VN Ren'Py Ruin Me [v0.61.1] [Phanes]

[Deleted member 5143451]

Background are very artistic, but the characters in front, totally don't fit, and are just ugly to look at. But that's just my opinion.

 

[PhanesDionysos]

What if you TEMPORARILY used AI generated images for the scenes?
Something like NovelAI is fully capable of generating consistent images for most part. As lazy as it sounds, it may be a lot more appealing than the current art.

AI art generation is glorified tracing: all the most-hyped GANs are using a corpus of other peoples' art/photography and just blending them together. While I don't have anything against collage, I try to only use art resources in my game that I have permission for, crediting when appropriate. That's impossible to do with most AI art generators, and I'm not inclined to try and track down one with a free-and-open corpus.

Also: using AI art would still involve a bunch of learning and work and tuning, plus overdrawing all the places that AI has trouble with, like hands and hair and fabric and teeth and lighting. Just the thought of getting consistent character models in different poses out of a GAN fills me with dread.

Plus, all the time spent generating and cleaning up AI art would be time when I wasn't improving my skills at doing my own art. I know that a lot of folks still don't care for my character drawing and I can see the flaws in it, but you can see a huge difference between

You must be registered to see the links

and

You must be registered to see the links

. If I'd somehow used NovelAI from the start, I'd still be using it and my art skills would have barely grown.

Any of the available software suites (that you would power with your local BEAST rig, if you have one) and online services (that run these ai networks on their servers) should work wonders on items, buildings and background. Which this game needs desperately.

This bit is surprising to me! I haven't seen much criticism of my (admittedly low-effort) backgrounds, and I'm interested to hear what turns you off about them. My backgrounds are mostly public domain or free licensed photographs passed through some filters (i.e. tracing) and I've always considered them perfectly fine set dressing in the style of various dating sims that use a similar technique. Can you speak more about what bugs you about them? You're welcome to DM me so the thread doesn't become a mess.

Background are very artistic, but the characters in front, totally don't fit, and are just ugly to look at. But that's just my opinion.

This is the sort of opinion I've come to expect, and for now I'm inclined to hope that other aspects of the game help to balance out any art issues.

That said, anyone who has an art background and is willing to give me specific criticism about systemic flaws (I'm already practicing fundamentals and studying anatomy and such), please DM me! I'm very interested in ways in which I can improve, especially if there's something I can fix more easily than just building general skill.

There are plenty of smutty games I love that have seriously wonky art, and I'm curious to know what it is about my art that gets a reaction that those games don't get. That said, I'd rather this thread not be a public litany of my flaws, so I'd appreciate tips in private.

 

[Fluxcarnal]

Just to give a bit of a different view, I quite like the art. It also seems fitting for the setting and atmosphere in a way that any soulless daz renders never could.

 

[tioafa4]

I don't want to be rude to the ones who have suggested this, so apologies in advance if it comes off that way, but I would absolutely hate AI art being used for this game. I mostly dislike renders in the first place; I usually find them off-putting and the only game which has renders that I really like (Heavy Five) has seen its development time balloon to absurd proportions because of the time and resources the dev has put in getting them up to that level of quality. Ruin Me's art is so obviously better and more appropriate for the game than renders or AI art would be that I was taken a little aback by the suggestion that it's somehow objectively bad or wrong for the game (like. objectively? What is the objective measurement being used here, exactly?). The art was OK to start with and has gotten quite good for most of the main characters, especially Theresa. Even the well-made renders done by a fan on this thread a while ago don't appeal to me as much as what we have in the game as it stands. AI art's only advantage for this game would be its ability to effortlessly conjure up nightmares.

 

[spikesagitta]

I actually find the art quite charming. I always like games with drawn art, instead of 3D renders.

 

[sadiyaflux]

This bit is surprising to me! I haven't seen much criticism of my (admittedly low-effort) backgrounds, and I'm interested to hear what turns you off about them.

I was not only talking about backgrounds, it's more about the entirely missing item artwork. And sure, new Backgrounds would be nice as well. The "tracing" you speak of looks to me just like a simple Photoshop Filter, this is also in the realm of personal preference. I think the current ones serve their purpose well.

The issue raised by you is a real one - and I missed this earlier. Yes, these neural networks are trained to "copy" a specific art style. However, you can determine what they are doing - to some degree. This requires more in-depth knowledge, but it's certainly doable with self-hosted networks. Website services will try everything to make a profit from this - obviously.
But it is by no means decided or made into law anywhere (to my knowledge) what these images should be, what they are and what they are not. But once these networks can mix and match styles, this debate will flare up again, I'm sure. Corridor Digital has a few nice opinion pieces on youtube about this. I can absolutely understand the issue here.

But I see you're very adamant about just using your drawn artwork. And sure, this game must be some kind of passion project - so I think I can relate. It's entirely your choice, obviously. And yes, you have made big advances in drawing character artwork, at least in my eyes (and have stated this before, I think). But the original, old artworks are still present nonetheless =)

At this point it seems like this kind of critique falls on somewhat deaf ears. If you cannot see the issue here - I can't really make you see it, y'know?

You do you. I thoroughly enjoyed what is there (Played a build from two or three (?) months ago) and I want nothing more than to buy the full version, content & feature complete, with the endings you envisioned. Just to escape into this wonderfully perverse, and weird story until the morning comes =)

@spikesagitta and @tioafa4

We're not talking about 3D rendering. The current topic is image generation with AI networks. In short, it is a collection of programs that can "learn" how to draw images, in 2D (it's artwork - mostly). Now, there is an ocean of more information on this topic here - so this is an extremely simplified version of it ^^
The gist of it is certainly not the dreaded 3d rendering you saw in other games. Image generation can give you (if you know what you are doing, and spend the time to train the networks further) a sexy vampire lady portrait in the style of any artist you can think of. In minutes, for "free". Is it really "artwork"? Cannot say.

Using this technology, one way or another, would ultimately be a filler to speed up development. I would at least consider it in earnest for the items and the backgrounds, regardless of what is already there.

Have gr8 day everyone!

 

[tioafa4]

I was not only talking about backgrounds, it's more about the entirely missing item artwork. And sure, new Backgrounds would be nice as well. The "tracing" you speak of looks to me just like a simple Photoshop Filter, this is also in the realm of personal preference. I think the current ones serve their purpose well.

The issue raised by you is a real one - and I missed this earlier. Yes, these neural networks are trained to "copy" a specific art style. However, you can determine what they are doing - to some degree. This requires more in-depth knowledge, but it's certainly doable with self-hosted networks. Website services will try everything to make a profit from this - obviously.
But it is by no means decided or made into law anywhere (to my knowledge) what these images should be, what they are and what they are not. But once these networks can mix and match styles, this debate will flare up again, I'm sure. Corridor Digital has a few nice opinion pieces on youtube about this. I can absolutely understand the issue here.

But I see you're very adamant about just using your drawn artwork. And sure, this game must be some kind of passion project - so I think I can relate. It's entirely your choice, obviously. And yes, you have made big advances in drawing character artwork, at least in my eyes (and have stated this before, I think). But the original, old artworks are still present nonetheless =)

At this point it seems like this kind of critique falls on somewhat deaf ears. If you cannot see the issue here - I can't really make you see it, y'know?

You do you. I thoroughly enjoyed what is there (Played a build from two or three (?) months ago) and I want nothing more than to buy the full version, content & feature complete, with the endings you envisioned. Just to escape into this wonderfully perverse, and weird story until the morning comes =)

@spikesagitta and @tioafa4

We're not talking about 3D rendering. The current topic is image generation with AI networks. In short, it is a collection of programs that can "learn" how to draw images, in 2D (it's artwork - mostly). Now, there is an ocean of more information on this topic here - so this is an extremely simplified version of it ^^
The gist of it is certainly not the dreaded 3d rendering you saw in other games. Image generation can give you (if you know what you are doing, and spend the time to train the networks further) a sexy vampire lady portrait in the style of any artist you can think of. In minutes, for "free". Is it really "artwork"? Cannot say.

Using this technology, one way or another, would ultimately be a filler to speed up development. I would at least consider it in earnest for the items and the backgrounds, regardless of what is already there.

Have gr8 day everyone!

I don't know very much about AI image generation, but I think i get the gist of it? It would be pretty interesting to see a neural net trained only on one person's art. You could make it generate images and see if someone unfamiliar with the artist can spot the fake. Anyway, the ethics are troubling (art should have to be licensed from artists in order for commercially-used AI art to be trained on it, as those artists are effectively acting as the AI's art tutors), but I can't say that I'm not interested in the idea of affordable asset production for those of us who can't draw at all.

 

[PyoT]

AI art is very impressive but I also feel that it's vastly understated how much you need to generate to get something consistent and that's not, say, got mangled ears, six fingers, a missing hand, etc. Too much pretense that you're just "this is what I want" and then you get it ... it's not like that, and personally I very much doubt we'll get over those "small issues" any time soon.

I don't think it's much of an option except for messy placeholder art (which this game doesn't need) and for somebody who then spends a ton of time on fixing problems (and clearly if you do that you suddenly aren't saving time anymore).

 

[sadiyaflux]

AI art is very impressive but I also feel that it's vastly understated how much you need to generate to get something consistent and that's not, say, got mangled ears, six fingers, a missing hand, etc. Too much pretense that you're just "this is what I want" and then you get it ... it's not like that, and personally I very much doubt we'll get over those "small issues" any time soon.

I don't think it's much of an option except for messy placeholder art (which this game doesn't need) and for somebody who then spends a ton of time on fixing problems (and clearly if you do that you suddenly aren't saving time anymore).

The models those website-services use have been trained extensively, hundo per cento. That is the reason why we see so much activity in this space as of late - there have been some exciting (open source aka free) releases with pre-trained models. For example Stable Diffusion or DALL-E (example:

You must be registered to see the links

or

You must be registered to see the links

), both are text-to-image workflows. The "small issues" have already been resolved. The problem with what you're describing is real, don't get me wrong, but with improvements in the software itself AND extensive training models, these are already very much not an issue anymore. Training always will be the limiting factor, not generation.

It all depends on what the software knows well and what it can understand. Honing in on one area might get you where you want to go. Especially if you set it up yourself (which is, granted, a pain in the... well, Theresa understands) and then focus on one training area with this one model. With AI generation you will always see these rogue, weird-ass pictures. The software itself doesn't know what it is creating, it just learns to attach forms, shades and structures to words and then composes them into one final image, y'know?

But enough of that from me ^-^

PS: That took like... 3 minutes, on a website, for free (however, they want credits preeeeetty fast, especially for higher res images) - it's just an example what a well trained network can do. They call this "style transfer". All the issues and problem apply, again - just an low effort example.

 

[PhanesDionysos]

It would be pretty interesting to see a neural net trained only on one person's art. You could make it generate images and see if someone unfamiliar with the artist can spot the fake. Anyway, the ethics are troubling (art should have to be licensed from artists in order for commercially-used AI art to be trained on it, as those artists are effectively acting as the AI's art tutors), but I can't say that I'm not interested in the idea of affordable asset production for those of us who can't draw at all.

Yeah, as someone with a computer science background, there's all sorts of really cool possibilities and potential for neural-net image processing! One purpose that's already good and getting even better is doing tasks like noise removal and upscaling; you can get super high-quality results and I'm much less concerned about the ethics of training that sort of software on a corpus that's not entirely freely-licensed.

I'd love to see more good neural net trained on CC-0 and public domain works only! Unfortunately, I bet that the money tradeoff isn't there for the really fancy software; it's a far more compelling product for companies to say "you can make something that looks like cool art that you'd otherwise have to pay to commission!" rather than "you can make something that looks like a 17th-century Flemish painter!"

It's not what we're talking about here, but the shit that especially bothers me is when someone opens commissions or a patreon for AI-generated art that's just mashing up other artists' work, especially when it involves prompts like "in the style of X artist" or "trending on Artstation". It's shitty when someone just traces art and resells it, and it's shitty when someone has a robot do the equivalent.

There's an extra little bummer for me, because I'm super turned on by the idea of a computer that makes porn that looks great at first until you notice the unsettling number of fingers or the fact that their lingerie is melting into their flesh. Maybe at some point I'll track down an ethically-sourced image generation AI with acceptable results to use for a horror expedition.

 

[tioafa4]

I saw a thread on a different porn forum the other day where someone was discussing the benefits and drawbacks of AI adult art. For the drawbacks, they simply posted an image where the AI had made a photorealistic triangle of shins, like a cross between an ouroboros and a triskelion. AI is definitely unmatched at fearlessly exploring the uncanny valley.

 

[Stroth]

AI art is very impressive but I also feel that it's vastly understated how much you need to generate to get something consistent and that's not, say, got mangled ears, six fingers, a missing hand, etc

Just playing devil's advocate, I'm of the opinion that people vastly overestimate how good AI generated art is right now and quite like the current art, but for this particular game the stated problems there would not necessarily be an issue in a lot of cases.

 

[PhanesDionysos]

Hello, perverts! It's time for a new version of Ruin Me.

The new version for supporters adds many fixes, a new minor horror story, and four special Halloween costumes! If you support me in the next few weeks, you'll be able to get the costumes for free; afterward they'll require in-game resources. The current free version has the opportunity for creeping on Theresa at night and an optional, opt-in soft-NTR mode.

You can always check out the latest free version at:

You must be registered to see the links

Version 0.50.0p is the new patron version and adds:

• Four new outfits for Theresa, all burlesque/cabaret and Halloween themed (and using the same basic shape). In increasing levels of sluttiness, they are The Black Cat (inspired by Poe), a Wicked Witch of the East (inspired by Baum), a Vamp of Dracula (inspired by Theda Bara and Musidora), and an Eldritch Whore (inspired by Kuroboshi Kouhaku's design for Ninomae Ina'nis). They're available from the Tailor, found by exploring out of town. Within two weeks of Halloween (Oct 31), they're available for free; for the rest of the year, they cost Ken like other outfits.
• A new minor exploration, the abandoned bedroom, found when exploring the mansion by spending five Ken. Within, you will explore a little mystery room that may shed some light on the history of the Radcliffes.
• A tweak so that you now earn one Ken for each horror expedition you complete. Existing saves will be credited retroactively.
• A reduction in the cost of outfits from the Tailor. Existing saves will not be refunded.
• Adjustments to the main mansion navigation menu: it's been reordered and you now visit Esther from there instead of from the "Elsewhere on the grounds..." menu.
• A text ambience for each horror expedition that can be displayed randomly upon visiting town once the expedition is complete. You shouldn't see these repeat too often.
• The ability to close the journal by clicking/tapping on the space behind it.
• Content warnings for each horror expedition that are visible in the journal if you haven't disabled content warnings in the preferences.
• Additional improvements to Theresa's main standing pose, including a redraw of her resting arms and better shading.
• An image for the "cork from a cheap bottle of wine" curio, which is also now shown and mentioned in the expedition "The Treacherous Splay".
• A change so that Theresa's graduation is only available once you've finished with the Birds, to make sure the plan is introduced in the correct order.
• An option to wait until dusk when appropriate in the expedition "The Adulterine Soap".
• A split of the Preferences menu between "System Preferences" and "Game Preferences".
• New art for radio buttons and check buttons (in the Preferences menus).
• An improvement so the game no longer plays audio while minimized.
• A fix for my pluralize helper that should improve pluralization of chosen nomenclature.
• A fix for an error when viewing the completed Chthonic Swelling expedition in the journal.
• Additional save slots per page on Android.
• An upgrade to Ren'Py 8.0.3 (currently using Python 3.9.10). This should cause no changes, so if you notice something weird, please let me know!
• Many minor enhancements and fixes.

0.49.2 is free for everyone, adding:

• The option to check in on Theresa at night, including consensual peeping and waking her up with sex. You can get permission with the behavior menu option in her chambers.
  • More options unlock as Theresa's stats increase.
  • The new bedtime menu can be disabled by telling her you'll leave her alone at night. Asking about it again will re-enable the prompt.
  • This feature may be expanded in a future release.
• A new option for sophisticated tastes, available from the "Examine your locked suitcase" option in your bedroom after Day 3. This is opt-in writing that is added to existing scenes. The new option, "Jealousy and Pride: How Inferiority Can Engender Ecstasy", adds teasing of the player, inadequacy, jealousy, and consensual cuckolding vibes to many scenes.
  • This is generally pretty "netorase" in feel as opposed to other forms of NTR and doesn't involve any raceplay or specific anatomical criticisms.
  • This is 100% optional and opt-in; you won't see any added material until you choose to.
  • You can set "Sophisticated Content" to "Highlight" in the preferences to see when text has been added or modified.
• A fix to the "Have Theresa dominate you in manual sex" play scene and its Theresa-initiated equivalent, which previously had unintended behavior when you use a safe word or otherwise end the scene early.
• An enhancement to "The Pileous Retort" with a unique line for each stories you choose to tell.
• Additional minor options for the ending of the sex scene when getting Theresa into the Reform Institute, providing opportunities for a bit of partner-sharing kinkiness.
• A change to the default ambience logic to make events like town ambiences avoid the same event recurring more than every three days. This should increase the likelihood of novel events, but please let me know if you're seeing too many repeats or other odd behavior.
• Additional forms of address and other nomenclature for you and Theresa to allow for different relationship dynamics ("Comrade", "friend", "cuck", "Goddess", and more).
• A correction to the spelling of "anilingus", which I never would have guessed was spelled that way!
• The changelog is now accessible from the main menu (from a link beside the version number in the corner).
• A new option in the "Bend Reality" cheat menu, available to $5+ patrons: you can enable the option to pick certain outcomes that are normally random. For the moment, you can choose ambiences (in town or at breakfast) and certain random events when checking in on Theresa at night. More places to control randomness may be added later.
• Minor enhancements and fixes.

The patron version also has the Reminisce feature, letting you replay old erotic sequences that otherwise can't be revisited and take new paths through them. $5 patrons can access a special cheat menu, guide development priorities with polls and suggestions, and make requests for specific additions to the game.

 

[Larsung]

ith this game pure maledom or are there 2 routes one femdom too?

 

[Larsung]

asking because i reat this in the changelog

• A fix to the "Have Theresa dominate you in manual sex" play scene and its Theresa-initiated equivalent, which previously had unintended behavior when you use a safe word or otherwise end the scene early.

 

[PhanesDionysos]

ith this game pure maledom or are there 2 routes one femdom too?

You can pick your own gender. There's only one main route: Theresa is generally submissive, as part of her personality, but there are options for her to be dominant at times as her corruption progresses, plus an optional soft NTR mode where the vibe is more complicated.

There are also various side characters that can take a dominant role with your character.

 

[DannyLucifer]

is there any plan to replace the art after a while by the dev?

 

[Darth Sidious]

Always amuses me how devs feel the need to make excuses for adding NTR in their own games Hushhhh... It's ok, and just...

 

[PhanesDionysos]

is there any plan to replace the art after a while by the dev?

I'm steadily and gradually updating all the art (I don't think there's any original character art left in the game from the first release) but I don't have any plans to change the style; it'll always be pretty cartoony.

Always amuses me how devs feel the need to make excuses for adding NTR in their own games

I will never make excuses! It's hot! But I did my best to try and answer some of the common questions people have about this stuff in the changelog.

 

[Retrofire]

This is really something. It's not my usual kink set, and the art isn't really my thing, but the writer has real talent!