3.80 star(s) 4 Votes

Ackerlight

Member
Donor
Jun 5, 2018
119
228
I know you are doing your own style, but I have been exerimenting some. This is my custom merge of Hassanblend v1.4 and Bondage v8. Results can be decent. These are raw unedited or inpainted outputs running locally on Stable Diffusion Automatic111 Web UI.
You don't have permission to view the spoiler content. Log in or register now.
 
Last edited:

Lumifel

Member
Game Developer
Apr 27, 2017
136
301
I know you are doing your own style, but I have been exerimenting some. This is my custom merge of Hassanblend v1.4 and Bondage v8. Results can be decent. These are raw unedited or inpainted outputs running locally on Stable Diffusion Web UI 111.
You don't have permission to view the spoiler content. Log in or register now.
Nice I really like the hair color, there's endless possibilities with AI image generation.
I like Hassanblend, results are great for NSFW.

I don't use blend actually. My most used model is F222/F111 for the bodies, dreambooth I've made for the poses and characters, 1.5 inpaint for the clothes and Waifu-diffusion for the final style and pretty much everything else. Waifu pointy ears are quite great, most natural I've seen.

I can't wait for the next version of Waifu-diffusion
 
  • Like
Reactions: Ackerlight

Ackerlight

Member
Donor
Jun 5, 2018
119
228
The Hassanblend I am using includes the F222 merge. The main thing it adds is the photorealism. This is an img2img I did. Definitely lots of possibilities if you are willing to put in the effort. If you or anyone wants my merge let me know.
You don't have permission to view the spoiler content. Log in or register now.
 
Last edited:
  • I just jizzed my pants
Reactions: Lumifel

Lumifel

Member
Game Developer
Apr 27, 2017
136
301
The Hassanblend I am using includes the F222 merge. The main thing it adds is the photorealism. This is another img2img I did. Definitely lots of possibilities if you are willing to put in the effort. If you or anyone wants my merge let me know.
You don't have permission to view the spoiler content. Log in or register now.
Indeed Stable diffusion is awesome.

A great tool for SD is Krita with interpause extension. It's sadly underappreciated by the community. But I'm doing pretty much every renders for Solaria with it
 

Ackerlight

Member
Donor
Jun 5, 2018
119
228
Yeah I haven't used that one yet. Next step for me is working on doing custom characters and things in Dreambooth. I really want to do custom embeddings and hypernets. The limiting thing for me right now is having an 3080 8gb on my laptop. So haven't decided to do cloud yet or anything since I can't currently run Dreambooth to train. I really miss Google Colab Pro when it was unlimited use for $9/mo.

I would be worried right now if I was a company like DAZ3D. Because the only thing really missing is a decent way to do animations easily. If AI can start doing that too from the generated single frames they will be in a world of hurt. It will just be a matter of time.
 
Last edited:
  • Like
Reactions: Lumifel

Lumifel

Member
Game Developer
Apr 27, 2017
136
301
Yeah I haven't used that one yet. Next step for me is working on doing custom characters and things in Dreambooth. I really want to do custom embeddings and hypernets. The limiting thing for me right now is having an 3080 8gb on my laptop. So haven't decided to do cloud yet or anything since I can't currently run Dreambooth to train. I really miss Google Colab Pro when it was unlimited use for $9/mo.

I would be worried right now if I was a company like DAZ3D. Because the only thing really missing is a decent way to do animations easily. If AI can start doing that too from the generated single frames they will be in a world of hurt. It will just be a matter of time.
i think for cloud, people use others provider like runpod. There's a lot of service better than Colab nowadays.

Regarding DAZ you are right. The biggest advantages with Ai is the price, relatively faster render times and variety. Daz models tends to all look alike. AI isn't there yet but it will get to a point where it will be easy to create characters that always look the same.
 
  • Like
Reactions: Ackerlight

Ackerlight

Member
Donor
Jun 5, 2018
119
228
I think AI will get there faster than we think. Thinking that I read somewhere that Stable Diffusion has only been around for what now? Eight months since its general public release? The 2.1 release added back in a lot of the NSFW keywords and once people can do custom merges on 2.x it will just get better. The fact that the AI can now understand context like 'A inside B', 'A next to B' and 'A on top of B' with 2.x is pretty big for realism.
 
  • Like
Reactions: Lumifel

Lumifel

Member
Game Developer
Apr 27, 2017
136
301
The new rooms are done! I just finished making the new rooms in the mansions and expanding some, like the library. I've also created several hidden areas in the rooms, but they won't be used yet. A fully explorable manor is taking shape, step by step !

In comparison to the background, the area where the characters sit or stand is more realistic (it will look strange when I inpaint the character there if it isn't).

Next step is finishing to write the plot for this version, this should be done by tomorrow. I will be able to start "coding" in Ren'py.
salon.png salon2.png
 

JLuix

Member
Jul 1, 2017
212
294
The new rooms are done! I just finished making the new rooms in the mansions and expanding some, like the library. I've also created several hidden areas in the rooms, but they won't be used yet. A fully explorable manor is taking shape, step by step !

In comparison to the background, the area where the characters sit or stand is more realistic (it will look strange when I inpaint the character there if it isn't).

Next step is finishing to write the plot for this version, this should be done by tomorrow. I will be able to start "coding" in Ren'py.
View attachment 2226267 View attachment 2226269
Thank you for all the effort, it looks magnificent.

Best regards.
 

Lumifel

Member
Game Developer
Apr 27, 2017
136
301
Some progress! I've checked and corrected the grammar of about 1/3 of the text, but it's boring, so I've started making the Ren'py code. It's less efficient, but it all feels more enjoyable.

Clicking on the Headphone pinup play a sample of the narrator voice.
1671214646329.png
 
  • Like
Reactions: JLuix

JLuix

Member
Jul 1, 2017
212
294
Some progress! I've checked and corrected the grammar of about 1/3 of the text, but it's boring, so I've started making the Ren'py code. It's less efficient, but it all feels more enjoyable.

Clicking on the Headphone pinup play a sample of the narrator voice.
View attachment 2240263
Nowadays there are people that does not realize the huge effort that is required to create a game.
Creating a game is a craft of art and passion, hard work and a lot of time.
The author needs to take care of numerous details that sometimes the players don't appreciate or value.
Some people take this for granted, but it is not simple, easy or quick.
Thanks to the author for making all of this possible.
 
Last edited:
  • Like
Reactions: Lumifel

Lumifel

Member
Game Developer
Apr 27, 2017
136
301
Nowadays there are people that does not realize the huge effort that is required to create a game.
Creating a game is a craft of art and passion, hard work and a lot of time.
The author needs to take care of numerous details that sometimes the players don't appreciate or value.
Some people take this for granted, but it is not simple, easy or quick.
Thanks to the author for making all of this possible.
Yeah I'm starting to fully contemplate this. Some details take a long time to make and when you use then in Ren'py, they are only used for a few seconds or a few renders.

The toggle for the narrator's voice is done when someone starts the game, and in the menu, customizing the name of the MC is done with a warning that the voice will keep the default name. If someone wants to enable the voice again, I think I should add an input to change the MC name back.
1671280900610.png

edit : This was easy !
1671283073159.png
 
Last edited:

Lumifel

Member
Game Developer
Apr 27, 2017
136
301
Merry Christmas to you all.

Just to let you know, development did not proceed as quickly as I had hoped, owing primarily to the ongoing war against AI waged by some artists.Several communities I'm in have been harassed, and mass reports (with false reports of CP or DeepFake) have temporarily shut down some resources, and some people have left the scene entirely because of it.

It had an effect on my motivation for a few days but it's coming back so I've resumed working on it as much as I can !
 
  • Sad
Reactions: hysepReC

JLuix

Member
Jul 1, 2017
212
294
Hello Lumifel
I hope you can recover your motivation and energy, the mental strength is fundamental for any project we want to work on.
Thanks for your consideration and communication.
 
  • Like
Reactions: Lumifel

Lumifel

Member
Game Developer
Apr 27, 2017
136
301
How I'm generating the narrator's voice.

The AI I'm using is with a unique voice I have created. It's a CLI tool.
I've also made a script that helps me generate the audio. It asks for the name, the mood, the number of candidates, and the text.


After a while, several files are generated, and I'm choosing the one that fits the best and has the best quality.
After It's time to use Audacity to remove the slight background noise of the generated file.
When it's done, the script will remove the leftover wav, backup the clean.wav to another disk, transcode it to opus, rename it, and move the audio to the game folder.

Exemple of candidats:



 
  • Like
Reactions: JLuix

JLuix

Member
Jul 1, 2017
212
294
How I'm generating the narrator's voice.

The AI I'm using is with a unique voice I have created. It's a CLI tool.
I've also made a script that helps me generate the audio. It asks for the name, the mood, the number of candidates, and the text.


After a while, several files are generated, and I'm choosing the one that fits the best and has the best quality.
After It's time to use Audacity to remove the slight background noise of the generated file.
When it's done, the script will remove the leftover wav, backup the clean.wav to another disk, transcode it to opus, rename it, and move the audio to the game folder.

Exemple of candidats:



That's a lot of work.
I vote for the #1 example.
Note the #2 and #3 have a glitch at the beginning.

Regards.
 
  • Like
Reactions: Lumifel

Lumifel

Member
Game Developer
Apr 27, 2017
136
301
That's a lot of work.
I vote for the #1 example.
Note the #2 and #3 have a glitch at the beginning.

Regards.
There's some crazy glitch sometimes, haha. Next time I have one I will post an example.
 
3.80 star(s) 4 Votes