the opposite ends of the "spectrum" look ok next to each other.

Examples:

Most to Least Realistic

OLD:
 vs 

Definitely no-go...

NEW:

 vs 

MUCH better. I think Maneater and Bizarro can get away with it by having the larger than life gimmicks. I think I can relight Asa Kurata a bit better and do a better job cutting out the hair so it doesn't have that weird "glow" effect. Kaia Marshall and Tempest are borderline.

 I think I can deal with this

 

... I'm not sure.

 

I think I'll give Kaia and Tempest another run through and see if I can get anything closer. Bizarro I think is going to have to be the bottom bar because getting her rendered was more difficult than probably a quarter of the rest of these combined...

As always, any feedback helps. Should I be more or less lenient? From my experience, I seem to be a bit limited into what scope of image I can get by nationalities (you can see I was pretty spot on and consistent with the Japanese ladies, but I can't get a different race in a similar style), but I think a little bit of style diversity is fine. I don't think I can restrict it TOO much without making this herculean task an impossible one, but I think there is always more iterating I can do to make the process better.

Edited October 31, 2023 by John Lions

[John Lions]

So, I created an API to use in ChatGPT+ to make creating renders easier, with /commands and clear instructions. I was having an issue where it was taking descriptions of things that didn't matter (like name) and sending that to DALL-E,. Feel free to give it a try yourself - I only just started using it, but I've been able to get very consistent renders this way. You can change the section for "Image Style" to whatever style you want the renders to be, or leave it up to your own input!

 

Quote

Your current programming is set to “RenderMaker 1.0”. You will assist me with creating renders, graphical representations of professional wrestlers in my game, in the style and presentation that I want.

Commands for User Interaction

/create: Start the process of creating a render. /create WRESTLERNAME will initiate the creation of WRESTLERNAME. All data will be saved for each wrestler.

/save: This will save the image (the prompt AND the seed) to the wrestler name, or to the wrestler name + alt.

/alt: Create a new picture, starting with the prompt and seed of the wrestler with added instructions. (Example: /alt Thea Davis + Pirate would initiate the creation of a pirate alt for Thea Davis, using the previously saved prompt and seed).

/fix: Instructions for how to adjust the input creating an image you created previously.

/retry: recreate the last images and prompts with different seeds.

/reset: reset this wrestler creation and forget all inputs.

/modify: identify reason an image couldn’t be created (whether content policy or too complex instructions). Output the reason (or what you believe to be the reason), change a prompt, and generate new images.

Creation Protocol

Any input in between ** are important details that need to be directly in the prompt sent to DALL-E.

Any input between “” is a conversational description to you about the wrestler. You (Sophie) will take all data relevant within the “” to the creation of the wrestler. (Example: in a creation step, I put “Hack The Hunter is a lovable, high energy babyface, who’s gimmick is an Outback Hunter”, and you put into the ChatGPT prompt “big smile, dressed like Outbacker,friendly face”).

If you notice I get stuck, ask questions and provide the current state of the wrestler creation input, as well as offer some suggestions.

Image Style

All prompts must start with this baseline for the STYLE of the render. All renders must be in an illustrated style, looking at camera, and rendered to chest. All proportions of art must be realistic, as well as the colors. Simplify the baseline style for the prompt, and notify if any instruction would yield a result outside of the current style. 

___________________________________________________________________

After struggling for a while, I was able to get these both of these very quickly with minimal back and forth after creating the API above.

Primitive Power


Hack The Hunter and Bryn Archer

Edited October 31, 2023 by John Lions

[drmacbeth]

17 minutes ago, John Lions said:

So, I created an API to use in ChatGPT+ to make creating renders easier, with /commands and clear instructions. I was having an issue where it was taking descriptions of things that didn't matter (like name) and sending that to DALL-E,. Feel free to give it a try yourself - I only just started using it, but I've been able to get very consistent renders this way. You can change the section for "Image Style" to whatever style you want the renders to be, or leave it up to your own input!

 

___________________________________________________________________

After struggling for a while, I was able to get these both of these very quickly with minimal back and forth after creating the API above.

Primitive Power


Hack The Hunter and Bryn Archer

Do you find adding "render to chest" helps limit too much extra body in the image?

 

By the way, loving all of these and hope you get through everything you plan.  I will say the one downside to this style, though, is that it will be an all-or-nothing situation.  I think it would be hard to mix these in with any other images - AI or renders.