While digging through my Steam Library, I almost forgot I had a free generative AI music program called Robo Maestro, maybe it could be used for something larger, so I far I only played around with it a bit. I don't have any saved examples yet, but the Steam page shows off what it can do, and its completely free with very low system requirements.
I wasn't sure if this program could be accepted in the useful tools section since it is AI based, so this seems like the right spot for now.
AI General Sharing Thread
● ARCHIVED · READ-ONLY
-
-
This looks amazing, but I believe you're not supposed to upload images to an AI, of which you don't have the right to share. This means you can not really use copyrighted content to upload to AI and ask for edits.I found this in an ad... he did some really interesting things with the 2k3 monsters
View attachment 356875
View attachment 356876View attachment 356877 -
After fanarts of my waifu, here are fanarts of my character Luciana Vincenti, aka the Queen of Adretana, this time not made me but by [GM]-Phantom, a player from my Discord. Luciana appeared in 8 games and started her adventures back in 2001 in my most enduring series (still running to this date), the Laxius saga B):rock-right:
FYI the original :
-
@Indinera Those are absolutely beautiful!
Honestly, anyone who says that AI can't make images that are as good or better than many human artists is either delusional or just plain ignorant because they refuse to focus on anything other than the negatives of AI.
And I honestly think we are at a point where many people can't really tell the difference between a really good AI generated image and one that is made by a human. I think they make assumptions and go with whatever they are told. And I bet there are a bunch of people who would believe that AI generated images are actually made by a human if that is what they were told. And they would think it was great...until they were told it was made by AI and then it would be terrible.
I would really like for the naysayers to take a really good AI image and then tell me exactly how they know it was made with AI. Take the images recently posted. How exactly can you tell that those images were made with AI other than being told?
Anyway, Day 4 of posting some AI generated images behind a SPOILER. These actually have a few icons that I said I was uploading yesterday, but I forgot. As well as a series of the same treasure chest. I wanted a normal looking treasure chest and then I wanted that same chest outlined with different colors. That way you could outline the chest in green as being something you can interact with, outline it in blue for say a Detect Magic spell, outline it in yellow for some sort of Divination spell, and outline it in red for say Detecting a Trap.
SPOILER - ICONS & CHEST VARIATIONS
I plan on using some of these pixel art images to try to use Aseprite to recreate them as part of learning to create pixel art by myself. These are all much larger images and would have to be scaled down for use in an actual game.
Have a great day! -
The pixel art renditions are pretty good for sure. It is indeed becoming more and more difficult to tell AI apart from man-made resources.
Those are absolutely beautiful!
Thank you :p I loved them too :wub
Honestly, anyone who says that AI can't make images that are as good or better than many human artists is either delusional or just plain ignorant because they refuse to focus on anything other than the negatives of AI.
Those who say that are generally also very ignorant and/or misled about other game making things, I've noticed. Probably not a coincidence.
And I honestly think we are at a point where many people can't really tell the difference between a really good AI generated image and one that is made by a human. I think they make assumptions and go with whatever they are told. And I bet there are a bunch of people who would believe that AI generated images are actually made by a human if that is what they were told. And they would think it was great...until they were told it was made by AI and then it would be terrible.
Yes that's my biggest gripe about AI. It's invasive and vastly unidentified.
It happens with videos too. Some are blatantly rubbish and can easily be disregarded but some are... you don't really know.
The other issue also is when a video is truly fantastic, even if it was man-made, people are going to say it's AI. "then it would be terrible", you said it.
I would really like for the naysayers to take a really good AI image and then tell me exactly how they know it was made with AI. Take the images recently posted. How exactly can you tell that those images were made with AI other than being told?
I've seen naysayers try to pick on very small details in an attempt to slap "slop" (pun intended) on everything AI. The problem is that even man-made art has often had small mistakes here and there (sometimes not so small, even). Very few people master their art so much that they can consistently provide perfect renditions.
Anyway.
More stuff of my 25 years old characters... not canon but pretty good as a fanfiction, it could almost be a VN lol I'm glad my work is inspiring people to make the most out of AI generations B)
-
Day 5 of sharing images made with Google Gemini.
These are non-pixel art experiments.
Assorted Character Scenes
Enjoy your day! -
3rd is best!
-
Hi friends! How's new year treating you so far, I hope you all are holding up alright.
First off, I would like to apologise for making a thread for my acapella songs which I converted into studio songs using Suno. Thank you to @slimmmeiske2 for reminding me about this AI thread.
Friends, pls consider checking out my songs ^^
Here's some tracks for you to hear from the get go:
My Songs
-
Here is some early concept art for my second game, Chronicles of Avineria. These are probably not going to be used in the actual game.
-
Some professional game developers used AI for prototyping and concept art that was later swapped out before the final release, its something to work with during production even if the completed game doesn't have it. Those results came out very well, it hard to tell if the images are raw or had any kind of clean up.Here is some early concept art for my second game, Chronicles of Avineria. These are probably not going to be used in the actual game.
Also I'm doing some prototyping with one of the the Perchance character description generators that also makes images after the backstory attempts based on my alternate persona, Nightfire.
I'm still trying to figure out good prompts to eventually test with the dedicated image generators, robot designs very wildly so its actually very difficult to get the right one with AI. -
Looks great. Perchance was actually exactly what I used for my images above as well.Some professional game developers used AI for prototyping and concept art that was later swapped out before the final release, its something to work with during production even if the completed game doesn't have it. Those results came out very well, it hard to tell if the images are raw or had any kind of clean up.
Also I'm doing some prototyping with one of the the Perchance character description generators that also makes images after the backstory attempts based on my alternate persona, Nightfire.
View attachment 357188View attachment 357189View attachment 357190View attachment 357191View attachment 357192
I'm still trying to figure out good prompts to eventually test with the dedicated image generators, robot designs very wildly so its actually very difficult to get the right one with AI.
Also, I forgot one character. This would be Dahlia, the protagonist of Chronicles of Avineria. -
Some professional game developers used AI for prototyping and concept art that was later swapped out before the final release, its something to work with during production even if the completed game doesn't have it
Sounds hypocritical to me, like let's happily use what AI makes, but then ask a human to redo it so we can wash our hands and (fakely) come out clean. And they only do it because they have the budget for it and probably evaluated they'll make more money this way than just taking the risk to keep the AI material and getting called out for it.
As for the latest pics - pretty good if a bit AI-generic.
Perchance tends to make all anime girls look alike in style. -
Don't really see the issue personally. If wouldn't count it as ai being in a game if none of it survive in the final version. I kinda did this for Dextera Mortis actually. This sort of artwork is quite good to show an artist as a reference so they have some idea what a character is supposed to look like, though I also tend to not want a carbon copy of the ai artwork. But I tend to be rather particular about how I envision my characters and sometimes it is hard to communicate to the artist what my vision for that character is without an example.Sounds hypocritical to me, like let's use what's AI makes, but then ask a human to redo it so we can wash our hands and (fakely) come out clean. And they only do it because they have the budget for it and probably evaluated they'll make more money this way than just taking the risk to keep the AI material.
As for the latest pics - pretty good if a bit AI-generic.
Perchance tends to make all anime girls look alike in style.
Though I am hardely a professional developer. Money is not the goal for me, rather I want as many people playing my game as possible. I released my first game for free for this very reason.
This would the the built in Character generator version of Dahlia by the way. It may or may not be the portrait I will use in the final version. I did commission artwork for my first game, but that was quite expensive. So I am not sure if I want to do so next time or not. Last time I could not afford custom artwork for everything. So the art ended up being a bit inconsistent, being a blend of commissioned art and character generator portraits. It is perhaps better to use only character generator portraits? -
Overall, I think these look nice. But I'm not sure if Perchance is as good as Google Gemini. One thing I would watch for and try to get the AI to fix is the most common areas of mistakes.Here is some early concept art for my second game, Chronicles of Avineria. These are probably not going to be used in the actual game.
For instance, in your first two images of that set, the first image seems to have 5 fingers on the character's right hand (thumb isn't visible either) and the second image looks like it has two right hands for the character.
I really try to scrutinize AI images and see if it can fix obvious mistakes. I haven't seen many problems with Google Gemini when it generates a new image, but it definitely can have some problems when you try to modify images. And I have seen it put parts of weapons through the character's body in certain poses.
As a general thought, one of the things that bothers me about the hypocrisy of the anti-AI image movement is the fact that many of the people spread complete misinformation about how AI generated images are made and many of them don't seem to have a problem when a human artist does something similar to what they think AI does.
For example, I'm a very visual person. I have a very hard time conceptualizing something in my head. So, when I practice drawing, I have to work off something. Either something real or another image.
If I "copy" and "modify" another person's drawing, as long as it is substantively different, many people don't seem to have a problem with that because it was done by a human. If I get too close to the original, then that is when there is a problem.
I mean after all, most drawing books I have seen have you practice recreating someone else's drawing. Humans have copied each other for thousands of years.
But suddenly, when AI supposedly does something similar, it's the end of the world.
Honestly, I think that by the time I ever am ready to release a game that uses AI generated images as placeholders, this whole AI image argument will be mute. The anti-AI group will continue to get smaller and smaller from the massive group of players who are more focused on the game and not whether AI was used in the process.
And the other thing that really bothers me is the whole argument about how AI is destroying the planet. I get it and I appreciate the concern, but there are thousands of other things that are far more dangerous and have been for decades that people haven't gotten anywhere near as concerned about.
The reality is that the potential harm that AI generated images could cause to the planet, specifically in small indie game making, is infinitesimally smaller than other uses for AI that apply across disciplines, sciences, medicine, government uses, etc.
Even if there were new, global laws passed tomorrow that eliminated AI generated images and movies (making them illegal) across the board, AI would still be used in millions of other ways in all aspects of life. The use of AI is not going away and will only exponentially increase. Image generation is just a tiny drop in a massive AI pool.
Most people are currently exposed to things related to AI every single day and they don't know it and they generally don't really care because it doesn't impact them personally. And they certainly don't think the benefits of AI in other contexts is "AI slop".
Anyway, Day 5 of a few more AI images with Google Gemini comparing pixel art and non-pixel art versions. I'm trying to decide whether I like pixel art or non-pixel art for a future RPG project.
Comparison of Pixel and Non-Pixel Art
I really like the non-pixel art images, but I also know that since I'm trying to learn how to make pixel art myself, I'm far more likely to be able to make usable pixel art than I am trying to learn how to draw a non-pixel art image.
Anyway, have a great day everyone and good luck on your projects! -
Thats a dope-looking castle. You do seem to have more experience prompting than I do. I can't get anywhere near that good of a result. Maybe I should try google gemini.Overall, I think these look nice. But I'm not sure if Perchance is as good as Google Gemini. One thing I would watch for and try to get the AI to fix is the most common areas of mistakes.
For instance, in your first two images of that set, the first image seems to have 5 fingers on the character's right hand (thumb isn't visible either) and the second image looks like it has two right hands for the character.
I really try to scrutinize AI images and see if it can fix obvious mistakes. I haven't seen many problems with Google Gemini when it generates a new image, but it definitely can have some problems when you try to modify images. And I have seen it put parts of weapons through the character's body in certain poses.
As a general thought, one of the things that bothers me about the hypocrisy of the anti-AI image movement is the fact that many of the people spread complete misinformation about how AI generated images are made and many of them don't seem to have a problem when a human artist does something similar to what they think AI does.
For example, I'm a very visual person. I have a very hard time conceptualizing something in my head. So, when I practice drawing, I have to work off something. Either something real or another image.
If I "copy" and "modify" another person's drawing, as long as it is substantively different, many people don't seem to have a problem with that because it was done by a human. If I get too close to the original, then that is when there is a problem.
I mean after all, most drawing books I have seen have you practice recreating someone else's drawing. Humans have copied each other for thousands of years.
But suddenly, when AI supposedly does something similar, it's the end of the world.
Honestly, I think that by the time I ever am ready to release a game that uses AI generated images as placeholders, this whole AI image argument will be mute. The anti-AI group will continue to get smaller and smaller from the massive group of players who are more focused on the game and not whether AI was used in the process.
And the other thing that really bothers me is the whole argument about how AI is destroying the planet. I get it and I appreciate the concern, but there are thousands of other things that are far more dangerous and have been for decades that people haven't gotten anywhere near as concerned about.
The reality is that the potential harm that AI generated images could cause to the planet, specifically in small indie game making, is infinitesimally smaller than other uses for AI that apply across disciplines, sciences, medicine, government uses, etc.
Even if there were new, global laws passed tomorrow that eliminated AI generated images and movies (making them illegal) across the board, AI would still be used in millions of other ways in all aspects of life. The use of AI is not going away and will only exponentially increase. Image generation is just a tiny drop in a massive AI pool.
Most people are currently exposed to things related to AI every single day and they don't know it and they generally don't really care because it doesn't impact them personally. And they certainly don't think the benefits of AI in other contexts is "AI slop".
Anyway, Day 5 of a few more AI images with Google Gemini comparing pixel art and non-pixel art versions. I'm trying to decide whether I like pixel art or non-pixel art for a future RPG project.
Comparison of Pixel and Non-Pixel Art
I really like the non-pixel art images, but I also know that since I'm trying to learn how to make pixel art myself, I'm far more likely to be able to make usable pixel art than I am trying to learn how to draw a non-pixel art image.
Anyway, have a great day everyone and good luck on your projects!
Also, with the extra 5 fingers, if you mean the Nyx image, my original plan was to crop in in a way so the fingers are not visible. If I were to use it in game it would need to be in a 144x144 format anyway. Not sure if I will or not. It is entirely possible that when Chronicles of Avineria is done in a couple of years, use of ai will not have the stigma it has now. Maybe I will do a combination of ai and commisioned art. -
@Darkmoon9 I've actually only been actively using AI to experiment with generating images for about a week now. But I've definitely found that Google Gemini gives me the best results so far. But I also haven't tried any paid options yet so those could be even better.
I've also found that Gemini has a problem with reducing a full body or larger portion of the body image down to a portrait. Of course, it took me a bit to realize that a portrait isn't necessarily limited to the Faces type of portrait that is used in RPG Maker.
I've found that specifically giving Gemini references to RPG Maker can be helpful. It definitely helps when learning Javascript.
Gemini seems better at generating a Faces portrait when specifically referenced to RPG Maker MZ or MV (or your version) and then expanding that portrait to a full body image than the other way around for some reason.
And I like Gemini because I can actually ask it why it generated an image the way it did and see its logic. Sometimes it refers back to something I used earlier in that thread and I have to tell it not to.
It is definitely a learning experience. But Gemini is nice because I can have an actual conversation about things and ask it why things are the way they are and it can be really helpful.
Any sometimes, Gemini just doesn't follow the prompt. I have been actively using the feedback option to let it know when it gave a good response or a bad one. -
I used Starryai until I found Perchance gave similar results with infinite generations instead of a daily credit system.@Darkmoon9 I've actually only been actively using AI to experiment with generating images for about a week now. But I've definitely found that Google Gemini gives me the best results so far. But I also haven't tried any paid options yet so those could be even better.
I've also found that Gemini has a problem with reducing a full body or larger portion of the body image down to a portrait. Of course, it took me a bit to realize that a portrait isn't necessarily limited to the Faces type of portrait that is used in RPG Maker.
I've found that specifically giving Gemini references to RPG Maker can be helpful. It definitely helps when learning Javascript.
Gemini seems better at generating a Faces portrait when specifically referenced to RPG Maker MZ or MV (or your version) and then expanding that portrait to a full body image than the other way around for some reason.
And I like Gemini because I can actually ask it why it generated an image the way it did and see its logic. Sometimes it refers back to something I used earlier in that thread and I have to tell it not to.
It is definitely a learning experience. But Gemini is nice because I can have an actual conversation about things and ask it why things are the way they are and it can be really helpful.
Any sometimes, Gemini just doesn't follow the prompt. I have been actively using the feedback option to let it know when it gave a good response or a bad one.
I also use online image resizers and background removers to crop the image and reduce it to a 144x144 png and then run it through a background remover. I do the same thing with commissioned art as well. So I don't need the ai or artist to produce the exact dimensions I need for RPG Maker.
Also good to deliberately keep the hands out of the image as ai often screws up the fingers, at least in the images I have generated. -
@Darkmoon9 Yeah, I still use GIMP for scaling and such. But I bought Aseprite recently to learn how to make my own pixel art so I'm going to check that out too.
BTW, is your user name associated in any way with the 2nd game in the Eye of the Beholder series? The Eye of the Beholder II: The Legend of Darkmoon?
Just curious. It's probably a reference to something else, but that is what it always reminds me of. -
Nope, that is just a coincidence. The only game I have made so far is called Dextera Mortis. Released the full version in Itch a couple of days ago. It might have been a reference to the Blades of the Darkmoon in the first Dark Souls though.@Darkmoon9 Yeah, I still use GIMP for scaling and such. But I bought Aseprite recently to learn how to make my own pixel art so I'm going to check that out too.
BTW, is your user name associated in any way with the 2nd game in the Eye of the Beholder series? The Eye of the Beholder II: The Legend of Darkmoon?
Just curious. It's probably a reference to something else, but that is what it always reminds me of.
Also, did an experiment with Gemini on one character from Chronicles of Avineria. Totally nailed it. Have trouble with Dahlia though. How do I get it to do a different variation from the same prompt? -
@Darkmoon9 I have had pretty good luck with Gemini with things like "Change the image..." if I'm referring to the last image is generated. Gemini references the entire chat thread unless you create a new thread (which I just discovered recently).
I also have had luck with copying and pasting the image I want to refer to in the next prompt and then using "Using this image, change the character's emotion to..." or "Using this image, make it so the character is wearing a hooded cloak."
Here are a few examples of using these types of prompts to generate emotion sets. I'm going to use GIMP (and maybe Aseprite) to make MZ Face sets for these and then post a larger example set.
Changing Emotions Non-Pixel Art
Changing Emotions Pixel Art