I found a bunch of ways online to go about this (Daz3D pops up almost always) but I'm wondering more what is the usual way a Ren'py porn developer in 2024 goes about creating renders/images for their game. I've been playing around with stable diffusion these last few days and been able to cook up some pretty nice images with it which have gathered quite a bit of upvotes on rule34 by now, how viable is it to use stable diffusion images to create the renders for you Ren'py game ? And how many people actually go this route. If this is a long shot I'm fine with trying to learn Daz3D but it seems quite a bit of work to setup an entire scene, procure the renders you want, etc etc. Whereas with stable difussion I can have some pretty specific and good looking 2D images in a few minutes.
Ren'Py How do most people make renders/images for Ren'py in 2024?
- Thread starter SamsiraTrigger
- Start date
- Jun 19, 2019
- 8,454
- 32,263
TK17, DAZ, Koikatu, Honey Select 1 & 2, AI, Kisaeki Dolls. Lots of ways.
Yes but I was more wondering a bit what the popular or even "go-to" way is nowadays, is it just generating some images using AI, import some renders in DAZ, actually go all the way to create your own renders from scratch. There seems to be a million ways to go about this but I don't have the time or energy to sift through all these, learn how they work, actually create something with them and then figure out if it's a viable thing or not if you get what I mean. If you were to start as a Game Developer right now what approach would you take ?
Man, you already seem to know the answer, you just don't like it.
If you want to know what's most popular, go check the Update Page. I set the search for game updates from the last 14 days and ordered them by views.
I can assure you, you will find no AI "Art" game on those pages, and my guess is the community will hate it too. It's not consistent enough to be pleasing. You can't make animations with it—at least not good ones. I'm not even going to talk about ethics here.
If you think you can be the first indie developer to make a good and famous AI game, go for it. But I believe it's easier to go with what people like. If you want to discover what that is, play the most famous titles on this site. They are not famous by accident.
What it looks like to me is that you are asking if you can make money using AI on this site. The answer to that is: probably no. But if you just want to make a game and maybe, someday, with a lot of luck, earn something from it, then yes, you will have to learn how to use some of the tools like the ones Icarus Media sends you.
Good look man, and remember, AI "Art" is not Art
If you want to know what's most popular, go check the Update Page. I set the search for game updates from the last 14 days and ordered them by views.
I can assure you, you will find no AI "Art" game on those pages, and my guess is the community will hate it too. It's not consistent enough to be pleasing. You can't make animations with it—at least not good ones. I'm not even going to talk about ethics here.
If you think you can be the first indie developer to make a good and famous AI game, go for it. But I believe it's easier to go with what people like. If you want to discover what that is, play the most famous titles on this site. They are not famous by accident.
What it looks like to me is that you are asking if you can make money using AI on this site. The answer to that is: probably no. But if you just want to make a game and maybe, someday, with a lot of luck, earn something from it, then yes, you will have to learn how to use some of the tools like the ones Icarus Media sends you.
Good look man, and remember, AI "Art" is not Art
Hi, I already played quite a lot of games on this site over the years but that doesn't change the fact that I don't know how those renders were created in the first place, I'm sorry if I upset you in any way I'm not looking for a get rich quick scheme or anything like that I just want to get my foot in the door of game development and wanted to poke around for what people think would be a good route to start. I know most games have 3D renders which still leaves my question how did they create those renders and did they make those renders themselve from scratch ? Icarus Media linked a bunch of useful stuff but I'm still looking for somone to point me a little more concrete in a direction on how to go about this.
I'll try and clarify what I'm asking for a little more:
- What is the recommended tool to start creating images/renders for Ren'py games for beginners ?
- Given the answer on the above question is some 3D rendering tool, do you recommend creating your own models from scratch or stick to models other people have created ? (Not looking to pay someone for models so other models would mean free models you can find online here)
I don't want to come across as lazy or someone looking to make a quick buck, there is just a lot of information about this out there and most threads surrounding this that I could find are atleast several years old
I'll try and clarify what I'm asking for a little more:
- What is the recommended tool to start creating images/renders for Ren'py games for beginners ?
- Given the answer on the above question is some 3D rendering tool, do you recommend creating your own models from scratch or stick to models other people have created ? (Not looking to pay someone for models so other models would mean free models you can find online here)
I don't want to come across as lazy or someone looking to make a quick buck, there is just a lot of information about this out there and most threads surrounding this that I could find are atleast several years old
Icarus Media literally told you on the 2nd post tho
You'll be a beginner with all of them. If you're asking for a magic tool that pops renders without you having to put the effort to learn it, there's none. The biggest bottleneck is your machine (Some of those tools require good hardware), and the art style you want to go for.
No one creates humanoid models from scratch, only big studios or teams do, and not even that.
You can model your own environment stuff if you go for a stylized look, but otherwise you purchase them.
- Sep 27, 2017
- 499
- 923
If you're not going to do complicated scenes, Stable Diffusion is the best choice. Like here, for example, a girl and a dick without a face.
If you want to create more detailed scenes with multiple characters in the frame, Daz3d will be much better. Although the developer of this AI game is doing something I'm talking about City of Dreams, but probably a lot of editing is done here to get the final result.
At the moment, I think Daz3d is better, but AI looks more promising.
Daz3d has a huge amount of different stuff that is freely available if you don't want to pay.
If you want to create more detailed scenes with multiple characters in the frame, Daz3d will be much better. Although the developer of this AI game is doing something I'm talking about City of Dreams, but probably a lot of editing is done here to get the final result.
At the moment, I think Daz3d is better, but AI looks more promising.
Daz3d has a huge amount of different stuff that is freely available if you don't want to pay.
- Mar 11, 2024
- 190
- 854
there are no answer to this, there are no "usual" or "recommended" ways.
Icarus mentioned the most popular options, and those are the "easy way". Which is using "premade tools", or games that has been turned into "premade tools".
But even then each and every single one of those still require you to put a few hundred of hours(atleast) into learning and messing around with it before you can put out something that doesnt look like an eldritch horror
On the question of whether or not you should make your own 3d models... The answer is no. Assuming you have 0 background in art and 3D, to learn everything from the very basic to the level where you could model your own human 3D model is gonna take a couple of years.
But modeling isnt everything you need, there's also texturing, rigging, retopology, posing, lighting and a million other things... it's time consumming and extremely expensive.
And nobody make everything from scratch, not even big games companies, even goddamn Fromsoftware still use premade assets in Elden Ring.
I believe most developers here are already somewhat familiar with one of these tools before they even thought of making a game, be it they have background in art/graphic designs... or they actually play those honey select/koikatsu games and know how to navigate it. If you are neither, then I guess you just have to choose on the style of renders you wanna make, and then spend the next 6 months or something drowning yourself in tutorials to see if it's even worth doing for you.
Icarus mentioned the most popular options, and those are the "easy way". Which is using "premade tools", or games that has been turned into "premade tools".
But even then each and every single one of those still require you to put a few hundred of hours(atleast) into learning and messing around with it before you can put out something that doesnt look like an eldritch horror
On the question of whether or not you should make your own 3d models... The answer is no. Assuming you have 0 background in art and 3D, to learn everything from the very basic to the level where you could model your own human 3D model is gonna take a couple of years.
But modeling isnt everything you need, there's also texturing, rigging, retopology, posing, lighting and a million other things... it's time consumming and extremely expensive.
And nobody make everything from scratch, not even big games companies, even goddamn Fromsoftware still use premade assets in Elden Ring.
I believe most developers here are already somewhat familiar with one of these tools before they even thought of making a game, be it they have background in art/graphic designs... or they actually play those honey select/koikatsu games and know how to navigate it. If you are neither, then I guess you just have to choose on the style of renders you wanna make, and then spend the next 6 months or something drowning yourself in tutorials to see if it's even worth doing for you.
There's no AI art games because it's a recent technology, and an even more recently added tag. The great majority of games at the top are old, way before AI was a thing, and it's way more likely someone will use common and easier to use tools to make something good than a new technology. It's also one of the reasons why Unity is not more popular despite being more powerful.
Sure, Unity is not a new technology, yet Ren'Py is so much easier to use. AI is hard and new, it's not going at the top anytime soon, but it's not due to bad quality, there are many good looking (and rated) games already:
https://f95zone.to/sam/latest_alpha/#/cat=games/page=1/tags=2265/sort=rating
Remember to screenshot that quote to prove you're fighting the good fight so you can attach it to your next CG order for a 50% discount, it'd be crazy not to pay only 150 euro instead of the full price.
Thank you all for the input, appreciate it! I'll probably see what I'm able to do using stable diffusion and some image editing and if not switch over to Daz3D using some already created humanoid renders
I have little to no experience with art (besides playing around with stable diffusion this last week), but I do have a lot of programming experience (CompSci graduate) but I'm always up for something new to learn
- Mar 11, 2024
- 190
- 854
Oh then you understand the amount of effort needed to learn it then. The biggest obstacles in learning art is false expectation.
dont know if it helps you but I intend on leveraging AI to create some games. My thoughts were models like Flux + controlnets for the character generation. Then I'd take the characters into photoshop to separate the character into movable parts. From there, I'd be doing looping state animations in live2D. Paired with the Unity SDK , it should make for a fairly good looking interactive game.
- Jun 28, 2024
- 36
- 61
2024 Newcomer developer here.
Completely understand how daunting it is to think about spending a lot of time learning software to create images. Indeed, inputting prompts and getting image output is convenient and quick. However, I think you also know that while AI-generated images can be beautiful, achieving consistency and adjusting angles can be difficult.
I also have no clue about those 3D software programs. I tried Blender for a few days, experimented with AI also, but in the end, I abandoned all these options. I've concluded that as a beginner, it's best to use ready-made tools, like Koikatu and Honey Select.
However, in the end, I also abandoned Koikatu and Honey Select and chose to go with physical, handmade production, which is actually more challenging than learning 3D software. But at least the method I chose provides more realthings feedback and allows for creating tangible, interactive objects. And I still continue to use AI tools to assist with the design of puppets and scenes, or to provide inspiration and references. Besides I chose this method considering its future applications.
But to be honest, if you have little to no experience in art like me, Koikatu and Honey Select would be the better choice. Or, you might find a way to combine them with AI tools. Moreover, it all depends on the type of game or VN you want to create.
Last edited:
- Jun 7, 2017
- 9
- 1
I am not sold on using AI for models specifically. There's a lot of context that is missing; (like if a model is lying down, what angle should it be at to look right on the bed at for example). I actually see potential on using AI specifically for generating the one off, "I'm traveling images" with some trickery to hide that that was used for the background. Those are literally one offs and thus are natural candidates for looking different.
Also in terms of looks, most AI I've seen can pretty consistently get to 7 out of 10 with some knowledge of what you want. The problem is ironically enough, variety. It ends up looking too similar too each other in a game, and that's really bad for making people like the game.
Also in terms of looks, most AI I've seen can pretty consistently get to 7 out of 10 with some knowledge of what you want. The problem is ironically enough, variety. It ends up looking too similar too each other in a game, and that's really bad for making people like the game.
- Aug 17, 2019
- 5,452
- 8,734
The one alternate option I'm not seeing here is Blender + AI. I'm not too familiar with how advanced it is, but it's definitely a thing. Might be worth a look, but I could definitely see a struggle with adult content depending on the model being used.
Just to not discourage anyone here, these tools (Daz3D, Blender, ZBrush, UE) used for top-tier games can be quite easy to learn. They require some time, but the outcome is unmatched by anything else. You do need a good PC to make the workflow faster and "viable", but if you have that, dive in and learn. The possibilities for poses, expressions, lighting, props, environments, and overall adaptability are endless. You can create exactly what you imagined
I still have my first two works with 3D modeling. Both were made on my first day of learning how to use Daz, and they are just the default Genesis 8 model altered in Blender
After months of experience, I see many, many, and many flaws in them that I now know how to avoid, which is one reason I keep them. These tools are easy to use and deliver incredible results. Yes, AI is easier, you just type a prompt and see the result, and if you don't like it, you can roll it again. But you won't get the same amount of adjustments and detail
Those were rendered in 4K with RayTracing, probably the upload will reduce the quallity
I still have my first two works with 3D modeling. Both were made on my first day of learning how to use Daz, and they are just the default Genesis 8 model altered in Blender
After months of experience, I see many, many, and many flaws in them that I now know how to avoid, which is one reason I keep them. These tools are easy to use and deliver incredible results. Yes, AI is easier, you just type a prompt and see the result, and if you don't like it, you can roll it again. But you won't get the same amount of adjustments and detail
Those were rendered in 4K with RayTracing, probably the upload will reduce the quallity
- Jul 4, 2017
- 2,597
- 4,736
I've been grinding away on my own little project for around 6 months now (omfg how can it be that long already!).
Using Stable Diffusion or similar tools is only a shortcut if you are willing to accept the many limitations:
- cannot force an exact "look" (face, hair, body, clothing) to be repeated in different poses/angles/situations
- deformed or unnatural body parts
- impossible or improbably physical designs for objects/environments
- difficult control over lighting and positioning in scenes, especially with multiple figures
- inconsistent art style from one image to the next
- an inescapable "that's AI" appearance
The only way I've found around some of these limitations is extensive image editing and rework, and rework, and rework.
Some aspects can be forced (geometry, scene composition) to some extent by using i.e. daz or blender to render simple 3d scenes and then using those as a base for img2img generation. Art style can be made somewhat consistent by using LORA or other fine-tune techniques. The rest can only be fixed by a combination of image edits (hand painting, and also editing together the best parts from multiple generations) followed by another "healing" pass of img2img to hide the edit seams.
But it still takes lots of time. LOTS of time.
Unless you are willing to accept that the output will be very obviously "dev spammed Stable diffusion images and threw them all together without much care", like we see in the worst cheap AI flip games, the old truism prevails: "there is no such thing as a free lunch". You're not getting that much, if any, speed advantage by using AI over using Daz.
Using Stable Diffusion or similar tools is only a shortcut if you are willing to accept the many limitations:
- cannot force an exact "look" (face, hair, body, clothing) to be repeated in different poses/angles/situations
- deformed or unnatural body parts
- impossible or improbably physical designs for objects/environments
- difficult control over lighting and positioning in scenes, especially with multiple figures
- inconsistent art style from one image to the next
- an inescapable "that's AI" appearance
The only way I've found around some of these limitations is extensive image editing and rework, and rework, and rework.
Some aspects can be forced (geometry, scene composition) to some extent by using i.e. daz or blender to render simple 3d scenes and then using those as a base for img2img generation. Art style can be made somewhat consistent by using LORA or other fine-tune techniques. The rest can only be fixed by a combination of image edits (hand painting, and also editing together the best parts from multiple generations) followed by another "healing" pass of img2img to hide the edit seams.
But it still takes lots of time. LOTS of time.
Unless you are willing to accept that the output will be very obviously "dev spammed Stable diffusion images and threw them all together without much care", like we see in the worst cheap AI flip games, the old truism prevails: "there is no such thing as a free lunch". You're not getting that much, if any, speed advantage by using AI over using Daz.
Last edited: